• Author / Uploaded
• Luis Armando Pando

• Categories
• Ct Scan
• Imagen
• Medicina
• Medicina CLINICA
• Especialidades Medicas

Citation preview

High-Resolution CT of the Lung

#152201  Cust: LWW  Au: Webb  Pg. No. i

DESIGN SERVICES OF

#152201  Cust: LWW  Au: Webb  Pg. No. ii

DESIGN SERVICES OF

High-Resolution CT of the Lung FIFTH EDITION

W. Richard Webb, MD

Professor Emeritus of Radiology and Biomedical Imaging Emeritus Member, Haile Debas Academy of Medical Educators University of California San Francisco San Francisco, California

Nestor L. Müller, MD, PhD

Professor Emeritus of Radiology Department of Radiology, University of British Columbia Vancouver, British Columbia, Canada

David P. Naidich, MD, FACR, FAACP Professor of Radiology and Medicine New York University Langone Medical Center New York, New York

#152201  Cust: LWW  Au: Webb  Pg. No. iii

DESIGN SERVICES OF

Senior Executive Editor: Jonathan W. Pine, Jr. Acquisitions Editor: Ryan Shaw Product Development Editor: Amy G. Dinkel Production Project Manager: David Orzechowski Senior Manufacturing Coordinator: Beth Welsh Marketing Manager: Dan Dressler Senior Designer: Joan Wendt Production Service: S4Carlisle Publishing Services Copyright © 2015 Wolters Kluwer Health Two Commerce Square 2001 Market Street Philadelphia, PA 19103 USA LWW.com 4th edition © 2009 by LIPPINCOTT WILLIAMS & WILKINS, a WOLTERS KLUWER business All rights reserved. This book is protected by copyright. No part of this book may be reproduced in any form by any means, including photocopying, or utilized by any information storage and retrieval system without written permission from the copyright owner, except for brief quotations embodied in critical ­articles and reviews. Materials appearing in this book prepared by individuals as part of their official duties as U.S. government employees are not covered by the above-mentioned copyright. Printed in China Library of Congress Cataloging-in-Publication Data Webb, W. Richard (Wayne Richard), 1945- author. High-resolution CT of the lung / W. Richard Webb, Nestor L. Müller, David P. Naidich. — Fifth edition.    p. ; cm. Includes bibliographical references and index. ISBN 978-1-4511-7601-8 (alk. paper) I. Müller, Nestor Luiz, 1948- , author. II. Naidich, David P., author. III. Title. [DNLM:  1. Lung—radiography.  2. Tomography, X-Ray Computed.  3. Lung Diseases—pathology.    WF 600] RC734.T64 616.2’407572—dc23 2014003388 Care has been taken to confirm the accuracy of the information presented and to describe generally accepted practices. However, the authors, editors, and publisher are not responsible for errors or omissions or for any consequences from application of the information in this book and make no warranty, expressed or implied, with respect to the currency, completeness, or accuracy of the contents of the publication. Application of the information in a particular situation remains the professional responsibility of the practitioner. The authors, editors, and publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accordance with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any change in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new or infrequently employed drug. Some drugs and medical devices presented in the publication have Food and Drug Administration (FDA) clearance for limited use in restricted research settings. It is the responsibility of the health care providers to ascertain the FDA status of each drug or device planned for use in their clinical practice. To purchase additional copies of this book, call our customer service department at (800) 638-3030 or fax orders to (301) 223-2320. International customers should call (301) 223-2300. Visit Lippincott Williams & Wilkins on the Internet: at LWW.com. Lippincott Williams & Wilkins customer service representatives are available from 8:30 am to 6 pm, EST. 10 9 8 7 6 5 4 3 2 1

#152201  Cust: LWW  Au: Webb  Pg. No. iv

DESIGN SERVICES OF

DEDICATION To my father, who encouraged my curiosity and taught me to figure things out ––WRW

To my wife, Isabela, and my children—Alison, Phillip, and Noah Müller ––NLM

To Jocelyn, whose constant love and support has always been my greatest inspiration ––DPN

#152201  Cust: LWW  Au: Webb  Pg. No. v

DESIGN SERVICES OF

#152201  Cust: LWW  Au: Webb  Pg. No. vi

DESIGN SERVICES OF

Contributing Authors Brett M. Elicker, MD

Associate Professor of Clinical Radiology and Biomedical Imaging Chief, Cardiac and Pulmonary Imaging University of California San Francisco San Francisco, California

Myrna C. B. Godoy, MD, PhD

Assistant Professor of Radiology University of Texas MD Anderson Cancer Center Houston, Texas

C. Isabela S. Müller, MD, PhD Department of Radiology Delfin Clinic Salvador, Bahia, Brazil

vii

#152201  Cust: LWW  Au: Webb  Pg. No. vii

DESIGN SERVICES OF

#152201  Cust: LWW  Au: Webb  Pg. No. viii

DESIGN SERVICES OF

Preface During the past 25 years, high-resolution CT (HRCT) has become established as an indispensable tool in the evaluation of patients with diffuse lung disease. HRCT is now commonly used in clinical practice to detect and characterize a variety of lung abnormalities. In the approximately 5 years since our fourth edition was published, considerable progress has taken place in the understanding of diffuse lung diseases and the recognition of new entities and their nature, causes, and characteristics. Without doubt, HRCT has played a fundamental role in contributing to this progress and has become essential to the diagnosis of a number of diffuse diseases. This fifth edition continues what the three of us, independently, in conjunction, and with each other’s encouragement and support, began some 30 years ago. The photograph of the three of us below was taken by a ­local resident at the 1989 Diagnostic Course in Davos, on a walk we took on the promenade above the Sweitzerhof on the day of our arrival, when as junior faculty, we were more than a little anxious about teaching along with such important and impressive chest radiologists as Fraser, Felson, Greenspan, Milne, Flowers, Heitzman, and many others.

At this meeting, we each spoke about the use of HRCT, which, at the time, was a little-known technique that was regarded with skepticism by many radiologists. We learned from each other as we spoke, compared slides in the speaker-ready room, and gained confidence from our shared opinions. At this meeting, we began thinking about a collaboration that would combine our experience and thoughts about this new modality and its potential uses. Our first edition of this book was published in late 1991, with a grand total of 159 pages. It was a quarter

of an inch thick, and, to our knowledge, referenced every known paper on HRCT. From our perspective, it was the most important thing we had ever done. That is how things start. Maybe that is the best way things should start. It was certainly fun and rewarding for each of us. And we three have stuck together over the years, out of our combined respect, admiration, friendship, and good humor. Each one of us believes that we learned more from our collaboration than we taught. In this edition, we have incorporated an update and review of numerous recent advances in the classification and understanding of diffuse lung diseases and their HRCT features. Recent technical modifications in obtaining HRCT have also been reviewed, most notably the use of helical HRCT and dose-reduction techniques. We hope the reader will find these changes and updates helpful. As is our wont, we have reorganized our discussions into new sections and chapters, which we feel best presents the most important topics in HRCT diagnosis for reference and learning. A new section has been added at the end of the book to provide a general review of HRCT, including an illustrated glossary of HRCT terms and a chapter providing a compilation of the common and typical appearances of the most common diffuse lung diseases encountered in clinical practice. These sections are intended to provide an illustrated index to the detailed descriptions of diseases found elsewhere in the book. It is with a great deal of pride that we complete our fifth edition of this book, which has occupied so much of our thoughts, efforts, and time over the years. This task is accomplished in the hope that this book will encourage future generations of thoracic imagers to develop mutually productive relationships with friends and colleagues, in order to explore important questions in our understanding of the role of imaging in the assessment of thoracic disease. To this end, we acknowledge the contributions of three esteemed colleagues, our former fellows, who have authored parts of this book. Their efforts have greatly inspired our own enthusiasm for the considerable task of bringing this edition to fruition. W. Richard Webb  Nestor L. Müller  David P. Naidich ix

#152201  Cust: LWW  Au: Webb  Pg. No. ix

DESIGN SERVICES OF

#152201  Cust: LWW  Au: Webb  Pg. No. x

DESIGN SERVICES OF

Acknowledgments We wish to gratefully acknowledge the many colleagues who have provided us with insights and inspiration over the years, and allowed us to use their illustrations for this

and prior editions of this book. Although they are too numerous to mention here, they are recognized throughout the following pages.

xi

#152201  Cust: LWW  Au: Webb  Pg. No. xi

DESIGN SERVICES OF

#152201  Cust: LWW  Au: Webb  Pg. No. xii

DESIGN SERVICES OF

Contents S E C T I O N

I

High-Resolution CT Techniques and Normal Anatomy  1

1 2

Technical Aspects of High-Resolution CT   2 Normal Lung Anatomy   47

S E C T I O N

II

Approach to HRCT Diagnosis and Findings of Lung Disease  71

3 4 5 6 7

HRCT Findings: Linear and Reticular Opacities  74 HRCT Findings: Multiple Nodules and Nodular Opacities   106 HRCT Findings: Parenchymal Opacification   140 HRCT Findings: Air-Filled Cystic Lesions   165 HRCT Findings: Decreased Lung Attenuation   .187

S E C T I O N

III

High-Resolution CT Diagnosis of Diffuse Lung Disease  207

8

The Idiopathic Interstitial Pneumonias, Part I: Usual Interstitial Pneumonia/ Idiopathic Pulmonary Fibrosis and Nonspecific Interstitial Pneumonia   208

9

The Idiopathic Interstitial Pneumonias, Part II: Cryptogenic ­Organizing Pneumonia, Acute Interstitial Pneumonia, Respiratory Bronchiolitis-­ Interstitial Lung Disease, Desquamative Interstitial Pneumonia, Lymphoid Interstitial Pneumonia, and Pleuroparenchymal Fibroelastosis   232

10 11

Collagen-Vascular Diseases   256 Diffuse Pulmonary Neoplasms and Pulmonary Lymphoproliferative Diseases  280

12 Sarcoidosis  312 13 Pneumoconiosis, Occupational, and Environmental Lung Disease   342 xiii

#152201  Cust: LWW  Au: Webb  Pg. No. xiii

DESIGN SERVICES OF

xiv

Contents

14 Hypersensitivity Pneumonitis and Eosinophilic Lung Diseases   376 15 Drug-Induced Lung Diseases and Radiation Pneumonitis   397 16 Miscellaneous Infiltrative Lung Diseases   411 17 Infections  429 18 Pulmonary Edema and Acute Respiratory Distress Syndrome   481 19 Cystic Lung Diseases   492 20 Emphysema and Chronic Obstructive Pulmonary Disease   517 21 Airways Diseases   552 22 Pulmonary Hypertension and Pulmonary Vascular Disease   622 S E C T I O N

IV

High-Resolution CT Review 

23 24

659

Illustrated Glossary of High-Resolution CT Terms   660 Appearances and Characteristics of Common Diseases   678

#152201  Cust: LWW  Au: Webb  Pg. No. xiv

DESIGN SERVICES OF

S E C T I O N

I

High-Resolution CT Techniques and Normal Anatomy

#152201  Cust: LWW  Au: Webb  Pg. No. 1

DESIGN SERVICES OF

1

Technical Aspects of High-Resolution CT I M P O R T A N T

T O P I C S

HIGH-RESOLUTION COMPUTED TOMOGRAPHY: FUNDAMENTAL TECHNIQUES  2

ADDITIONAL TECHNICAL MODIFICATIONS  31

TECHNIQUES OF SCAN ACQUISITION: SPACED AXIAL SCANNING VERSUS VOLUMETRIC SCANNING  10

HIGH-RESOLUTION COMPUTED TOMOGRAPHY PROTOCOLS 36

RADIATION DOSE  20

SPATIAL RESOLUTION OF HIGH-RESOLUTION COMPUTED TOMOGRAPHY  38

EXPIRATORY HIGH-RESOLUTION COMPUTED TOMOGRAPHY 24 QUANTITATIVE COMPUTED TOMOGRAPHY  30

Abbreviations Used in This Chapter ASIR BOS COPD CTDI DLP ECG FBP FOV HU kV kV(p) MIP mA mAs mGy mSv MinIP MBIR MDCT MD-HRCT NSIP ROI 3D 2D

adaptive statistical iterative reconstruction bronchiolitis obliterans syndrome chronic obstructive pulmonary disease CT dose index dose length product electrocardiographic filtered back projection field of view Hounsfield units kilovolt kilovolt peak maximum-intensity projection milliampere milliampere seconds milligray millisievert minimum-intensity projection model-based iterative reconstruction multidetector helical computed tomography multidetector helical HRCT nonspecific interstitial pneumonia region of interest three-dimensional two-dimensional

High-resolution computed tomography (HRCT) is capable of imaging the lung with excellent spatial resolution, providing anatomical detail similar to that available from gross pathologic specimens and paper-mounted lung slices (1–4). HRCT can readily demonstrate the normal and abnormal

IMAGE DISPLAY  33

HIGH-RESOLUTION COMPUTED TOMOGRAPHY ARTIFACTS 39

lung interstitium and morphologic characteristics of both localized and diffuse parenchymal abnormalities; in this regard, HRCT is clearly superior to plain radiographs. The first use of the term high-resolution computed ­tomography has been attributed to Todo et al. (5), who, in 1982, described the potential use of this technique for assessing lung disease. The first reports of HRCT in English date to 1985, including landmark descriptions of HRCT findings by Nakata et al., Naidich et al., and Zerhouni et al. (6–8). Since then, HRCT has become established as an important diagnostic tool in pulmonary medicine and has significantly contributed to our understanding of diffuse lung diseases. Although many of the HRCT techniques used in these initial studies are still appropriate today, the recent development of multidetector helical computed tomography (MDCT) scanners capable of volumetric high-resolution scanning has significantly changed the manner in which HRCT may be obtained. In this chapter, we review computed tomography (CT) techniques that are appropriate for obtaining HRCT in patients with suspected lung disease, scan protocols recommended in specific clinical settings, the spatial resolution and radiation dose associated with HRCT, and common HRCT artifacts.

HIGH-RESOLUTION COMPUTED TOMOGRAPHY: FUNDAMENTAL TECHNIQUES This section reviews the effect of various technical factors on the appearance of HRCT and summarizes our recommendations for obtaining appropriate examinations.

2

#152201  Cust: LWW  Au: Webb  Pg. No. 2

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

Although each author performs HRCT in a different manner, we generally agree as to what fundamental techniques constitute a “high-resolution” CT study. Quite simply, these include (a) the use of thin-collimation axial scans or thin-section reconstruction of volumetric data obtained using MDCT and narrow detector width (0.5– 1.25 mm) and (b) image reconstruction with a high spatial frequency (sharp or high-resolution) algorithm. Sufficient radiation (in milliampere seconds [mAs] or effective mAs [mAs/pitch for helical scans]) (9) must be used to keep image noise at a level low enough to allow accurate image interpretation, while keeping patient exposure at appropriate levels; keep in mind that dose reduction techniques can be used while obtaining diagnostic scans (Table 1-1) (1–4,10–12). Targeted image reconstruction may be used to reduce pixel size, but is not necessary for clinical diagnosis in most settings (Table 1-1) (1–4,10–12).

Slice Thickness The use of thin sections (0.5–1.5 mm) is essential if spatial resolution and lung detail are to be optimized (4,6,8,10) (Table 1-1). Generally, 1-mm-thick slices are adequate for diagnosis; a clear-cut advantage for thinner slices has not been shown (13). With slices thicker than 1 to 1.5 mm, volume averaging within the plane of scan significantly

reduces the ability of CT to resolve small structures. The use of 2.5- to 5-mm slice thickness should not be considered adequate for HRCT. In an early study, Murata et al. (12) compared the ability of axial HRCT performed with 1.5- and 3-mm collimation to allow the identification of small vessels, bronchi, interlobular septa, and some pathologic findings. With 1.5-mm collimation, greater contrast was evident between vessels and surrounding lung parenchyma, more branches of small vessels were seen, and small bronchi were more often recognizable than with 3-mm collimation (12). Also, slight increases in lung attenuation (as may be seen in early interstitial lung disease), or decreases in attenuation (as in emphysema), were better resolved with 1.5-mm collimation. However, the authors concluded that certain pathologic findings, such as thickened interlobular septa, were similarly visible on images with 1.5- and 3-mm collimation (12). There are several differences in how lung structures are visualized on scans performed with thin (e.g., 1-mm) and thick (e.g., 5-mm) sections. With thin slices, it is more difficult to follow the courses of vessels and bronchi than it is with thick slices. With thick slices, for example, vessels that lie in the plane of scan look like vessels (i.e., they appear cylindrical or branching) and can be clearly identified as such. With thin slices, vessels can appear round or

TABLE 1-1  Summary of HRCT Techniques Recommended Slice thickness: thinnest available (0.5–1.5 mm) Reconstruction algorithm: high spatial frequency or “sharp” algorithm kV(p) 120; 100 or 80 for small or pediatric patients mA less than 250; mAs (effective) of 100 or less Scan (rotation) time: as short as possible (e.g., 0.3–0.5 s) Pitch (MD-HRCT): 1-1.5 Inspiratory level: full inspiration Position: supine; prone scans routinely in patients with suspected interstitial lung disease; in patients with minimal or unknown chest film abnormalities, or monitor supine scans for dependent density Acquisition: spaced axial imaging or MD-HRCT Expiratory imaging: postexpiratory scans at three or more levels in patients with obstructive disease Reconstruction: transaxial; entire thorax Windows: at least one consistent lung window setting is necessary. Window mean/width values of 600–700 HU/1,000–1,500 HU are appropriate. Good combinations are 700/1,000 HU or 600/1,500 HU. Soft-tissue windows of approximately 50/350 HU should also be used for the mediastinum, hila, and pleura. Image display: workstation (optimal) or photography of lung windows 12 on 1 Optional Reduced mAs: low-dose axial HRCT or MD-HRCT best for follow-up studies Acquisition: ECG gating or segmented reconstruction to reduce motion artifacts Expiratory imaging: dynamic, volumetric, or spirometrically triggered expiratory scans Contrast injection: patients with suspected vascular disease Reconstruction: targeted (15- to 25-cm FOV; 2D or 3D reconstruction; MIP or MinIP reconstructions) Windows: windows may need to be customized; a low window mean (800–900 HU) is optimal for diagnosing emphysema. For viewing the mediastinum, 50/350 HU is recommended. For viewing pleuroparenchymal disease, 600/2,000 HU is recommended

#152201  Cust: LWW  Au: Webb  Pg. No. 3

3

DESIGN SERVICES OF

4

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

B

A

FIGU RE 1-1  Effect of slice thickness on resolution. A: Helical CT with 5-mm slice thickness, reconstructed with the standard algorithm in a normal subject. A number of

branching pulmonary vessels are visible (arrows). B: Helical CT at the same level with 1.25-mm slice thickness reconstructed with the same scan data and algorithm. Pulmonary arteries seen as branching or cylindrical on the thicker scan appear “nodular” on the scan with 1.25-mm slice thickness (arrows). The resolution is clearly improved with thin slices.

oval (i.e., nodular) because only short segments may lie in the plane of scan (Fig. 1-1). With experience, this difficulty is easily avoided. Also, with thin slices, the diameter of a vessel that lies in or near the plane of scan can appear larger than it does with thicker slices because less volume averaging occurs between the rounded edge of the vessel and the adjacent air-filled lung; thin scans more accurately reflect vessel diameter in this setting, analogous to the better estimation of the diameter of a lung nodule that is possible with thin slices. Furthermore, with thin slices, bronchi that are oriented obliquely relative to the scan plane are much better defined than they are with thicker slices, and their wall thicknesses and luminal diameters are more accurately assessed (14). The diameters of vessels or bronchi that lie perpendicular to the scan plane appear the same with both thin and thick collimation.

Reconstruction Algorithm The inherent or maximum spatial resolution of a CT scanner is determined by the geometry of the data-collecting system and the frequency at which scan data are sampled during the scan sequence (10). The spatial resolution of the image produced is less than the inherent resolution of the scan system, depending on whether axial or volumetric (helical) imaging is used, the reconstruction algorithm, the matrix size, and the field of view (FOV), all of which in turn determine pixel size. In HRCT, these parameters are optimized to increase the spatial resolution of the image. With body CT, scan data are usually reconstructed with a relatively low spatial frequency algorithm (e.g., “standard” or “soft-tissue” algorithms) that smoothes the image, reduces visible image noise, and improves the contrast resolution to some degree (11,15). Low spatial frequency simply means that the frequency of information

#152201  Cust: LWW  Au: Webb  Pg. No. 4

recorded in the final image is relatively low; it is the same as saying that the algorithm is low resolution rather than high resolution. Reconstruction of images using a sharp, high spatial frequency, or high-resolution algorithm reduces i­mage smoothing and increases spatial resolution, making structures appear sharper (Figs. 1-2 to 1-4) (6,10,12,16). Using a high-resolution algorithm is a critical element in performing HRCT (Table 1-1) (11,15). In one study of HRCT techniques (10), the use of a high spatial frequency algorithm to reconstruct scan data resulted in a quantitative improvement in spatial resolution when compared to a standard algorithm (Fig. 1-3); in this study, subjective image quality was also rated more highly with the high spatial frequency algorithm. In another study of HRCT (12), small vessels and bronchi were better seen when images were reconstructed with a high-resolution algorithm than when the standard algorithm was used. The use of a sharp algorithm has also been recommended to improve spatial resolution for routine chest CT reconstructed with thicker slices (17).

Kilovolts (Peak), Milliamperes, and Scan Time Using a sharp or high-resolution reconstruction algorithm, in addition to increasing image detail, increases the visibility of noise in the CT image (11,15). This noise usually appears as a graininess, mottle, or streaks that can be distracting and may obscure anatomical detail (Fig. 1-4) (10). Because much of this noise is quantum related, it is inversely proportional to the number of photons absorbed (precisely, it is inversely proportional to the square root of the product of mA and scan time) (16). Consequently, it increases with decreasing mAs or kilovolt peak (kV(p)) and decreases with increased mAs or kV(p) (Fig. 1-5) (10,16). For example, in one study

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

A

5

B

FIGU RE 1-2  Effect of reconstruction algorithm on resolution. MD-HRCT obtained with 1.25-mm slice thickness in a patient with usual interstitial pneumonia has been reconstructed using a high-resolution (sharp) algorithm (A) and a smooth (standard) algorithm (B). Lung structures, reticular opacities, and traction bronchiectasis are much more sharply defined with the high-resolution algorithm.

using an early-generation scanner (10), a measure of image noise was reduced by approximately 30% when kV(p)/mAs were increased from 120/200 to 140/340 (Fig. 1-5), and the scans with increased kV(p) and mAs settings were rated as being of better quality in 80% of cases (Fig. 1-6) (10). Although increasing mAs or kV(p) above routine values can reduce image noise, it is not necessary for obtaining adequate HRCT images, and maintenance of patient radiation dose at a reasonable level is considered to be more important (16). With current scanners and reconstruction algorithms, diagnostic scans can be obtained using mAs and kV(p) techniques considered routine for chest CT. Scan techniques with a kV(p) of 120 are generally used, although a reduced kV(p) of 100 or 80 may be

A

used in small or pediatric patients (i.e., less than 80 or 60 kg) (13). Using mAs (or effective mAs) values of 100 or less has proven satisfactory for obtaining HRCT in most patients with current-generation scanners (13,18). Increased patient size and increased chest wall thickness are associated with increased image noise; this may be reduced with increased mA (Fig. 1-7) (10). Reducing mA to 40 (i.e., low-dose CT) may be used to reduce image dose, but this should generally be reserved for small or pediatric patients. Image noise may be excessive with low mA settings in large patients (Fig. 1-8). Specific mA, kV(p), pitch (with helical scanning), and gantry rotation times most appropriate for HRCT vary with different scanners. When obtaining helical HRCT,

B

FIGU RE 1-3  Effect of reconstruction algorithm on spatial resolution. A: HRCT of a line-pair phantom obtained with 1.5-mm collimation and reconstructed with the standard algorithm. Numbers indicate the resolution in line pairs per centimeter. The resolution with this technique is 6 line pairs per centimeter. B: When the same scan is reconstructed using the high-resolution (i.e., bone) algorithm, spatial resolution improves. Also, in contrast to the scan reconstructed using the standard algorithm, 7.5 line pairs are easily resolved (arrow), and edges are considerably sharper. (From Mayo JR, Webb WR, Gould R, et al. High-resolution CT of the lungs: an optimal approach. Radiology 1987;163:507, with permission.)

#152201  Cust: LWW  Au: Webb  Pg. No. 5

DESIGN SERVICES OF

6

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

FIGU RE 1-4  Effect of reconstruction algorithm on resolution and image noise. A 1.25-mm MD-HRCT has been reconstructed with high-resolution (A) and standard (B) algorithms. A: The image reconstructed with the high-resolution algorithm is sharper and shows more detail, but streak artifacts due to aliasing and noise are more apparent. B: Resolution is diminished with this algorithm. The image appears smoother with this algorithm, and noise is less apparent.

the use of dynamic, modulated, or adaptive mA that varies with body thickness should generally be used to keep radiation dose low, without sacrificing image quality (19). In large patients, a reasonable maximum mA should be set when using this technique, to avoid inappropriately high exposures. Because of artifacts related to patient motion, breathing, and cardiac pulsation, it is desirable to minimize scan or gantry rotation time. A scan time or gantry rotation time of 0.5 s or less is optimal for HRCT and, if available, is recommended (Table 1-1). Most current scanners have gantry rotation times of 300 to 500 ms.

Field of View and Targeted Reconstruction Scanning should be performed using the smallest FOV that will encompass the patient (e.g., 35 cm), as this reduces pixel size. Retrospectively targeting image reconstruction to a single lung instead of the entire thorax significantly

A

60 50 Bone algorithm

Noise (HU)

40

120 kV(p) 140 kV(p)

30 20 10

60

Standard algorithm

80

100

120 mA

140

160

120 kV(p) 140 kV(p)

180

FIGU RE 1-5  Effect of algorithm, kV(p), and mA on image noise. Graph of HRCT image noise (SD of HU measurements) in an anthropomorphic CT phantom as related to the reconstruction algorithm and scan technique. Noise increases when the bone (high-resolution) algorithm is used instead of the standard algorithm. With the bone algorithm, noise decreases approximately 30% with increased kV(p) and mA settings. (From Mayo JR, Webb WR, Gould R, et al. High-resolution CT of the lungs: an optimal approach. Radiology 1987;163:507, with permission.)

#152201  Cust: LWW  Au: Webb  Pg. No. 6

B FIGU RE 1-6  A and B: Effect of kV(p) and mA on image noise. Axial HRCT obtained with a tube current of 100 mA (A) and 400 mA (B) in a patient with atypical mycobacterial infection. There is a relative increase in noise in A, which is evident both in the soft tissues and lung. Note, however, that the lower-dose scan (A) is still of diagnostic quality.

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

60

7

Bone algorithm

50 Thick chest wall Noise (HU)

40 120 kV(p) 30

140 kV(p)

20

120 kV(p) 140 kV(p)

Thin chest wall

10

60

A 80

100

120 mA

140

160

180

FIGU RE 1-7  Relationship of noise to patient size. Graph of image noise measured using an anthropomorphic chest phantom, with simulated thick and thin chest walls. Noise significantly increases with the thick chest wall. (From Mayo JR, Webb WR, Gould R, et al. High-resolution CT of the lungs: an optimal approach. Radiology 1987;163:507, with permission.)

reduces the FOV and image pixel size, and thus increases spatial resolution (Figs. 1-9 and 1-10) (10,20,21). For ­example, with a 40-cm reconstruction circle (FOV) and a 512 × 512 matrix, pixel size measures 0.78 mm. With targeted image reconstruction using a 25-cm FOV, pixel size is reduced to 0.49 mm, and the spatial resolution is correspondingly increased (Fig. 1-9). Using a 15-cm FOV further reduces pixel size to 0.29 mm, but this FOV is usually insufficient to view an entire lung and is not often used clinically. It should be recognized, however, that the improvement in resolution obtainable by targeting is limited by the intrinsic resolution of the detectors used. The use of targeted reconstruction is often a matter of personal preference. In clinical practice, the use of image targeting is uncommon because it requires additional reconstruction time, the raw scan data must be saved until targeting is performed, and display of the individual lung images is somewhat cumbersome. With a nontargeted reconstruction, the ability to see both lungs on the same image allows a quick comparison of one lung to the other; this can be quite helpful in diagnosis and is preferred to the marginal increase in resolution achieved with targeting.

Inspiratory Level Routine HRCT is obtained during suspended full inspiration, which (a) optimizes contrast between normal structures, various abnormalities, and normal aerated lung parenchyma; and (b) reduces transient atelectasis, a finding that may mimic or obscure significant abnormalities. Selected scans obtained during or after forced expiration may also be valuable in diagnosing patients with

#152201  Cust: LWW  Au: Webb  Pg. No. 7

B

C F I G U R E 1-8  A–C: Low-dose (40 mA) axial HRCT in a large patient. Images through the upper (A), mid- (B), and lower (C) lungs are shown from a normal HRCT obtained at 1-cm intervals in the supine position in inspiration using a fixed tube current (40 mA). Dynamic expiratory images were also obtained at three selected levels. The estimated effective dose for this examination was 0.2 mSv. However, image noise is excessive, and subtle abnormalities may be difficult to detect.

obstructive lung disease or airway abnormalities. The use of expiratory HRCT is discussed later in this chapter, and in Chapters 2 and 7.

DESIGN SERVICES OF

8

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

FIGU RE 1-9  Effect of targeted reconstruction on resolution. A: HRCT image in a patient with end-stage sarcoidosis obtained with a 38-cm FOV and 1.5-mm collimation, and reconstructed using a high-resolution algorithm and a 38-cm reconstruction circle. B: The same CT scan has been reconstructed using a targeted FOV (15 cm), reducing image pixel diameter. Image sharpness is improved compared to A.

Patient Position and the Use of Prone Scanning Scans obtained with the patient supine are adequate for diagnosis in most instances. However, scans obtained with the patient positioned prone are sometimes necessary for diagnosing subtle lung abnormalities. Atelectasis is commonly seen in the dependent lung (i.e., posterior lung on supine scans) in both normal and abnormal subjects, resulting in a so-called dependent density or

A

subpleural line (Fig. 1-11) (22,23). These normal findings can closely mimic the appearance of early lung fibrosis, and they can be impossible to distinguish from true pathology on ­supine scans alone. However, if scans are obtained in both supine and prone positions, dependent density can be easily differentiated from true pathology. Normal dependent density disappears in the prone position (Fig. 1-11); a true abnormality remains visible regardless of whether it is dependent or nondependent (Figs. 1-12 and 1-13).

B

FIGU RE 1-10  Effect of targeted reconstruction on spatial resolution. A: HRCT of a line-pair phantom. The scan was obtained with a 40-cm FOV, and reconstructed using a targeted FOV of 25 cm. The resolution with this technique is 7.5 line pairs (arrow). B: The same scan viewed without targeting shows the effects of larger pixel size. Only 6 line pairs can be resolved (arrow), and the margins of the lines appear jagged or wavy. (From Mayo JR, Webb WR, Gould R, et al. High-resolution CT of the lungs: an optimal approach. Radiology 1987;163:507, with permission.)

#152201  Cust: LWW  Au: Webb  Pg. No. 8

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

A

9

B

FIGU RE 1-11  Transient dependent density. A: Supine scan shows ill-defined opacity in the posterior lungs (arrows). B: On a prone image, the posterior lung appears normal. Note that some dependent opacity is now visible in the anterior lung.

Dependent density results in a diagnostic dilemma only in patients who have normal lungs or subtle lung abnormalities. In patients with obvious abnormalities, such as honeycombing, or in patients with diffuse lung disease, dependent density is not usually a diagnostic problem. Thus, if the patient being studied has evidence of moderate-­to-severe lung disease on plain radiographs, prone scans are not likely to be needed. However, if the patient is suspected of having an interstitial abnormality and the plain radiograph is normal or near normal, or the results of chest radiographs are unknown, prone scans may prove helpful. In addition, even in patients with obvious lung disease on supine scans, prone scans may prove useful in identifying specific important diagnostic findings (i.e., subtle posterior lung honeycombing), not clearly seen on the supine images. Volpe et al. (24) assessed the usefulness of prone scans in patients who had chest radiographs read as normal, possibly abnormal, or definitely abnormal. Overall, prone scans were considered helpful in 17 of 100 consecutive

patients having HRCT (24). Prone HRCT scans were helpful in confirming or ruling out posterior lung abnormalities in 10 of 36 (28%) patients who had normal findings on chest radiographs, 5 of 18 (28%) patients who had possibly abnormal findings on chest radiographs, and only 2 of 46 (4%) patients who had definitely abnormal findings on chest radiographs. The proportion of patients who benefited from prone scans was significantly lower among the patients with abnormal findings on chest ­radiographs than among the patients with normal (p = 0.008) or possibly abnormal (p = 0.02) findings. The two patients who had abnormal findings on radiographs and in whom CT scans obtained with the patient prone were helpful had minimal radiographic abnormalities. Some investigators (21,25) obtain HRCT in the prone position only when dependent lung collapse is problematic (26); however, this approach requires that the scans be closely monitored or that the patient be called back for additional scans. Others use prone scanning in specific clinical settings, for example, when

B

A

FIGU RE 1-12  Persistent opacity in the posterior lung in a patient with mild pulmonary fibrosis. A: Supine scan shows ill-defined opacity in the posterior lungs and in a subpleural region anteriorly. B: On a prone image, the posterior lung is unchanged in appearance, indicative of lung disease.

#152201  Cust: LWW  Au: Webb  Pg. No. 9

DESIGN SERVICES OF

10

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

FIGU RE 1-13  Persistent posterior lung ground opacity on prone scans in a patient with scleroderma and NSIP. A: Supine scan shows ill-defined opacity in the posterior lungs. B: On a prone image, the posterior subpleural lung opacity is unchanged in appearance, and the presence of true lung disease can be diagnosed.

asbestosis or early lung fibrosis is suspected, whereas still others obtain prone scans routinely (22,27). In patients who are suspected of having emphysema, airways disease such as bronchiectasis, or another obstructive lung disease, dependent atelectasis is not usually a diagnostic problem, and prone scans are not usually needed. Spaced axial prone scans, prone scans clustered near the lung bases, or volumetric helical imaging in the prone position may all be used. Some protocols call for prone volumetric imaging only (i.e., no supine scans are obtained) (28); this would be most useful in a patient suspected of having a disease with a posterior lung predominance, such as asbestosis or idiopathic pulmonary fibrosis.

TECHNIQUES OF SCAN ACQUISITION: SPACED AXIAL SCANning VERSUS VOLUMETRIC SCANNING

When spaced axial scanning is chosen for HRCT, we consider scans obtained at 1-cm intervals, from the lung apices to bases, to be the most appropriate routine

A

Before the introduction of MDCT scanners, HRCT was performed by obtaining individual scans at spaced intervals. This technique remains in use today. However, the development of MDCT scanners, capable of rapidly imaging the thorax using an isotropic technique, has greatly expanded the ways in which a HRCT study may be ­obtained (13).

Spaced Axial Scans HRCT may be performed with individual axial scans being obtained at spaced intervals, usually 1 to 2 cm, without table motion (Figs. 1-14 and 1-15). In this manner, HRCT is intended to “sample” lung anatomy, with the assumptions being that (a) a diffuse lung disease will be visible in at least one of the levels sampled and (b) the findings seen at the levels scanned will be representative of what is present throughout the lung. These assumptions have proven valid during more than 20 years of ­experience with HRCT (29).

#152201  Cust: LWW  Au: Webb  Pg. No. 10

B FIGU RE 1-14  A and B: Comparison of prone 1.25-mm spaced axial HRCT (A) and 1.25-mm MD-HRCT (B) in a patient with scleroderma and fibrotic NSIP. Two prone HRCT images at the same level are shown in a patient with sclerodermarelated NSIP. While of similar diagnostic quality, the axial HRCT (A) has slightly better resolution and the structures and abnormalities appear sharper than on the helical HRCT (B).

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

A

11

B

FIGU RE 1-15  Comparison of 1.25-mm spaced axial HRCT (A) and

1.25-mm MD-HRCT (B and C) in a patient with mixed connective tissue disease and NSIP. A: Axial 1.25-mm HRCT shows irregular reticulation, ground-glass opacity, and traction bronchiectasis with lower-lobe predominance. Subpleural sparing is present. These findings are typical of NSIP. 2D and 3D reconstructed images from the MD-HRCT are also shown in Fig. 1-16. B and C: Comparable levels from the MD-HRCT show identical findings. There is no significant difference in diagnostic value of the axial and MD-HRCT images, although the MD-HRCT images appear slightly smoother.

C

scanning protocol, allowing an adequate sampling of the lung and lung disease regardless of its distribution. In early reports, HRCT scanning was sometimes performed with scans at 2-, 3-, and even 4-cm intervals (3,26); at three preselected levels (25); or at one or two levels through the lower lungs (21). Although such wide spacing may be sufficient for assessing some patients and some lung diseases, in many cases, these protocols would prove inadequate for initial diagnosis. It should be pointed out, however, that in patients with a known disease, a limited number of HRCT images may be sufficient to assess disease extent. For example, in one study (30), the ability of HRCT obtained at three selected levels (limited HRCT) to show features of idiopathic pulmonary fibrosis was compared to that of HRCT obtained at 10-mm increments (complete HRCT). HRCT fibrosis scores strongly correlated with pathology fibrosis scores for both the complete (r = 0.53, p = 0.0001) and limited (r = 0.50, p = 0.0001) HRCT examinations. HRCT ground-glass opacity scores also correlated with the histologic inflammatory scores on the complete (r = 0.27, p  = 0.03) and limited (r = 0.26, p = 0.03) HRCT examinations. Similarly, in evaluating patients with asbestos exposure, several investigators have suggested that a limited number of scans should be sufficient for the diagnosis of asbestosis (22,27,31–34). Obtaining four or five scans near the lung bases has proved to have good sensitivity in patients with suspected asbestosis (35). Thickslice CT, combined with a few HRCT images, has also

#152201  Cust: LWW  Au: Webb  Pg. No. 11

been applied to patients with suspected diffuse lung disease and has been shown to be clinically efficacious (35); HRCT scans obtained at the levels of the aortic arch, carina, and 2 cm above the right hemidiaphragm allow the assessment of the lung regions in which lung biopsies are most frequently ­performed (11). In patients who are likely to require prone images, prone scans can be added to the routine supine sequence obtained with 1-cm scan spacing; a reasonable protocol would include additional prone scans at 2-cm intervals. Although axial imaging is a low-dose technique, if further radiation dose reduction is a concern, scans could be obtained at 2-cm intervals from lung apices to bases, in both supine and prone positions. Because the prone and supine images will be slightly different, even if an attempt is made to obtain the scans at exactly the same levels, the number of different levels scanned will be equivalent to a supine position scan protocol using 1-cm spacing. Another dose reduction technique used with axial imaging is to customize the number or location of scans, depending on the patient’s suspected disease, clinical findings, or the location of plain radiographic abnormalities. For example, if the lung disease being studied predominates in a certain region of lung, as determined by chest radiographs, conventional CT (21), or other imaging studies, it makes sense that more scans should be obtained in the most abnormal area. In patients with suspected asbestosis, it has been recommended that more scans be performed near the diaphragm than in the upper

DESIGN SERVICES OF

12

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

lobes because of the typical basal distribution of this disease, even if the chest radiograph does not suggest an abnormality in this region (22,27). Some support for this approach has been lent by a paper (36) describing theoretical methods useful in selecting the appropriate number of HRCT images for estimating any quantitative parameter of lung disease. A marked reduction in the number of images necessary for quantification of a desired parameter can be achieved by using a stratified sampling technique based on prior knowledge of the disease distribution. Spaced axial HRCT scans may be obtained in combination with a volumetric helical CT study, in which the entire thorax is imaged (11,21,25), although this is rarely needed in current practice. If both volumetric imaging and HRCT are needed for diagnosis, it is usually adequate to reconstruct the volumetric scan with thin slices and a sharp algorithm. However, Leswick et al. (37) compared the patient radiation dose from a combination of spaced axial HRCT and volumetric helical MDCT to that from a volumetric helical HRCT having a noise level similar to that of the axial scans. The authors found that the volumetric helical HRCT had a radiation dose 32% higher than that in the combined study (37).

Volumetric High-Resolution Computed Tomography The use of MDCT scanners capable of rapid scanning and thin-slice acquisition has revolutionized HRCT technique. Volumetric HRCT using thin detectors (0.5–0.625  mm) has become the routine in many institutions. In an early attempt at volumetric imaging (38), four contiguous HRCT scans were obtained without using helical technique at each of three locations (the aortic arch, carina, and 2 cm above the right hemidiaphragm) in 50 consecutive patients with interstitial lung disease or bronchiectasis. At each level, the diagnostic information ­obtainable from the set of four scans was compared to that obtainable from the first scan in the set of four. When the full set of four scans was considered, more findings of disease were identified. The sensitivity of the first scan as compared to the set of four was 84% for the detection of bronchiectasis, 97% for ground-glass opacity, 88% for honeycombing, 88% for septal thickening, and 86% for nodular opacities (38). However, it is more likely that the improvement in sensitivity found using the set of four scans reflects the number of scans viewed rather than the fact that they were obtained in contiguity. Although volumetric HRCT images appear slightly smoother than axial HRCT (Figs. 1-14 and 1-15), the technique has several advantages. It allows (a) complete imaging of the lungs and thorax, (b) viewing of contiguous slices for the purpose of better defining lung abnormalities, (c) reconstruction of scan data in any plane or using maximum-intensity projections (MIPs) or minimumintensity projections (MinIPs), (d) precise level-by-level

#152201  Cust: LWW  Au: Webb  Pg. No. 12

comparison of studies obtained at different times for evaluation of disease progression or improvement, and (e) the diagnosis of additional thoracic abnormalities (Figs. 1-16 to 1-24). On the other hand, the use of volumetric multidetector helical HRCT (MD-HRCT) results in a greater radiation dose than does spaced axial imaging. It is not unusual for only one or two slices from a volumetric HRCT study to provide the key observations necessary for diagnosis (39–44). For example, in a patient with suspected idiopathic pulmonary fibrosis, the presence of honeycombing, which is necessary for a definite diagnosis, may be visible only on a few images. This finding could be missed on spaced axial images. MDCT scanners make use of multiple adjacent detector rows that acquire scan data simultaneously and may be used independently or in combination to generate images of different thickness (45). Current MDCT scanners are capable of imaging the entire thorax within a few seconds, with the volumetric reconstruction of thin, high-resolution slices. For example, using a 64-detector scanner, data may be simultaneously acquired from sixtyfour 0.625-mm-thick detector arrays, a pitch of 1 to 1.5, and a gantry rotation time of 0.5 s or less. The volumetric data resulting from this mode of scanning allow isotropic imaging and HRCT assessment of lung morphology in a continuous fashion from lung apex to base, the viewing of the scan volume in nontransaxial planes or with three-dimensional (3D) reconstruction (Figs. 1-16A–C to 1-18), and the production of MIP and MinIP images at any desired level or in alternate planes (Figs. 1-16D–F and 1-19 to 1-24). This technique also allows a volumetric CT examination of the thorax to be easily combined with HRCT. Even with a rapid scanner, dyspneic patients with diffuse lung disease may not be able to hold their breath for the duration of a volumetric study. In such patients, if optimal resolution is desired, the scan protocol may be modified according to the distribution of the disease suspected. For diseases likely to have a basal predominance, such as idiopathic pulmonary fibrosis, scanning should begin near the diaphragm and proceed cephalad. In this manner, the more important basal lung will be imaged at the beginning of the scan sequence, and if the patient begins to breathe during scanning, only images through the less important upper lobes will be degraded by respiratory motion. For the same reason, in a patient suspected of having a disease with an upper-lobe predominance (e.g., sarcoidosis), it is appropriate to begin scanning in the lung apices. Because lung movement with respiration is greatest at the lung bases, an alternative approach would be to scan from the bases to apices in all patients. If the patient breathes during the scan, the upper lobes would be less affected. The helical acquisition of HRCT data results in some broadening of the scan profile as compared to detector width. Using a low value of pitch (e.g., 1) is recommended to minimize this effect (29). However, the effective slice thickness obtained using MD-HRCT is clearly sufficient for

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

13

A

B

C

D

FIGU RE 1-16  Reconstructed MD-HRCT in a patient with mixed connective tissue disease and NSIP. This is the same patient as shown in Fig. 1-15. A: 2D coronal reconstruction shows findings identical to those in Fig. 1-15. The basal distribution of the abnormalities is well shown, but the same information would be available from review of the transaxial images. B: 2D sagittal reconstruction clearly shows the posterior and lower-lobe predominance of the abnormalities, lower-lobe volume loss as evidenced by posterior displacement of the major fissure (white arrows), and subpleural sparing (black arrows). C: 3D surface-display reconstruction with a perspective from below the lung bases shows the distribution of basal-predominant lung disease, but otherwise is of little diagnostic value. D: Transaxial MIP image at the same level as shown in Fig. 1-15C. MIP imaging in this patient obscures detail and is of little diagnostic value. E and F: 3D MinIP coronal (E) and sagittal (F) reconstructions show the airways and lower-lobe traction bronchiectasis to best advantage. Ground-glass opacity is less apparent in the lung bases than on the routine images.

E

#152201  Cust: LWW  Au: Webb  Pg. No. 13

DESIGN SERVICES OF

14

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

F FIGU RE 1-16  (Continued)

HRCT diagnosis when thin detectors are used (Figs. 1-14 and 1-15). With 0.625-mm detector width and a pitch of 1, the effective slice thickness is 1 mm or less. Practically speaking, in most situations, using thin detectors and standard pitch is adequate for diagnosis. Depending on the technique used and how data from the various detector rows are combined, images of different thickness may be produced retrospectively from the same study. Using the protocol described previously, in addition to viewing images generated from data acquired by the individual detectors, data from the detector rows may be combined to produce images representing thicker slices (i.e., 2.5 or 5 mm). Thus, this technique enables HRCT and “routine” or thick-section chest imaging to be combined as a single examination, blurring the distinction between these studies. Combining a volumetric chest CT examination with HRCT by using MDCT may be of value in patients being studied primarily for diffuse lung disease, for which HRCT would be the examination of choice, and in patients being evaluated for a disease or abnormality usually studied using a thicker-slice helical CT. For example, in patients with hemoptysis, both thin and thick image reconstruction may be of value in demonstrating both

#152201  Cust: LWW  Au: Webb  Pg. No. 14

small or large airways disease and vascular abnormalities (46). Another advantage of MDCT would be in patients requiring CT for the diagnosis of thoracic disease such a lung carcinoma. In such patients, scan data may be reconstructed with a thickness appropriate for the detection of lung nodules and bronchial abnormalities and for assessment of mediastinal and hilar lymph nodes. At the same time, and without additional scanning, high-resolution images could be reconstructed for the purpose of delineating nodule morphology and attenuation, or for the ­diagnosis of associated lymphangitic spread of carcinoma. Similarly, in patients with suspected pulmonary vascular disease, HRCT with contrast enhancement may be obtained using MDCT, allowing the detailed assessment of both vasculature and pulmonary parenchyma (Figs. 1-18 and 1-19) (46,47). In patients having helical CT for the diagnosis of acute or chronic pulmonary embolism or pulmonary hypertension, scan data can be reconstructed using different algorithms to look for vascular abnormalities and lung disease that could be associated with similar symptoms. Although it is clear that MDCT scanners produce diagnostic helical HRCT examinations, the results of studies comparing MD-HRCT to spaced axial HRCT have been mixed, and neither imaging method appears to offer a clear advantage. For example, Sumikawa et al. (48) found that the quality of MD-HRCT images was equivalent to that of axial HRCT in 11 autopsy lungs; visualization of abnormal structures and diagnostic efficacy with MD-HRCT (0.75-mm collimation, pitch of 1) was equal to that of axial scans with 0.75-mm collimation. Also, S­ choepf et al. (49) compared MD-HRCT using a 1.25-mm detector and a pitch of 1.5 with spaced axial images (1-mm slices) in two groups of patients. No significant difference (p = 0.986) was found between multislice and single-slice axial HRCT sections in an overall score of image quality, spatial resolution, subjective signal-to-noise ratio, diagnostic value, depiction of bronchi and parenchyma, and motion and streak artifacts (49). In contrast, Honda et al. (50) compared the image quality and diagnostic e­fficacy of MD-HRCT (1.25-mm slices) to axial HRCT (1-mm slices) in imaging cadaveric lungs. The image quality of  axial HRCT was considered superior to that of MD-HRCT obtained with 1.25-mm detectors and a pitch of 0.75 or 1.5, and less image noise was present with axial HRCT. However, the diagnostic efficacy of MD-HRCT with a pitch of 0.75 was equal to that of axial HRCT (50). Kelly et al. (51) found that MD-HRCT may be associated with significantly greater motion artifact compared with axial HRCT obtained in the same patient. However, MD-HRCT scans were obtained using 4- or 8-detector scanners, and scan time was undoubtedly longer than that with current MDCT scanners. On the other hand, Studler et al. (52) found that motion artifacts were significantly more common on axial HRCT scans (1-mm collimation) than on MD-HRCT (1.5-mm detectors, pitch 1.25) images (p < 0.001). However, the authors believed that the assessment of ground-glass opacity was superior on axial

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

15

A

B FIGU RE 1-17  Reconstructed images in a patient with interstitial

pneumonia. This represents the same patient as shown in Fig. 1-2. A: Transaxial MD-HRCT with 1.25-mm slice thickness shows reticulation, traction bronchiectasis, and ground-glass opacity with a posterior and subpleural distribution. The upper lobes were less abnormal. B: Sagittal reconstruction (0.7 mm thick) shows these abnormalities to predominate in the posterior and basal subpleural lung (arrows). C: Coronal reconstruction (0.7 mm thick) through the posterior lung shows a similar distribution.

C

HRCT. The effective radiation doses were 3.8 millisievert (mSv) for MDCT and 0.9 mSv for axial HRCT. When considering the relative value of spaced axial HRCT and MD-HRCT, the greater radiation dose involved in MDHRCT should be kept in mind. Benaoud et al. (53) compared 1-mm-thick images reconstructed contiguously through the chest with images spaced at 1-cm intervals in the evaluation of chronic bronchopulmonary diseases. The spaced reconstructions would have resulted in a 79% radiation reduction, and there was almost perfect agreement (kappa = 0.83–1) for both the detection and distribution and findings between the volumetric and spaced reconstructions (53). In the screening of patients with scleroderma, Winklehner et al. (54) showed equal sensitivity for interstitial lung disease comparing volumetric MD-HRCT and spaced 1-mm sections reconstructed at 1-cm intervals. Certain HRCT findings, however, that have a heterogenous distribution may be better detected and quantified using volumetric HRCT. For instance, in the assessment of bronchiolitis obliterans syndrome (BOS) in lung transplant recipients, Dodd et al. (55) showed that volumetric MDCT correlated with the stage of BOS, whereas spaced HRCT images did not.

#152201  Cust: LWW  Au: Webb  Pg. No. 15

Despite these limitations, volumetric HRCT is becoming standard in many patients, for many indications, and in many institutions. At least partially, this reflects recent advances in radiation dose reduction with volumetric CT. Often, only the supine images will be obtained with volumetric imaging technique.

Reconstruction Techniques with Volumetric High-Resolution Computed Tomography Sagittal and Coronal Reformations MD-HRCT produces isotropic scans, allowing contiguous 3D ­visualization of the lung parenchyma and the capacity to create high-quality two-dimensional (2D) and 3D reformatted images (Figs. 1-16 to 1-20) (20). Honda et al. compared the quality of coronal multiplanar reconstructions obtained from an MDCT data set (0.5-mm collimation, 0.5-mm reconstruction interval) with the quality of direct coronal MD-HRCT (0.5-mm collimation) scans in 10 normal autopsy lung specimens. Image quality was considered equal (56). It is clear that MD-HRCT

DESIGN SERVICES OF

16

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

D

C FIGU RE 1-18  Contrast-enhanced MD-HRCT in an AIDS patient with pulmonary hypertension, obtained with 1.25-mm slice thickness. The differential diagnosis

included chronic pulmonary embolism, vasculitis, and lung disease. Transaxial (A and B) and sagittal reconstructed (C) HRCTs obtained during a single breath hold show normal findings. D: Transaxial image shows enlargement of main pulmonary artery consistent with pulmonary hypertension, but no evidence of pulmonary embolism. The presence of pulmonary hypertension in the absence of pulmonary embolism or lung disease suggests AIDS-related pulmonary hypertension with plexogenic arteriopathy.

reconstructions may provide additional information in selected cases (20), largely in regard to lung disease distribution, but routine transverse images are adequate for diagnosis in the large majority of cases. It has been suggested that the use of 2D coronal reconstructions may be useful for the primary interpretation of thoracic CT, but at present, it would seem most appropriate to use multiplanar reconstructions as a compliment to axial images. Kwan et al. (57) compared the accuracy and efficiency of primary interpretation of thoracic MDCT (5-mm slice thickness) using coronal reformations to that of routine transverse images. Each image set was assessed for 58 abnormalities of the lungs, mediastinum, pleura, chest wall, diaphragm, abdomen, and

#152201  Cust: LWW  Au: Webb  Pg. No. 16

skeleton. The mean detection sensitivity of all lesions was significantly (p = 0.001) lower on coronal (44% ± 26% [SD]) than on transverse (51% ± 22%) images, whereas the mean detection specificity was significantly (p = 0.005) higher (96% ± 5% vs. 95% ± 6%, respectively). Also, reporting findings for fewer coronal images took significantly (p = 0.025) longer (mean, 263 ± 56 s vs. 238 ± 45 s, respectively) (57). Arakawa et al. (58) evaluated the diagnostic utility of coronal MD-HRCT reformations (1.9-mm thickness) to axial HRCT (2-mm collimation) in diffuse and focal lung diseases. In 22.1% of cases, coronal MD-HRCT reformations were regarded as superior to axial HRCT or provided additional information, whereas in 72.4%, coronal MD-HRCT was regarded

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

17

A B

C

D

FIGU RE 1-19  Contrast-enhanced MD-HRCT obtained with 1.25-mm slice thickness in a 19-year-old woman with hypoxemia. Transaxial (A and B) images show numerous very small subpleural arteriovenous malformations (arrows). Onecentimeter-thick MIP images in the transaxial (C, D) and coronal (E) planes show the malformations (arrows) and their vascular supply to better advantage. She was subsequently found to have Osler-Weber-Rendu disease.

E

#152201  Cust: LWW  Au: Webb  Pg. No. 17

DESIGN SERVICES OF

18

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

FIGU RE 1-20  Coronal (A) and sagittal (B) MinIP reconstruction from 1.25-mm MD-HRCT in a patient with lymphangiomyomatosis. The MinIP images optimize visualization of the lung cysts and their distribution.

as comparable to axial HRCT, and in 5.5% it was considered inferior to axial images (58). Remy-­Jardin et al. (59) assessed the diagnostic accuracy of coronal reconstructions as an alternative to transverse MD-HRCT in the diagnosis of infiltrative lung disease. No significant

difference was found between the transverse and coronal images in the identification of CT features of disease or their distribution in the central, peripheral, anterior, and/ or posterior lung zones. However, in patients with extensive lung disease, the cephalocaudal distribution of lung

B

A

FIGU RE 1-21  MIP image in a patient with small lung nodules obtained using a multidetector-row helical CT scanner with 1.25-mm detector width and a pitch of 6. A: A single HRCT image shows two small nodules (arrows) that are difficult to distinguish from vessels. B: An MIP image consisting of eight contiguous HRCT images, including A, allows the two small nodules to be easily distinguished from surrounding vessels.

#152201  Cust: LWW  Au: Webb  Pg. No. 18

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

19

B

A

FIGU RE 1-22  MIP image in a patient with extensive abnormalities due to alveolar proteinosis obtained using MD-HRCT and 1.25-mm slice thickness. A: A single HRCT image shows a typical patchy distribution of interlobular septal thickening and ground-glass opacity (i.e., crazy paving) typical of alveolar proteinosis. B: A MIP image consisting of five contiguous HRCT images, including A, results in a confusing superimposition of opacities. Septal thickening is more difficult to diagnose.

abnormalities was more precisely assessed with coronal reconstructions (59). Nishino et al. (60) attempted to determine whether sagittal reformations of volumetric MD-HRCT provide additional information in evaluating lung abnormalities, when compared to axial HRCT images. Additional findings of diagnostic significance were identified on the sagittal reconstructions in 2 or 22 patients, principally related to the relationship of a nodule or mass to the fissures, pleura, or pericardium (60).

Maximum- and Minimum-Intensity Projections Several studies have used helical HRCT with thin collimation and MIPs or MinIPs to acquire and display volumetric HRCT data for a slab of lung (20,61–63). In a study by Bhalla et al. (61), when compared to conventional HRCT, volumetric MIP and MinIP images demonstrated additional findings in 13 of 20 (65%) cases. However, the authors found that the conventional HRCT scans

FIGU RE 1-23  MinIP image in the patient shown in Fig. 1-21A at the same anatomical level. Normal lung parenchyma appears relatively homogeneous. Pulmonary vessels disappear on MinIP images.

#152201  Cust: LWW  Au: Webb  Pg. No. 19

showed fine linear structures, such as the walls of airways and interlobular septa, more clearly than either MIP or MinIP images. MIP imaging has been used to best advantage for the diagnosis of nodular lung disease. MIP images increase the detection of small lung nodules and can be helpful in demonstrating their anatomical distribution (Fig. 1-21). Coakley et al. (62) assessed the use of MIP images in the detection of pulmonary nodules by helical CT. In this study, 40 pulmonary nodules of high density were created by placement of 2- and 4-mm beads into the peripheral airways of five dogs. MIP images were generated from overlapped slabs of seven consecutive 3-mm slices, reconstructed at 2-mm intervals, and acquired at pitch 2. MIP imaging increased the odds of nodule detection by more than two, when compared to helical images, and reader confidence for nodule detection was significantly higher with MIP images. In a study by Bhalla et al. (61), the use of helical HRCT and MIP images was compared in patients with nodular lung disease. Because of the markedly improved visualization of peripheral pulmonary vessels and improved spatial orientation, MIP images were considered superior to helical scans for identifying pulmonary nodules and specifying their location as peribronchovascular or centrilobular, a finding of great value in differential diagnosis. In another study (63), sliding-thin-slab MIP reconstructions were used in 81 patients with a variety of lung diseases associated with small nodules. In this study, patients were studied using 1- and 8-mm-thick conventional CT and helical CT with production of 3-, 5-, and 8-mmthick MIP reconstructions. When conventional CT findings were normal, MIPs did not demonstrate additional abnormalities. When conventional CT findings were ­inconclusive, MIP enabled the detection of micronodules (i.e., nodules 7 mm or less in diameter) involving less than 25% of the lung. When conventional CT scans showed micronodules, MIP showed the extent and distribution of micronodules and associated bronchiolar abnormalities

DESIGN SERVICES OF

20

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

B

A

FIGU RE 1-24  MD-HRCT image with contrast enhancement in a patient with bronchiolitis obliterans and a clinical suspicion of pulmonary embolism. No pulmonary embolism was found. A: A single HRCT image with 1.25-mm detector width shows bronchiectasis and patchy lung attenuation with reduced artery size in lucent lung regions due to air trapping and mosaic perfusion. B: A 10-mm-thick MinIP image at the same level as A accentuates the differences in attenuation between normal lung and lucent lung, but pulmonary arteries cannot be assessed. Bronchiectasis is well seen using MinIP imaging. C: MIP at the same level as B shows reduced vessel size in the lucent lung regions. Inhomogeneous lung attenuation is also visible. The bronchiectasis is difficult to see on the MIP image.

C

to better advantage. The sensitivity of MIP (3-mm-thick MIP, 94%; 5-mm-thick MIP, 100%; 8-mm-thick MIP, 92%) was significantly higher than that of conventional CT (8-mm-thick, 57%; 1-mm-thick, 73%) in the detection of micronodules (p < 0.001). The authors (63) concluded that sliding-thin-slab MIPs may help detect micronodular lung disease of limited extent and may be considered a valuable tool in the evaluation of diffuse infiltrative lung disease. Sakai et al. attempted to determine whether MIP images assisted in the diagnosis of the distribution of micronodules in a variety of focal and diffuse infiltrative lung diseases. Ten-millimeter-thick MIP image slabs at ­10-mm intervals were produced from MD-HRCT. Radiology residents interpreting the images benefited significantly from the use of MIPs, while board-certified radiologists had equal accuracy with and without the MIPs (64). Although MIP imaging may be valuable in the detection and diagnosis of lung nodules or nodular lung disease and in the demonstration of vascular abnormalities (Fig.  1-19), in patients with other abnormalities, MIP imaging may result in a confusing superimposition of opacities that tends to obscure anatomical detail. This is particularly true in patients with extensive ground-glass opacity or reticular opacities (Figs. 1-16D and 1-22). The utility of MinIP images (Figs. 1-20, 1-23, and 1-24) has also been evaluated. MinIP images are most useful in

#152201  Cust: LWW  Au: Webb  Pg. No. 20

the demonstration of abnormalities characterized by low attenuation (Figs. 1-20 and 1-24) (61). In one study (61), MinIP images were more accurate than routine HRCT scans in identifying (a) the lumina of central airways (Figs. 1-16E,F and 1-24B), (b) areas of abnormal low attenuation (e.g., emphysema or air trapping) (Figs.  1-20 and 1-24B), and (c) ground-glass opacity.

RADIATION DOSE The radiation dose associated with thoracic CT has received increased attention in recent years, as have attempts at CT dose reduction (16,19,45,65–70). At the same time, the development of volumetric MD-HRCT for diagnosing diffuse lung disease has resulted in an increased patient radiation dose, as compared to spaced axial HRCT. As pointed out by Aziz et al. (29), our enthusiasm for MDHRCT should be tempered by an understanding of the increased radiation dose involved. Before the use of spaced axial imaging is abandoned, there should be evidence that volumetric HRCT is superior (29). Effective dose is a widely used measure of radiation exposure from medical imaging (70). Effective dose is calculated by summing the absorbed doses to individual organs weighted for their radiation sensitivity; the unit of measurement is the sievert or millisievert. Determining the effective dose requires the measurement of absorbed

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

dose to each body organ multiplied by their radiation sensitivity, which is impractical in the clinical setting. However, a simpler calculation may be made in order to estimate the effective dose, based on several assumptions (70). Scanner manufacturers use dose data derived from measurements of radiation dose in phantoms to determine a weighted CT dose index (CTDI) for each CT scanner model, at all available selections of tube voltage (kV(p)), tube current (mA), and rotation time. The selected pitch value is then incorporated to produce a CT dose index called the CTDIVOL, measured in grays (Gy) or milligrays (mGy). The CTDIVOL allows a comparison of the amount of radiation associated with different scanners and scan parameters, but does not take into account the length of the scan or the radiation sensitivity of affected tissues and organs. The CTDIVOL is multiplied by the scan length in centimeters to calculate the dose length product (DLP). The DLP is a measure of the overall radiation dose delivered to the patient during the scan. An estimated effective dose for a specified CT scan can be calculated by multiplying the DLP by a normalized effective dose coefficient for the scanned body part (chest = 0.014 mSv/mGy/cm or 1.4%) (71). The effective dose coefficient accounts for the radiation sensitivity of the body region scanned, although specific organs are not considered, and several assumptions are made; tissue-weighting factors are averaged over sex and age and patient size, or in other words, this value assumes an average patient (70). Although it does not provide a precise measurement, it allows a general comparison of imaging studies. Yearly background radiation is approximately 2.5 to 3 mSv. HRCT performed with spaced axial images results in a low-radiation dose as compared to MD-HRCT obtained with volumetric image acquisition (Fig. 1-25, Table 1-2) (16,72). For example, Mayo et al. (72) compared the thoracic radiation dose associated with spaced axial HRCT to that of conventional CT with contiguous slices. In this study, using an early-generation scanner and scan technique of 120 kV(p), 200 mA, and 2-s scan time, the mean skin radiation dose was 4.4 mGy for 1.5-mm HRCT scans at 10-mm intervals, 2.1 mGy for scans at 20-mm intervals, and 36.3 mGy for conventional 10-mm scans at 10-mm intervals. Thus, HRCT scanning at 10- and 20-mm intervals, as done in clinical imaging, resulted in 12% and 6%, respectively, of the radiation dose associated with conventional CT. Schoepf et al. (49) compared MD-HRCT to HRCT obtained with spaced axial images considered to have equal image quality, spatial resolution, subjective signal-to-noise ratio, diagnostic value, depiction of bronchi and parenchyma, and motion and streak artifacts. Radiation dose measured 5.55 mSv for MDCT and 1.25 mSv for the series of 24 axial HRCT slices obtained (49). Leswick et al. found that if image noise is equalized, MD-HRCT may result in a higher radiation dose than the combination of a routine MDCT sufficient for volumetric imaging and spaced axial HRCT images (37).

#152201  Cust: LWW  Au: Webb  Pg. No. 21

21

A

B

C FIGU RE 1-25  A–C: Axial HRCT with a modulated mA of 100 to 150. Axial

HRCT images were obtained in the supine (A) and prone (B) positions at 1-cm intervals using a modulated tube current varying between 100 and 150 mA. Dynamic expiratory images (C) were also obtained at three selected levels using a tube current (mA) of 50. Estimated effective dose for the examination was 1.5 mSv.

TABLE 1-2  Comparison of Radiation Dose for Chest Imaging Techniques Procedure

Effective radiation dose (mSv)

Annual background radiation

2.5

PA chest radiograph

0.05

Spaced axial HRCT (10-mm spacing)

0.7

Spaced axial HRCT, supine, prone (10-mm spacing), expiratory

1.5

Spaced axial HRCT (20-mm spacing)

0.35

Low-dose spaced axial HRCT

0.02

MD-HRCT (standard technique)

4–7

MD-HRCT (modulated mA of approx. 100)

2–3

Modified from Mayo JR, Aldrich J, Muller NL. Radiation exposure at chest CT: a statement of the Fleischner Society. Radiology 2003;228:15–21.

DESIGN SERVICES OF

22

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

Attempts at dose reduction with MD-HRCT using a decrease in mAs may be achieved by choosing a reduced fixed mA value, body weight-based formulas, or scannerbased dynamic tube current modulation (19,70,73,74). Tube current modulation can provide excellent HRCT studies with reduced radiation dose (Fig. 1-26). However, as pointed out by Mayo et al. (16,75), dose reduction may have an adverse effect on image quality

A

B

C FIGU RE 1-26  A–C: Volumetric HRCT with a modulated mA of

approximately 100. Representative images through the upper (A), mid- (B), and lower (C) lungs are shown from a supine volumetric HRCT acquired with 120 kV(p), a fixed tube current of 100 mA, and reconstructed at 1.25-mm thickness. Dynamic expiratory images were also obtained at three selected levels. The estimated effective dose for the examination was 2 mSv. The dose would be increased by the inclusion of prone images.

#152201  Cust: LWW  Au: Webb  Pg. No. 22

and reader interpretations. For example, Yi et al. (76) assessed image noise and subjective image quality with respect to the radiation dose delivered by MDCT in 20 patients with suspected bronchiectasis. Images were ­obtained using 120 kV(p), 2.5-mm collimation, pitch of 1.5, 2.5-mm reconstruction intervals, and sharp reconstruction algorithm. The quality of the images obtained using six mA settings (170, 100, 70, 40, 20, and 10 mA) was assessed, and it was graded using a 5-point scale (5 = excellent to 1 = nondiagnostic) at both lung and mediastinal window settings. Also, radiation doses were measured at each of the six mA settings using a thoracic phantom. The mean image quality scores at exposures of 170, 100, 70, 40, 20, and 10 mA were 3.9, 3.7, 3.8, 3.2, 2.5, and 1.6 at lung window settings, and images obtained at 70 mA were rated significantly better than those obtained at 40  mA or less (p < 0.01). The average image noise (SD of pixels measured in blood) was 39, 42.7, 53.6, 69.2, 98.5, and 157.2 H, respectively, at 170, 100, 70, 40, 20, and 10 mA, and the mean radiation doses measured at these mA values were 23.72, 14.39, 10.54, 5.41, 2.74, and 1.50 mGy, respectively (76). The authors point out that the dose resulting from MDCT obtained with 70 mA (10.54 mGy) is five times that reported for spaced axial HRCT (2.17  mGy with parameters of 120 kV(p), 170 mA, 1-mm collimation, and 10-mm intervals) for the diagnosis of bronchiectasis (77). Das et al. (19) compared the image quality of thoracic MDCT obtained with a standard protocol (effective mAs = 100) to three methods of dose reduction, including a dynamic tube current modulation, effective mAs equals body weight in kilograms, and a combination of these. The mean effective doses for these protocols, respectively, were 6.83, 5.92, 4.73, and 3.97 mSv. Although there was a correlation between decreased dose and increased image noise, the image quality for all techniques was graded as excellent (19). Tube current modulation allows for dynamic changes of the tube current (mA) in the craniocaudal (Z plane) and transaxial (X and Y) planes. Tube output varies depending upon the attenuation profiles of specific anatomical locations. For example, tube current will be lowered in regions of the body that have less attenuation, such as levels at which the lungs comprise a large portion of the cross-sectional area of the chest. Tube current modulation attempts to maintain fixed image noise at all anatomical levels, reducing radiation exposure without sacrificing image quality. Using tube current modulation, Kalra et al. (78) showed a dose reduction of 18% to 26% compared to a fixed mA in patients undergoing routine chest CT. Angel et al. (79) demonstrated a 16% reduction in the absorbed dose to the lung with tube current modulation, an effect that was most pronounced in smaller patients. In larger patients, there was an increase in the dose of up to 33% (79). When tube current modulation is used, changing other parameters such as pitch and gantry rotation speed will have limited impact on radiation dose. For example, increasing the pitch will result in a subsequent

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

elevation in tube current so that noise image remains constant. The primary exception is when tube current is ­already at its maximum level, such as in large patients. With tube current modulation, and a state-of-the-art scanner having sensitive detectors, HRCT (supine volumetric, prone axial at 1-cm intervals, and dynamic expiratory imaging at three levels) may be performed using 300-ms rotation and an mA averaging about 100, with an estimated dose of about 2 to 3 mSv, and excellent image quality. Supine and prone axial imaging (1-cm spacing) and dynamic expiratory imaging at three levels can be performed with an estimated dose of about 1 mSv. Another dose reduction strategy is the use of adaptive statistical iterative reconstruction (ASIR). ASIR uses a postprocessing algorithm that represents an adjunct to standard filtered back projection (FBP). In usual clinical practice, a combination of FBP and ASIR are used to produce the final data set with typical blends, including 30% to 40% ASIR. Reconstructions using ASIR have reduced image noise compared to those using only FBP, allowing images to be acquired with parameters that lower the radiation dose (70). However, the use of ASIR can affect quantitative CT measurements (80). In one study (81), images were acquired at various tube current-time products (40–150 mAs) and then reconstructed using FBP and blended ASIR/FBP. At 40 and 75 mAs, the images reconstructed with FBP had unacceptable levels of noise, whereas the ASIR/FBP images had acceptable noise levels (81). In the evaluation of diffuse lung disease, Prakash et  al. (82) showed that images reconstructed with ASIR in a high-definition mode were superior in quality to FBP in 64% of cases. ASIR’s primary disadvantages are an increased postprocessing time (30% longer than FBP), edge definition artifacts, and the production of oversmoothed images. These disadvantages are limited by the blending of ASIR and FBP in the final reprocessing. Model-based iterative reconstruction (MBIR) is a more advanced form of iterative reconstruction that allows for further reduction in radiation dose at the expense of significantly increased postprocessing time, but is not in common use at this time. Radioprotective bismuth shields are another dose reduction technique that allows for a decrease in the specific target dose to radiosensitive organs such as the breasts and thyroid. Bismuth shields enable a reduction in the target organ dose at the expense of increased artifacts. With breast shields, in particular, this artifact is most pronounced in the anterior lungs (83). In a study by Colombo et al. (84), bismuth shielding allowed for a 34% reduction in the dose to the breast during chest CT with only a slight degradation of image quality. The contemporaneous use of tube current modulation and bismuth shields may result in an increase in the tube current or image noise depending upon when the shield is applied, before or after the scout image (85,86). Leswick et al. (86) showed that z-axis automatic tube current modulation was more effective than shielding in reducing the radiation exposure to the thyroid. A combination of shielding

#152201  Cust: LWW  Au: Webb  Pg. No. 23

23

and automatic tube current modulation reduced the thyroid dose slightly compared to tube current modulation alone; however, this was at the expense of increased artifact (86). The American Association of Physicists in Medicine recommends using alternative methods of dose reduction, in lieu of shields, because of their unpredictable effects on image quality and radiation exposure, particularly when automatic tube current modulation is used (87).

Low-Dose Axial High-Resolution Computed Tomography Spaced axial HRCT with reduced mAs can allow a ­diagnosis of diffuse lung disease with very limited radiation exposure. Obtaining spaced axial HRCT at 20-mm intervals (40 mAs) or at three levels (80 mAs) results in an average skin dose comparable to that associated with chest radiography (72,88–91). Low-dose HRCT should not be routinely used for the initial evaluation of patients with lung disease, although it can be valuable in following patients with a known lung abnormality or in screening large populations at risk for lung disease. Optimal low-dose techniques will likely vary with the clinical setting and indication for the study, and they remain to be established. The efficacy of low-dose spaced axial HRCT has been assessed in several studies (88,89,92,93). In a study by Zwirewich et al. (88), scans with 1.5-mm collimation and 2-s scan time at 120 kV(p) were obtained using both 20  mA (low-dose HRCT) and 200 mA (conventionaldose HRCT) at selected levels in the chests of 31 patients. Observers evaluated the visibility of normal structures, various parenchymal abnormalities, and artifacts using both techniques. Low- and conventional-dose HRCT were equivalent for the demonstration of vessels, lobular and segmental bronchi, and structures of the secondary pulmonary lobule, and in characterizing the extent and distribution of reticular abnormalities, honeycomb cysts, and thickened interlobular septa. However, the low-dose technique failed to demonstrate ground-glass opacity in 2 of 10 cases, and emphysema in 1 of 9 cases, although they were evident but subtle on the usual-dose HRCT. Linear streak artifacts were also more prominent on images acquired with the low-dose technique, but the two techniques were judged equally diagnostic in 97% of cases. The authors concluded that HRCT images acquired at 20 mA yield anatomical information equivalent to that obtained with 200-mA scans in the majority of patients without significant loss of spatial resolution or image degradation due to streak artifacts. In a subsequent study (89), the diagnostic accuracies of chest radiographs, low-dose HRCT (80 mAs, 120 kV(p)), and conventional-dose HRCT (340 mAs, 120 kV(p)) were compared in 50 patients with chronic infiltrative lung disease and 10 normal controls. For each HRCT technique, only three images were used, obtained at the levels of the aortic arch, tracheal carina, and 1 cm above

DESIGN SERVICES OF

24

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

the right hemidiaphragm. A correct first-choice diagnosis was made significantly more often with either HRCT technique than with radiography; the correct diagnosis was made in 65% of cases using radiographs, 74% of cases with low-dose HRCT (p < 0.02), and 80% of conventional HRCT (p < 0.005). A high confidence level in making a diagnosis was reached in 42% of radiographic examinations, 61% of the low-dose HRCT examinations (p < 0.01), and 63% of the conventional-dose HRCT examinations (p < 0.005), and it was correct in 92%, 90%, and 96% of the studies, respectively. Although conventional-dose HRCT was more accurate than lowdose HRCT, this difference was not significant, and both techniques provided quite similar anatomical information (Figs. 1-25 and 1-26) (89). In a comparison of standard (150 mAs) and low (40 mAs)-dose thin-section volumetric chest CT, Christie et al. (94) found that there was significantly increased detection of ground-glass opacities, ground-glass nodules, and interstitial opacities with the higher-dose scan. The detection of solid nodules, airspace disease, and airways disease was equivalent using low- and high-dose images. Majurin et al. (92) compared a variety of low-dose techniques in 45 patients with suspected asbestos-related lung disease. Of the 37 patients with CT evidence of lung fibrosis, HRCT images obtained with mAs as low as 120 clearly showed parenchymal bands, curvilinear opacities, and honeycombing. However, reliable identification of interstitial lines or areas of ground-glass opacity required a minimum technique of 160 mAs. Furthermore, these authors showed that using the lowest possible dosage (60 mAs) HRCT was sufficient only for detecting marked pleural thickening and areas of gross lung fibrosis. An additional factor in obtaining low-dose HRCT is a consideration of the anatomical distribution of suspected disease. Significant dose reduction can be achieved by limiting scanning to the most appropriate lung regions and the most appropriate patient positions for obtaining the scans. As an example, in screening for asbestosis, scanning in the prone position and the posterior lung bases is most helpful in diagnosis (Fig. 1-27).

EXPIRATORY HIGH-RESOLUTION COMPUTED TOMOGRAPHY As an adjunct to routine inspiratory images, expiratory HRCT scans have proved useful in the evaluation of patients with a variety of obstructive lung diseases (95,96). On expiratory scans, focal or diffuse air trapping may be diagnosed in patients with large or small airway obstruction or emphysema. It has been shown that the presence of air trapping on expiratory scans (a) correlates to some degree with pulmonary function test abnormalities (97,98), (b) can confirm the presence of obstructive airway disease in patients with subtle or nonspecific abnormalities visible on inspiratory scans, (c) allows the diagnosis of significant lung disease in some patients with normal inspiratory scans (99), and (d) can help distinguish between

#152201  Cust: LWW  Au: Webb  Pg. No. 24

FIGU RE 1-27  Low-dose HRCT for asbestos screening. HRCT images were obtained at 1-cm intervals in the prone position using a fixed tube current of 40 mA. No supine or expiratory images were obtained. Estimated effective dose for the examination was 0.2 mSv.

obstructive disease and infiltrative disease as a cause of inhomogeneous lung opacity seen on inspiratory scans (100). In most lung regions of normal subjects, lung parenchyma increases uniformly in attenuation during expiration (8,101–105), but in the presence of air trapping, lung parenchyma remains lucent on expiration and shows little change in volume. Focal, multifocal, or diffuse air trapping is visible as areas of abnormally low attenuation on expiratory or postexpiratory CT. On expiratory scans, visible differences in attenuation between normal and obstructed lung regions are visible using standard lung window settings and can be quantitated using regions of interest. Differences in attenuation between normal lung regions and regions that show air trapping often measure more than 100 Hounsfield units (HU) (106). Air trapping visible using expiratory or postexpiratory HRCT techniques has been recognized in patients with emphysema (107–110), chronic airways disease (98), asthma (111– 115), cystic fibrosis (116), bronchiolitis obliterans and BOS (99,108,117–127), the cystic lung diseases associated with Langerhans histiocytosis and tuberous sclerosis (128), bronchiectasis (108,129), airways disease related to AIDS (130), and small airways disease associated with thalassemia (131). Expiratory HRCT has also proved valuable in demonstrating the presence of bronchiolitis in patients with primarily infiltrative diseases such as hypersensitivity pneumonitis (132,133), sarcoidosis (134–137), and pneumonia. Some investigators obtain expiratory scans routinely in all patients who have HRCT, whereas others limit their use to patients with inspiratory scan abnormalities or suspected obstructive lung disease (95). We recommend the routine use of expiratory scans in a patient’s initial HRCT evaluation because the functional cause of respiratory disability is not always known before HRCT is performed. Furthermore, even in patients with a known restrictive abnormality on pulmonary function tests, or obvious HRCT findings of fibrosis, expiratory HRCT

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

25

may show air trapping, a finding of potential value in differential diagnosis (136). For example, the presence of air trapping in a patient with HRCT findings of fibrosis and honeycombing excludes the diagnosis of usual interstitial pneumonia and idiopathic pulmonary fibrosis (138). Limiting expiratory HRCT to patients with evidence of airway abnormalities on inspiratory scans will result in some missed diagnoses. Expiratory HRCT may also show findings of air trapping in the absence of inspiratory scan abnormalities (99). The use of expiratory scans may be of value in the follow-up of patients at risk of developing an obstructive abnormality. For example, expiratory scans are valuable in detecting bronchiolitis obliterans in patients being followed for lung transplantation (123,125,139–142). Expiratory HRCT scans may be obtained during suspended respiration after forced exhalation (postexpiratory CT), during forced exhalation (dynamic expiratory CT) (95,104,108,143), at a user-selected respiratory level controlled during exhalation with a spirometer (spirometrically triggered expiratory CT) or by using other methods (126,144–149). Generally, with these techniques, expiratory scans are obtained at selected levels. Three scans, five scans, or scans at 4-cm intervals have been used by different authors. Expiratory imaging may also be performed using helical technique and 3D volumetric reconstruction (150,151).

single scan. This maneuver is practiced with the patient before the scans are obtained to ensure an adequate level of expiration. Postexpiratory scans can be performed at several predetermined levels (e.g., aortic arch, carina, lung bases), at 2- to 4-cm intervals, or at levels appearing abnormal on the inspiratory images. Scans at two to five levels have been used by different authors (100,111,112,120,140,152). Expiratory scans at three selected levels (aortic arch, hila, and lower lobes) are generally sufficient for showing significant air trapping and may be used routinely, in addition to the inspiratory scan series, in patients with suspected airways or obstructive lung diseases. Although targeting postexpiratory scans to lung regions that appear abnormal on the inspiratory scans would seem advantageous, using preselected scans allows the same lung regions to be routinely imaged on follow-up examinations and, in some patients, can show air trapping when inspiratory scans are normal. Each postexpiratory scan is compared to the inspiratory scan that most closely duplicates its level to detect air trapping. Anatomical landmarks such as pulmonary vessels, bronchi, and fissures are most useful for localizing corresponding levels. Because of diaphragmatic motion occurring with expiration, attempting to localize the same scan levels by using the scout view is difficult and sometimes misleading.

Postexpiratory High-Resolution Computed Tomography

Dynamic Expiratory High-Resolution Computed Tomography

Postexpiratory HRCT scans, obtained during suspended respiration after a forced exhalation, are easily performed with any scanner and are most suitable for a routine examination (Fig. 1-28). The primary advantage of this technique is its simplicity. In obtaining expiratory HRCT, the patient is instructed to forcefully exhale and then hold his or her breath for the duration of the

Scans obtained dynamically during forced expiration can be obtained using an electron-beam scanner (Fig. 1-29) or a helical scanner (Figs. 1-30 to 1-33). There is some evidence to suggest that a greater increase in lung attenuation occurs with dynamic expiratory imaging than with simple postexpiratory HRCT and that, consequently, air trapping is more easily diagnosed (Fig. 1-33).

A

B

FIGU RE 1-28  Postexpiratory air trapping in a patient with idiopathic scoliosis and normal inspiratory scans. A: An inspiratory scan shows homogeneous lung attenuation without evidence of airways disease. B: Routine postexpiratory scan shows patchy air trapping (arrows) indicative of small airways disease.

#152201  Cust: LWW  Au: Webb  Pg. No. 25

DESIGN SERVICES OF

26

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

B

A

FIGU RE 1-29  Dynamic expiratory HRCT in a normal subject obtained using an electron-beam scanner. A: The 10-image dynamic ultrafast HRCT sequence acquired during a single forced vital capacity maneuver is shown, with the FOV limited to the left upper lobe. These ten 100-ms images were obtained at 600-ms intervals. They are shown in sequence, in a clockwise fashion, from the left upper corner (1) to the lower left corner (10). Images at full inspiration (insp) and full expiration (exp) are visible. Note the increase in lung attenuation and decrease in lung volume that occur as the subject exhales. As in most normal subjects, lung attenuation increase on expiration is relatively homogeneous. B: A time-attenuation curve is produced by measuring the mean lung attenuation (HU) for a specific ROI. In this subject, for an ROI in the anterior lung, attenuation decreases to approximately –870 HU at maximum inspiration (insp) and increases in attenuation to –670 HU at maximum expiration (exp) for an overall attenuation increase of approximately 200 HU. Each point on the time-attenuation curve represents one image from the dynamic sequence. (From Webb WR, Stern EJ, Kanth N, et al. Dynamic pulmonary CT: findings in normal adult men. Radiology 1993;186:117, with permission.)

Dynamic scanning with an electron-beam scanner has been termed dynamic ultrafast HRCT (108,128,153,154). This technique is performed using a scanner capable of obtaining a series of images with a 100-ms scan time (­ 500-ms interscan delay, 1.5- to 3-mm collimation, 150 kV(p), 650 mA) (108,128,153,155). In general, when using this technique, a series of 10 scans are obtained at a single level during a 6-s period, as the patient first inspires and then forcefully exhales. Patients are instructed to breathe in deeply and then breathe out as rapidly as possible (Fig. 1-29). Images are reconstructed using a high spatial frequency algorithm. Usually, dynamic expiratory CT scan sequences are obtained at several selected levels through the lungs. In papers describing this technique, three levels were used (e.g., at the level of the aortic arch, carina, and lung bases), although the protocol can be varied in individual cases, with imaging limited to a specific region. During expiration, the diaphragm ascends, and the lungs move cephalad. Lung motion is most significant on scans through the lung bases. Although slightly different regions of the lung are imaged on sequential scans obtained at the same level, the effect of diaphragmatic motion on the assessment of lung attenuation has been regarded as inconsequential (104,108,153). Little

#152201  Cust: LWW  Au: Webb  Pg. No. 26

motion-related image degradation is visible on dynamic ultrafast HRCT scans because of the very rapid scan time used (128,155). Dynamic scans can also be obtained using a helical CT scanner with a gantry rotation time of 1 s or less. Because of the continuous scanning that is possible with the helical technique, scans can be reconstructed at any point during the scan sequence, thus providing a temporal resolution equivalent or superior to that of dynamic ultrafast HRCT. However, because of the longer time required to obtain each image, some degradation of anatomical detail can be expected on individual images. In performing dynamic expiratory CT, although one or more images obtained during the rapid phases of expiration will show significant motion-related artifact (Fig. 1-30), images near to and at full expiration show little artifact and ­allow optimal assessment of lung attenuation (Figs. 1-30 to 1-33) (156). The use of a dynamic helical technique may be combined with a reduced mAs (e.g., 40 mAs) so that the sequence of images obtained represents a radiation dose similar to that associated with a single routine expiratory image (Figs. 1-32 and 1-33). Using such a technique, continuous imaging is performed for 6 to 8 s, as the patient

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

A

27

B

C FIGU RE 1-30  Dynamic expiratory HRCT obtained using a helical scanner in a patient with bilateral lung transplantation. A: The initial scan from the dynamic HRCT sequence is at full inspiration and appears normal. No respiratory motion is present. B: An image from the midportion of the dynamic expiration shows significant motion artifacts, with degradation of image quality. Note that this image appears to be at a more caudal level than in A because the diaphragm and lungs have moved cephalad relative to table position (C). The final image from the dynamic sequence is at expiration, with no respiratory motion visible. Image quality is good, and marked patchy air trapping is visible as a result of bronchiolitis obliterans.

A

B

FIGU RE 1-31  Dynamic expiratory HRCT obtained using a helical scanner. A: In a patient with bilateral lung transplantation, inspiratory HRCT shows stenosis of the bronchus intermedius. The lungs appear normal. B: An image from the expiratory phase shows marked air trapping in right middle and lower lobes, with little change in attenuation. Note that the right major fissure (white arrow) is bowed forward in comparison to the left major fissure and its position on the inspiratory scan. This also reflects local air trapping. The right upper lobe (black arrow) and left lung increase normally in attenuation.

#152201  Cust: LWW  Au: Webb  Pg. No. 27

DESIGN SERVICES OF

28

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

C

D

E

F

FIGU RE 1-32  Normal low-dose dynamic expiratory helical HRCT. A: Axial HRCT was obtained using 120 kV(p) and 200 mA. B–F: Sequential scans from a dynamic expiratory sequence obtained with 120 kV(p) and 40 mA. Scans show artifacts from respiratory motion and increased image noise as compared to the axial image shown in A. However, scans are of sufficient quality for diagnosis and show increased lung attenuation and decreased lung volume during the sequence. Lung attenuation on scans (B–F), measured using a 2-cm ROI in the left upper lobe, was –832 HU (B), –789 (C), –770 HU (D), –736 HU (E), and –700 HU (F).

rapidly exhales. Although image quality is reduced using the low-dose technique because of noise, images adequate for the diagnosis of air trapping are obtained (Figs. 1-32 and 1-33). In a group of lung transplant recipients studied using both postexpiratory HRCT and low-dose dynamic expiratory helical HRCT (156), lung attenuation was noted to increase significantly more with the dynamic technique (204 HU vs. 130 HU, p = 0.0007), and in one patient, air trapping was diagnosed only on the dynamic images. Lucidarme et al. (157) also compared the utility of dynamic expiratory scans (obtained during a 10-s ­expiratory maneuver) to scans obtained at end expiration in 49 ­patients with airways disease. Air trapping was

#152201  Cust: LWW  Au: Webb  Pg. No. 28

noted in 36 patients using dynamic expiratory CT and 35 patients at suspended end expiration. The extent of air trapping and the relative contrast between normal lung and regions of air trapping were significantly greater when scans were obtained with dynamic expiration (p = 0.001 and 0.007, respectively) (157). Regardless of the technique used, the dynamic scan sequence is viewed with attention to changes in lung ­attenuation and regional lung volume during the forced expiration. The images can be evaluated quantitatively or qualitatively, with measurement of lung attenuation during different phases of the respiratory maneuver, calculation of time-attenuation curves, or simple viewing of the serial scans in sequence or in cine mode. Air

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

A

29

B

FIGU RE 1-33  Postexpiratory and low-dose dynamic expiratory HRCT in a patient with bronchiolitis obliterans resulting from smoke inhalation. A: Inspiratory axial HRCT shows subtle lung inhomogeneity due to mosaic perfusion. B: Postexpiratory axial HRCT (240 mA) shows findings of air trapping with some lung regions remaining lucent. C: Low-dose (40 mA) dynamic expiratory HRCT shows increased streak artifacts due to aliasing. Several regions of air trapping are more easily seen on the dynamic scan.

C

trapping is considered to be present when the lung fails to increase normally in attenuation during exhalation (108,128,153). The image sequence can be analyzed both quantitatively and qualitatively (Figs. 1-29 and 1-32). The mean HU attenuation for a specific region of interest (ROI) in the lung can be measured and plotted for each scan, producing a time-attenuation curve graphically demonstrating the changes in lung attenuation that have occurred during a single expiration and inspiration (128). The use of dynamic expiratory HRCT is discussed further in ­Chapters 2 and 7.

Spirometrically Triggered Expiratory High-Resolution Computed Tomography Spirometrically triggered expiratory HRCT is a technique by which expiratory scanning can be done at specific, reproducible, user-selected lung volumes (126,144–146,149). With this technique, the patient breathes through a small handheld spirometer while positioned on the CT table. Before scanning, a spirometric measurement of the vital capacity is obtained, and trigger level (e.g., 90% of vital capacity) is chosen. During exhalation, the spirometer and associated microcomputer measure the volume of gas expired and trigger CT after a specific volume is reached. When the trigger signal is generated, airflow is inhibited by closure of a valve attached to the spirometer, and a scanning starts. Two or three different levels in the chest are typically selected and evaluated with respect to lung attenuation at specific lung volumes. Using this method, quantitative assessment of CT images with respect to lung attenuation

#152201  Cust: LWW  Au: Webb  Pg. No. 29

can be performed with excellent precision (144,145). This technique may also be used with a helical scanner or an electron-beam scanner (147). Spirometrically gated or controlled imaging may be particularly valuable in pediatric patients (147); with inhibition of respiration, respiratory motion artifacts may be avoided. Motion-free inspiratory and expiratory imaging can also be obtained in pediatric patients by using a positive-­ pressure ventilation device and controlled pauses in spontaneous respiration (148). As an example of the use of this technique, Camiciottoli et al. (158) studied the relationship between spirometrically gated inspiratory and expiratory HRCT and respiratory dysfunction in patients with chronic obstructive pulmonary disease (COPD). The authors found that both inspiratory and expiratory measurements were important in patient assessment. Measurements of lung attenuation at inspiration reflected the extent of emphysematous tissue loss, while expiratory measurements were related to airflow limitation and lung hyperinflation (158).

Volumetric Expiratory Computed Tomography Kauczor et al. (150) first used helical CT (slice thickness, 8 mm; pitch, 2; increment, 8 mm) with 2D and 3D postprocessing to assess lung volume at deep inspiration and expiration. Both 2D and 3D reconstructions were found to correlate with measured lung volumes. In another study, 3D volumetric reconstructions of total lung volume at inspiration and at expiration, as well as quantitation of regions of low attenuation (lung attenuation measuring

DESIGN SERVICES OF

30

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

less than –896 HU on inspiratory CT and –790 HU on expiratory CT), were correlated with pulmonary function test results (151); in this study, an excellent correlation was found between the volume of low-attenuation lung and pulmonary function test findings of obstruction, such as the ratio of forced expiratory volume in 1 s (FEV1) to the forced vital capacity. The use of volumetric MD-HRCT following expiration has also been used for the assessment of volumetric lung-attenuation changes, identification of air trapping and associated airway abnormalities, and 2D reconstruction of expiratory scans (60,159–162). This technique is combined with volumetric MD-HRCT. The use of this technique has been assessed in several studies (60,159–162). Nishino et al. reviewed 41 patients with suspected diffuse airway abnormalities. The volumetric expiratory HRCT was diagnostically acceptable in 83% to 93% of patients, depending on the level scanned. There was no significant difference in the detectability of air trapping with volumetric imaging when compared to six evenly spaced end-expiratory HRCT images, but the airway leading to a region of air trapping was better identified using volumetric imaging than with spaced scans (p < 0.0001) (161). In another study, the use of coronal reformations of expiratory MD-HRCT was compared to transverse MD-HRCT. Although air trapping was visible on both transverse and coronal images, and there was no difference in diagnostic confidence and the size, distribution, and extent of areas of air trapping identified, the borders of areas of air trapping were better shown in some patients on the coronal images (162). Sagittal reconstruction of expiratory MD-HRCT has also been reported (60), but does not appear to offer any significant advantage in the diagnosis of air trapping. MinIP image reconstruction may be used with volumetric postexpiratory HRCT to improve the visibility of air trapping (163). It should be pointed out that the use of volumetric expiratory imaging results in a greater radiation dose than spaced axial images. In one study, the total effective radiation dose for MD-HRCT obtained in both inspiration and expiration was estimated as 11.61 mSv, even when a reduced mAs was used (161,162). More recently, Bankier et al. (164) reported that volumetric expiratory HRCT may be performed with reduced radiation dose, without impairing the visual quantification of air trapping. In their study, volumetric expiratory HRCT was performed using MD-HRCT with 140 kV(p) and 80 mAs (effective) and simulated reduced effective mAs values of 60, 40, and 20. They found that lower mAs values did not result in a significant change in air trapping scores, although diagnostic confidence and interobserver agreement both decreased. The mean effective dose at 140 kV(p) and 80 mAs (effective) was estimated as 4.7 mSv in women and 3.8 mSv in men; at simulated mAs (effective) of 20, the estimated dose was 1.2 mSv in women and 1.0 mSv in men (164).

#152201  Cust: LWW  Au: Webb  Pg. No. 30

QUANTITATIVE COMPUTED TOMOGRAPHY Traditional analysis of HRCT relies primarily on the subjective recognition and interpretation of findings and patterns. Quantitative computed tomography (QCT) represents an alternative, by which computerized analysis allows for a more objective assessment of disease abnormalities or extent. QCT may be used to detect, characterize, and quantify the severity of HRCT abnormalities. It may also be used in the longitudinal follow-up of abnormalities over time. HRCT images may be quantified in a variety of ways, from the application of density masks to the use of more advanced computer algorithms that attempt to determine the presence and severity of CT findings that are spatially complex. These methods are usually applied to a volumetric HRCT data set. The main indications for quantitative lung imaging are in the evaluation of COPD, airways diseases, and interstitial lung diseases. A brief discussion of these indications will be presented here with a more detailed explanation in the pertinent chapters. The assessment of COPD has been the most rigorously studied use of QCT (see Chapters 7 and 20) (165–184). COPD is a multifactorial disease that is the result of several pathologic abnormalities including emphysema, large airways disease, and small airways disease. QCT has been used in COPD patients to determine COPD phenotypes (i.e., emphysema-predominant, airway-predominant, or mixed), the severity and distribution of disease, and longitudinal changes in abnormalities over time. Emphysema may result in areas of lung with decreased attenuation on CT and may be quantified by calculating the relative lowdensity area or the percentile of the frequency-­attenuation distribution. Both techniques utilize a density mask that measures the attenuation of specific voxels within an ROI. Low attenuation is measured using the percentage of voxels below a specific threshold, typically –950 HU. The frequency-attenuation distribution determines the attenuation value, below which a specific percentage of the low-attenuation voxel densities are distributed, typically the lowest 15%. These measurements have been shown to correlate with clinical symptoms (185), pulmonary function test abnormalities (186), and histologic scoring of emphysema (187). Given substantial interobserver variability in visual assessment of emphysema (188), quantitative CT may provide a more objective and reliable assessment of the severity of COPD. Bankier et al. (189) showed better agreement between objective quantification of emphysema using density masks (r = 0.555–0.623) than subjective grading (r = 0.439–0.505), compared to histology. On the other hand, Kim et al. (190) demonstrated that semiquantitative visual assessment performed just as well as computerized quantitative methods. As functional impairment in COPD is multifactorial and includes pathologic abnormalities other than emphysema, subjective visual assessment may provide details that density masks do not. While QCT in the setting of COPD has several potential advantages over visual

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

assessment of CT images, its exact clinical role in COPD patients is not well defined. Quantitative CT of the airways may be used independently or as a complement to emphysema quantification. Direct measurements of the large airways, including bronchial diameter and wall thickness, may be performed. Measurements of the small airways are performed indirectly by quantifying air trapping on expiratory CT. The presence of air trapping is performed by measuring the percentage of pixels below a threshold of –850 or –856 HU. Comparison of inspiratory and expiratory images in COPD patients allows the determination of the predominant pattern of abnormality: emphysema, airway obstruction, or mixed. These measurements have been shown to correlate with pulmonary function test abnormalities, and may even show better correlation (r = –0.077) with FEV1 compared to the quantification of inspiratory images (r = –0.67) (186). Quantification of both emphysema and airway abnormalities may be superior than either used in isolation (191). Quantification of airways abnormalities may also be used in patients with other diseases such as asthma (192) and bronchiolitis obliterans (193). Quantification of interstitial lung disease requires more complex, texture-based computer algorithms that are able to differentiate findings on the basis of their morphology. Because there is variability in the experience and ­accuracy of individual radiologists in the assessment of diffuse lung disease, QCT has the potential to provide objective measurements, independent of radiologists’ experience. Most studies to date have focused upon the feasibility of ­applying these algorithms to patients with interstitial lung disease. Early studies have shown acceptable agreement between quantitative measurements and visual assessment (194,195) or pulmonary function tests (196). Inaccuracies in CT quantification may arise from several sources. Differences in the level of inspiration or ­expiration may have a significant impact on lung density

A

and measurements of emphysema (177). Variations in the CT protocol such as tube current, slice thickness, reconstruction algorithm, and the use of ASIR may also impact quantification (73,80,197). Additionally, QCT may not reflect the complex and multifactorial interplay of multiple factors that contribute to CT abnormalities. For instance, COPD patients who quit smoking show a rapid increase in the low-attenuation area on longitudinal QCT (198). This is hypothesized to be due to the decrease in inflammatory factors associated with active smoke inhalation as opposed to a rapid increase in emphysema. This example underlies the complexity of diseases that may not be captured by quantification.

ADDITIONAL TECHNICAL MODIFICATIONS Reduction of Cardiac Motion Artifacts HRCT scans obtained in a routine fashion may be degraded by cardiac motion. Several motion-related artifacts may be seen, particularly in the left paracardiac region (see High-Resolution Computed Tomography Artifacts section). HRCT using electrocardiographic (ECG) triggering of scan acquisition, reduced gantry rotation time, and segmented reconstruction of scan data have all been used in an attempt to reduce these artifacts (199–203).

Electrocardiographically Triggered High-­Resolution Computed Tomography Electrocardiographically triggered HRCT may be used to reduce motion-related artifacts (Fig. 1-34), but has little effect on diagnosis and results in an increased radiation exposure. In a study using a helical scanner capable of 0.75-s gantry rotation, 500-ms HRCT scans, representing a 240-degree rotation of the gantry, were initiated at 50% of the R-R interval (199). Because of the shorter-than-routine scan time, images were reconstructed using a smoother algorithm than is usually used

B

FIGU RE 1-34  ECG-gated MD-HRCT obtained with 0.625-mm slice thickness, a pitch of 1, and targeted reconstruction. Scans at the level of the upper (A) and lower (B) heart show excellent spatial resolution without artifacts related to cardiac pulsation. Small lobular vessels in the lung periphery are clearly seen.

#152201  Cust: LWW  Au: Webb  Pg. No. 31

31

DESIGN SERVICES OF

32

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

for HRCT. In studying 35 patients using this technique, Schoepf et al. (199) found that ECG triggering significantly reduced artifacts caused by cardiac motion, such as distortion of pulmonary vessels, double images, or blurring of the cardiac border, when compared to routine images. Furthermore, in patients with a heart rate of 75 beats per minute or less, ECG triggering significantly improved image quality. It should be noted, however, that this technique was not found to improve diagnostic accuracy. Also, Boehm et al. (201) studied 45 patients referred for HRCT with routine MD-HRCT and prospectively ECG-triggered HRCT. ECG triggering resulted in a significant reduction in motion artifacts in the middle lobe, lingula, and left lower lobe, but no differences in diagnostic outcome were found between triggered and nontriggered techniques. The authors conclude that ECG-triggered thin-section CT of the lung is not recommended for routine clinical practice (201).

Segmented Reconstruction Partial or segmented reconstruction of scan data can serve to reduce effective scan time and can result in a significant reduction in motion artifacts without increasing radiation dose, albeit at the expense of increased image noise. Arac et al. (202) studied HRCT images obtained using a scanner capable of 1-s rotation and reconstruction using a full gantry ­rotation and a 225-degree rotation segment. Segmented reconstruction reduced cardiac motion artifacts (202). Ha et al. (203) evaluated the effects of partial (0.3-s) reconstruction to reconstruction obtained using a full rotation (0.75 s). The use of partial reconstruction resulted in reduced cardiac motion artifacts on HRCT, but image noise was increased. For example, image noise in air (38.0 ± 9.2) and lung parenchyma (86.0 ± 23.1) were greater for 0.3-s images than for 0.75-s images (35.6 ± 9.6 and 76.0 ± 20.3, respectively; p < 0.01) (203).

A

Gantry Angulation When HRCT is obtained using spaced axial images, angling the top of the CT gantry 20 degrees caudally with the patient supine (i.e., the gantry is angled toward the feet) improves visibility of the segmental and subsegmental bronchi, particularly in the middle lobe and lingula, by aligning them parallel to the plane of scan (Fig. 1-35) (204). This technique may be valuable in assessing patients with bronchiectasis (205). However, in the majority of patients with bronchiectasis, spaced HRCT images without gantry angulation are sufficient for diagnosis, and there would seem to be little use for this technique when volumetric HRCT is obtained. With MD-HRCT, images could be reconstructed in any desired plane to demonstrate bronchi to best advantage.

Use of Contrast Agents At present, there is no routine indication for the use of contrast agents with HRCT, except when studying a focal lung lesion or solitary nodule (206) or in patients being studied for pulmonary vascular disease (Figs. 1-18 and 1-19). Because the lung window settings routinely used for HRCT are intended to accentuate the contrast between air and tissue, vascular opacification is not visible in patients receiving an injection of intravenous contrast. Using a soft-tissue window, however, opacification of most segmental and subsegmental vessels may be seen on both spaced and volumetric HRCT (207).

Use of Dual-Energy kV(p) Dual-energy CT utilizes two different kV(p) settings, typically 80 to 100 and 140, generated by two separate x-ray tubes or rapid switching of a single x-ray tube. Attenuation differences at the different energies allow the determination of the composition of various tissue components

B

FIGU RE 1-35  Gantry angulation in a patient with right middle-lobe bronchiectasis. A: HRCT image obtained with the gantry vertical shows bronchial wall thickening in the right middle lobe. B: HRCT image obtained with the gantry angulated 20 degrees allows right middle-lobe bronchi (arrows) to be imaged along their axes.

#152201  Cust: LWW  Au: Webb  Pg. No. 32

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

33

(air, iodine, soft tissue) within a specific voxel (208). Most studies to date have focused on the ability of dualenergy CT to create iodine maps reflecting lung perfusion (209,210). Another potential clinical use is the ability to map and quantify lung ventilation using inhaled xenon as a contrast agent (210,211). This has the potential to provide anatomical and functional information that is complementary to the visual assessment of CT images.

IMAGE DISPLAY Use of Workstations The use of an electronic workstation to view HRCT images is optimal and recommended. Scans may be ­ viewed at a larger size than generally possible on film, making small or subtle abnormalities much easier to see. Monitors with a resolution of 1,000 to 2,000 lines are adequate for viewing. Workstations used for viewing CT are capable of interpolating or smoothing the CT data, producing ­ smaller pixels in the resultant image than are present in the scan itself. Although individual CT pixels are clearly visible on close inspection of an unprocessed HRCT ­image, this is not generally the case when viewing a study on a workstation. In fact, if unprocessed CT images (raw CT pixels) are viewed, the appearance can be disconcerting, and the images may be difficult to read. Cameras are capable of photographing CT scans using a range of settings, from sharp to smooth. If the camera is set on sharp, individual CT pixels will be visible; on a smooth setting, the data are interpolated, and image pixel size is reduced (Fig. 1-36). Although it might seem that a sharp setting would be best for HRCT, this is not the case. Resolution of fine structures is better with a smooth setting, and ­image interpretation is easier. Image compression may be used to reduce the quantity of digital data involved in the transmission and storage of images. The use of so-called lossy image compression (JPEG 2000), which is now part of the Digital Imaging and Communications in Medicine standard, ­allows the compression ratio to be adjusted. Ringl et al. (212) assessed the use of different degrees of compression on the quality of HRCT images. The authors found that images compressed with a ratio of 3:1 were indistinguishable from uncompressed images, while compression ratios of 7:1 or more resulted in substantial degradation of image quality and the potential loss of diagnostic information (212).

Window Settings The window mean and width used for image display have a significant impact on the appearance of the lung parenchyma and the dimensions of visualized structures (Fig. 1-37) (14,213). If the display technique used is not appropriate, normal structures can be made to look abnormal, or subtle abnormalities may be overlooked.

#152201  Cust: LWW  Au: Webb  Pg. No. 33

A

B FIGU RE 1-36  Image interpolation and pixel size. A: Actual CT pixels are displayed on this HRCT image in a patient with interlobular septal thickening. The thickening septa have a stair-step appearance, and centrilobular arteries appear square. B: With interpolation, the appearance of the image is considerably improved. Note that a small centrilobular bronchiole clearly seen on this image (arrow) cannot be recognized on the original image (A).

The most important window setting to use in display is the so-called lung window. It should be emphasized that there is no single correct or ideal window setting for the demonstration of lung anatomy on HRCT, and several combinations of window mean and width may be appropriate (214). Within limits, the precise window width and levels chosen are a matter of personal preference; the values indicated here should serve only as guidelines. However, it is important that a single lung window setting be used consistently in all patients. Unless this is done, it is difficult to compare one case to another, develop

DESIGN SERVICES OF

34

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

—300/500 HU (window mean/window width)

C

—500/500 HU (window mean/window width)

D

—700/500 HU (window mean/window width)

E

—900/500 HU (window mean/window width)

F

—300/1,000 HU (window mean/window width)

—500/1,000 HU (window mean/window width)

FIGU RE 1-37  A–L: Effects of window mean and width on the visibility of bronchi and vessels in a normal subject. Using a narrow window width (500 HU), a high window mean (e.g., –300 HU) makes bronchi and bronchial walls difficult to see, whereas a low window mean (e.g., –900 HU) accentuates the apparent thickness of bronchial walls and the diameter of vessels. This effect decreases with increasing window width (i.e., 1,500 HU) (I–L). A window mean of approximately –450 HU and width of 1,000 to 1,400 HU have been shown to be best suited to measuring bronchial wall thickness.

#152201  Cust: LWW  Au: Webb  Pg. No. 34

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

G

H

—700/1,000 HU (window mean/window width)

—900/1,000 HU (window mean/window width)

J

I —300/1,500 HU (window mean/window width)

K

—500/1,500 HU (window mean/window width)

L

—700/1,500 HU (window mean/window width)

—900/1,500 HU (window mean/window width)

FIGU RE 1-37  (Continued)

#152201  Cust: LWW  Au: Webb  Pg. No. 35

DESIGN SERVICES OF

35

36

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

an understanding of what appearances are normal and abnormal, and compare sequential examinations in the same patient. Although it is not inappropriate to use some additional window settings in specific cases, depending on what abnormality is being sought, the effects of the variations in window settings on the appearance of the resulting images must be kept in mind. Window level settings ranging from –600 to –700 HU and window widths of 1,000 to 1,500 HU are appropriate for a routine lung window (Fig. 1-37). The use of an extended window width (i.e., 2,000 HU) reduces contrast between lung parenchymal structures, such as vessels, bronchi, and the air-containing alveoli, and may make interstitial structures appear less conspicuous or thinner than they actually are. In contrast, extended windows may be of some value in detecting abnormalities of overall lung attenuation (215,216) and are also useful in evaluating the relationship of peripheral parenchymal abnormalities to the pleural surfaces. A window width of less than 1,000 HU is not generally appropriate for viewing lung parenchyma because it unnecessarily increases contrast and may result in an apparent increase in the size of soft-tissue structures. For example, the effect of window mean and level on the HRCT appearance of bronchial walls has been assessed using inflation-fixed lungs (213). In this study, window widths less than 1,000 HU resulted in a substantial overestimation of bronchial wall thickness, whereas window widths greater than 1,400 HU resulted in an underestimation of bronchial wall thickness (Fig. 1-37). Viewing soft-tissue windows is also important in reading HRCT. Window level/width settings of 40–50/350– 450 HU are best for evaluation of the mediastinum, hila, and pleura. Mediastinal and pleural abnormalities are sometimes of value in interpreting HRCT of the lung. For example, the presence of lymph node enlargement, esophageal dilatation, calcification, or pleural thickening may be helpful in making a correct diagnosis of lung disease. When performing an HRCT study, images are routinely displayed using both lung and soft-tissue windows. As stated, choosing different window levels can be advantageous in individual cases, despite the fact they may not be optimal for all indications (Fig. 1-37). Low window settings (–800 to –900 HU) with narrow window widths (500 HU) can be valuable in contrasting emphysema or air-filled cystic lesions with normal lung parenchyma. With such a low window mean, normal lung parenchyma looks gray, whereas areas of emphysema remain black. However, using this same window to image the lung interstitium would be improper. Such a low window mean, particularly combined with a narrow window width, would make the lung interstitium appear much more prominent than it really is and could make a normal case appear abnormal. This window would also result in overestimations of the size of vessels and of bronchial wall thickness. A window width of 2,000 HU is not generally suitable for viewing lung parenchyma because contrast is significantly reduced. However, window settings of –500 to

#152201  Cust: LWW  Au: Webb  Pg. No. 36

–700/2,000 HU may be used and are particularly useful when pleuroparenchymal abnormalities are being evaluated (3,21). It is also of diagnostic value, when using a workstation, to vary window settings or toggle rapidly between different preset window settings (e.g., lung window, wide lung window, soft-tissue window) at a given scan level. Having preset windows available is important; in one study assessing the utility of workstation viewing, interpreting HRCT studies with a fixed window (–500/2,000 HU) setting proved to be more accurate than viewing them with operator-varied window settings (215). Having preset windows available also markedly reduces the time required to interpret the images.

HIGH-RESOLUTION COMPUTED TOMOGRAPHY PROTOCOLS HRCT may be obtained in a number of different clinical settings, and to some extent, the manner in which the examination is obtained is varied according to the diseases suspected. Either spaced axial scans or MD-HRCT may be used (13). The following protocols are provided as guides, but these may be varied in individual cases.

Suspected Emphysema, Airways, or Obstructive Disease In patients suspected of having emphysema, airways disease such as bronchiectasis (217), or obstructive disease on the basis of clinical, pulmonary function, or plain radiographic findings, axial HRCT or MD-HRCT may be performed (96,143). Axial scanning should be obtained at full inspiration, at 1-cm intervals from lung apices to bases, and with the patient supine (Table 1-3); prone scans are not usually needed. Expiratory scans obtained at three or more levels are also recommended to detect air trapping. MD-HRCT, using 0.5- to 1.25-mm detector width, would be ideal for assessing airways disease and emphysema. Expiratory images may be obtained at selected levels or by using volumetric postexpiratory MD-HRCT. In patients with emphysema being evaluated for lung transplantation or volume reduction surgery, obtaining volumetric MD-HRCT would also be important for the detection of associated lung carcinoma, which has an incidence of up to 5% in this patient population (218). TABLE 1-3  Scan Protocols: Suspected Emphysema, ­ irways Disease, or Obstructive Lung Disease A Full inspiration Supine position only Axial scans with 1-cm spacing from lung apices to bases or MD-HRCT Expiratory scans at three or more levels Option: dynamic, volumetric, or spirometrically triggered expiratory scans

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

Suspected Fibrotic or Restrictive Disease, or Unknown Lung Disease In patients suspected of having a fibrotic or restrictive lung disease on the basis of clinical, pulmonary function, or plain radiographic findings, or in patients with an unknown type of respiratory disability, it would be appropriate to obtain axial HRCT scans at 1-cm intervals with the patient supine. If the chest radiograph appears normal or subtle lung disease is present, or if chest radiographs are unavailable for review, additional prone scans should be obtained, or the scans should be monitored for the presence of problematic dependent opacity (Table 1-4). If the plain radiograph shows a distinct abnormality, prone scans will not likely be needed. Prone scans at 2-cm intervals, in combination with the supine scans, are recommended when obtaining prone scans routinely; they obviate the need for reviewing plain radiographs or monitoring scans. Scans at 2-cm intervals, in both supine and prone positions, have proven to be a useful protocol for prone and supine imaging and provide the same number of images to review as obtained when scanning a patient with obstructive disease (24) (Table 1-4). In patients having their initial diagnostic evaluation, obtaining expiratory scans at three or more levels is recommended but not essential. In most patients with restrictive or fibrotic lung disease, expiratory imaging is of no diagnostic value. However, in a patient with an unknown lung disease, airway obstruction may be the cause of the patient’s disability. Furthermore, in a patient with a restrictive or fibrotic disease, the presence of air trapping visible on expiratory images may be helpful in differential diagnosis (100,219). Air trapping may be seen in several fibrotic lung diseases, most notably, hypersensitivity pneumonitis and sarcoidosis. MD-HRCT using 0.5- to 1.25-mm detector width obtained in the supine position may also be used in this TABLE 1-4  Scan Protocols: Suspected Restrictive or Fibrotic Lung Disease, or Diffuse Lung Disease of ­Unknown Type Chest radiograph abnormal Full inspiration Supine position Axial scans with 1-cm spacing from lung apices to bases or MD-HRCT Expiratory scans at three or more levels (initial examination only) Chest radiograph normal, minimally abnormal, or unavailable Full inspiration Supine position Axial scans with 1- to 2-cm spacing or MD-HRCT Prone scans with 2-cm spacing or volumetrically, or monitor scans for dependent density Expiratory scans at three or more levels (initial evaluation only)

#152201  Cust: LWW  Au: Webb  Pg. No. 37

37

TABLE 1-5  Scan Protocols: Hemoptysis Full inspiration MD-HRCT or volumetric CT with spaced axial HRCT at 1-cm intervals Contrast infusion optional Expiratory scans at three or more levels (optional for initial evaluation only)

clinical setting. Monitoring the study to determine the need for prone scans would be appropriate, although some investigators routinely obtain spaced prone scans or volumetric prone scans in this setting (13). Expiratory imaging, either axial or volumetric, may be used if needed or desired (13). In patients with restrictive disease who are having follow-up HRCT examinations, inspiratory images may be obtained at fewer levels than are appropriate for the initial diagnostic examination (30), and expiratory scans are not usually necessary. Follow-up examinations may be obtained using a low-dose technique to reduce radiation exposure.

Hemoptysis In patients who present with hemoptysis, possibly related to airway abnormalities or an endobronchial lesion, it is appropriate to obtain volumetric imaging for the detection of large airway abnormalities. This may be done using MD-HRCT or volumetric imaging in addition to spaced axial HRCT. In either instance, HRCT is needed to assess possible airways disease or to detect regions of hemorrhage appearing as ground-glass opacity or findings of vasculitis (Table 1-5) (220,221). In some situations, the injection of contrast agents may also be used to identify vascular abnormalities, such as arteriovenous fistula, pulmonary artery aneurysm, or bronchial artery enlargement (46,47). An ideal examination would be ­contrast-enhanced MD-HRCT.

Suspected Pulmonary Vascular Disease Some patients may have symptoms or signs (e.g., hypoxemia, pulmonary hypertension) that may result from lung disease (e.g., emphysema), pulmonary vascular disease (e.g., chronic pulmonary embolism, vasculitis), or a combination of these (222–226). In such patients, combining HRCT with a contrast-enhanced helical CT may be necessary for diagnosis. The HRCT study is used to detect findings of lung disease or small vessel disease, and the contrast-enhanced helical CT is used to detect large vessel abnormalities and vascular obstruction. The use of contrast-enhanced MD-HRCT with 0.5- to 1.25-mm detector width would be ideal for this indication, allowing the detailed assessment of both large and small vessel abnormalities and associated lung disease (Figs. 1-18 and 1-19) (46,47,227–230) (Table 1-6).

DESIGN SERVICES OF

38

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

TABLE 1-6  Scan Protocols: Suspected Pulmonary ­Vascular Disease Full inspiration MD-HRCT or volumetric CT with spaced axial HRCT at 1-cm intervals Contrast infusion Expiratory scans at three or more levels (optional for initial evaluation only)

SPATIAL RESOLUTION OF HIGH-RESOLUTION COMPUTED TOMOGRAPHY A fundamental relationship exists between pixel size and the size of structures that can be resolved by CT. For optimal matching of image display to the attainable spatial resolution of the scanner, there should be two  pixels for the smallest structure resolved (11). Using an FOV is sufficient to image the entire thorax, pixel size is approximately 0.5 mm, and current scanners with thin detectors are capable of providing a resolution of 10 to 12 line pairs per centimeter using a high-resolution algorithm, with similar line-pair resolution in the z-axis (10,199,231,232). Structures smaller than the pixel size should be difficult to resolve on HRCT; however, this is sometimes possible. This likely occurs because of the large differences in attenuation between the soft-tissue structures imaged and the air-filled alveoli surrounding them, and the use

of a high spatial frequency algorithm for reconstruction, which often results in some edge enhancement. The ability of HRCT to resolve fine lung structures also depends on their orientation relative to the scan plane (Fig. 1-38). Structures measuring 0.1 to 0.2 mm in thickness can be seen if they are oriented perpendicular to the scan plane and extend through the thickness of the scan plane or voxel (e.g., 1 mm) (4,10,233,234). For example, interlobular septa as thin as 0.1 mm and vessels with a diameter of 0.3 mm are sometimes visible on HRCT using a small FOV, and when oriented correctly. Similarly sizedstructures (i.e., 0.1–0.3 mm) that are oriented horizontally within the scan plane will not be visible because of volume averaging with the air-filled lung, which occupies most of the thickness of the voxel. Bronchi or bronchioles measuring less than 2 to 3 cm in diameter and having a wall thickness of approximately 0.3 mm are usually invisible in peripheral lung because they have courses that lie roughly in the plane of scan. Bronchi or bronchioles of similar sizes are sometimes visible when oriented perpendicular to the plane of scan. It should be kept in mind that, although soft-tissue structures can be resolved when they are thinner or smaller than the pixel size, their apparent size in the final HRCT image will be determined, at least partially, by the pixel size and the interpolation algorithm used in the workstation or camera and not by their actual dimensions. This can make the measurement of such small structures on HRCT difficult and prone to inaccuracies.

HRCT image

Volume scanned

Voxel diameter 0.3 mm

1.5-mm-thick scan

Pixel diameter 0.3 mm

1.5 mm

Tissue plane Perpendicular visible cylinder Tissue plane (0.1 mm thick) perpendicular to scan plane

Cylinders (0.2 mm in diameter) perpendicular and horizontal relative to scan plane

FIGU RE 1-38  Resolution of structures relative to their size, shape, and orientation. The tissue plane, 0.1 mm thick, and the perpendicular cylinder, 0.2 mm in diameter, are visible on the HRCT scan because they extend through the thickness of the scan volume or voxel. The horizontal cylinder cannot be seen.

#152201  Cust: LWW  Au: Webb  Pg. No. 38

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT

Resolution in the transverse plane using MD-HRCT is similar to that reported with axial scans. The resolution of lung structures on images reconstructed in the sagittal or coronal planes depends on the detector width and pitch used. With a pitch of 1 and the use of thin ­(0.5–0.625-mm) detectors or collimation, fine lung structures can be resolved on reconstructed images, and the data set appears to be nearly isotropic (20). For example, in a study of an anthropomorphic line-pair phantom, coronal reconstructions obtained using 16-detector MDCT and 0.625-mm detector width with a pitch of 1.75 were found to be nearly identical to transverse images in spatial resolution, up to 9.8 line pairs per centimeter (232); coronal reconstructions of MDCT using 1.25-mm detectors and a pitch of 1.375 showed a decrease in spatial resolution. In a study of the appearances of interlobar fissures on sagittal reconstructions from MD-HRCT, it was found that a collimation of 0.5 mm was necessary for visualization of the minor fissure as a sharp line (235). It has also been shown that the resolution of anatomical structures on reconstructed coronal images from MD-HRCT was identical to that of direct coronal images when 0.5-mm collimation was used for the MD-HRCT study (56). Resolution with reconstructed MD-HRCT was inferior when 1-mm collimation was used.

HIGH-RESOLUTION COMPUTED TOMOGRAPHY ARTIFACTS Several confusing artifacts can be seen on HRCT. However, familiarity with their appearances should eliminate potential misdiagnoses (3,10,214,236–238).

Streak Artifacts Streak artifacts that radiate from the edges of or adjacent to sharply marginated, high-contrast structures such as bronchial walls, ribs, or vertebral bodies are common on HRCT. On HRCT, streak artifacts are often visible as fine, linear, or netlike opacities that can be seen anywhere but are most commonly found overlying the posterior lung, paralleling the pleural surface and posterior chest wall (10). Although streak artifacts degrade the image, they do not usually mimic pathology or cause confusion in image interpretation. Streak artifacts are thinner and less dense, and have a different appearance than the normal or abnormal interstitium (interlobular septa) visible in this region. Streak artifacts can result from several mechanisms: beam hardening, photon starvation, and aliasing. Streak artifacts are more evident on scans obtained with low mA (88,214,238). Photon starvation results in prominent streak artifacts and is most notable in the paravertebral regions, adjacent to the highly attenuating vertebral bodies. It is related to insufficient photons reaching the CT detectors (11,238). This type of artifact is strongly related to radiation dose and can be minimized by increasing kV(p) and mA.

#152201  Cust: LWW  Au: Webb  Pg. No. 39

39

Adaptive or automatic tube current modulation may be used to reduce this type of streak artifact (238). Aliasing is a geometric phenomenon that occurs because of undersampling of spatial information and is related to detector spacing and scan collimation (11,238). It may appear as fine stripes radiating from the edge of a dense structure or at a distance from it (238). Because it is independent of radiation dose, increasing scan technique is of no value in reducing this type of artifact. The use of MDCT for obtaining HRCT introduces the possibility of a variety of helical CT artifacts, but most are reduced by using narrow detector width and low pitch (as in MD-HRCT) (238). Partial volume artifact occurs when a dense object is located off-center and is incompletely scanned by the x-ray beam when it is directed in different directions. This can result in the presence of lighter and darker shading adjacent to the object, but is minimized by using thin collimation or high pitch. When axial imaging is performed with a helical scanner, some distortion of the shape of objects may occur because of the helical reconstruction algorithm. This is greatest when an object rapidly changes shape along the z-axis. Ring artifacts and windmill artifacts may be seen with helical CT and are familiar to most radiologists (238). If 2D or 3D reconstructions are performed, additional artifacts may be introduced. These include stair-step artifacts, which occur when using scan data obtained with thick detectors or collimation and nonoverlapping reconstructions, and appear as jagged edges; this artifact is not conspicuous with MD-HRCT because of the thin detectors used. Zebra artifact results in the presence of horizontal stripes of varying density in the reconstructed image, corresponding to the thickness of detectors used, because of noise inhomogeneity in the z-axis.

Motion-Related Artifacts Pulsation or star artifacts are commonly visible, particularly at the left lung base, adjacent to the heart (Figs. 1-39 and 1-40). With pulsation artifacts, thin streaks radiate from the edges of vessels or other visible structures, which therefore resemble stars, and small areas of apparent lucency may be seen between these streaks. These lucent areas, if not recognized as artifactual, may be mistaken for dilated bronchi (237). On MD-HRCT, images reconstructed in the sagittal or coronal planes may show stairstep artifacts due to cardiac pulsation. The major fissure, usually on the left (Figs. 1-39 to ­1-41), or other parenchymal structures such as vessels and bronchi may be seen as double because of cardiac pulsation or respiration during the scan (214,236). This appearance of doubling artifacts can mimic bronchiectasis (Fig. 1-41). It results when a linear structure, such as the fissure or vessel, is in a slightly different position when scanned by the gantry from opposite directions (180  ­ degrees apart) (Fig.  1-41). As with image noise, these artifacts are much more conspicuous when

DESIGN SERVICES OF

40

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

Left major fissure

Detector

X-ray tube

X-ray beam tangent to fissure

A Original position of major fissure

Major fissure displaced because of cardiac motion

Detector

X-ray tube

FIGU RE 1-39  “Double fissure” artifact. The left major fissure (arrows) appears to be double. Fine streak artifacts are visible posteriorly. Pulsation artifacts are also visible adjacent to the left heart border.

B

Fissure appears double on resulting image

C FIGU RE 1-41  Mechanism of “double fissure” artifact. The major fissure is seen by the scanner only when the X-ray beam is tangent to it. If the position of the fissure is slightly altered by cardiac pulsation during the period in which the gantry has rotated 180 degrees (A, B), it appears to be seen in two different locations on the resulting image (C).

high-resolution techniques are used, simply because they are more sharply resolved. Motion-related artifacts can be reduced by ECG gating of scan acquisition (199), by using scanners with very rapid scan times (100 ms) (153), or by spirometrically controlled respiration during scanning (147,148,238).

REFERENCES

FIGU RE 1-40  Bronchiectasis artifact (“pseudobronchiectasis”). Several linear structures (arrows) appear double, mimicking bronchiectasis.

#152201  Cust: LWW  Au: Webb  Pg. No. 40

1. Müller NL, Miller RR. Computed diffuse infiltrative lung disease: part 1990;142:1206–1215. 2. Müller NL, Miller RR. Computed diffuse infiltrative lung disease: part 1990;142:1440–1448.

tomography of chronic 1. Am Rev Respir Dis tomography of chronic 2. Am Rev Respir Dis

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT 3. Webb WR. High-resolution CT of the lung parenchyma. Radiol Clin North Am 1989;27:1085–1097. 4. Zerhouni E. Computed tomography of the pulmonary parenchyma: an overview. Chest 1989;95:901–907. 5. Todo G, Itoh H, Nakano Y, et al. High-resolution CT for the evaluation of pulmonary peripheral disorders [Article in Japanese]. Jpn J Clin Radiol 1982;27:1319–1326. 6. Nakata H, Kimoto T, Nakayama T, et al. Diffuse peripheral lung disease: evaluation by high-resolution computed tomography. ­Radiology 1985;157:181–185. 7. Naidich DP, Zerhouni EA, Hutchins GM, et al. Computed tomography of the pulmonary parenchyma: part 1. Distal air-space disease. J Thorac Imaging 1985;1:39–53. 8. Zerhouni EA, Naidich DP, Stitik FP, et al. Computed tomography of the pulmonary parenchyma: part 2. Interstitial disease. J Thorac Imaging 1985;1:54–64. 9. Theocharopoulos N, Perisinakis K, Damilakis J, et al. Dosimetric characteristics of a 16-slice computed tomography scanner. Eur Radiol 2006;16:2575–2585. 10. Mayo JR, Webb WR, Gould R, et al. High-resolution CT of the lungs: an optimal approach. Radiology 1987;163:507–510. 11. Mayo JR. High resolution computed tomography: technical aspects. Radiol Clin North Am 1991;29:1043–1049. 12. Murata K, Khan A, Rojas KA, et al. Optimization of computed tomography technique to demonstrate the fine structure of the lung. Invest Radiol 1988;23:170–175. 13. Sundaram B, Chughtai AR, Kazerooni EA. Multidetector highresolution computed tomography of the lungs: protocols and applications. J Thorac Imaging 2010;25:125–141. 14. Webb WR, Gamsu G, Wall SD, et al. CT of a bronchial phantom: factors affecting appearance and size measurements. Invest Radiol 1984;19:394–398. 15. Mayo JR. The high-resolution computed tomography technique. Semin Roentgenol 1991;26:104–109. 16. Mayo JR, Aldrich J, Muller NL. Radiation exposure at chest CT: a statement of the Fleischner society. Radiology 2003;228:15–21. 17. Zwirewich CV, Terriff B, Müller NL. High-spatial-frequency (bone) algorithm improves quality of standard CT of the thorax. AJR Am J Roentgenol 1989;153:1169–1173. 18. Remy-Jardin M, Giraud F, Remy J, et al. Pulmonary sarcoidosis: role of CT in the evaluation of disease activity and functional impairment and in prognosis assessment. Radiology 1994;191:675–680. 19. Das M, Mahnken AH, Muhlenbruch G, et al. Individually adapted examination protocols for reduction of radiation exposure for 16-MDCT chest examinations. AJR Am J Roentgenol 2005;184:1437–1443. 20. Beigelman-Aubry C, Hill C, Guibal A, et al. Multi-detector row CT and postprocessing techniques in the assessment of diffuse lung disease. Radiographics 2005;25:1639–1652. 21. Murata K, Khan A, Herman PG. Pulmonary parenchymal disease: evaluation with high-resolution CT. Radiology 1989;170:629–635. 22. Aberle DR, Gamsu G, Ray CS, et al. Asbestos-related pleural and parenchymal fibrosis: detection with high-resolution CT. Radiology 1988;166:729–734. 23. Lee KN, Yoon SK, Sohn CH, et al. Dependent lung opacity at thinsection CT: evaluation by spirometrically-gated CT of the influence of lung volume. Korean J Radiol 2002;3:24–29. 24. Volpe J, Storto ML, Lee K, et al. High-resolution CT of the lung: determination of the usefulness of CT scans obtained with the patient prone based on plain radiographic findings. AJR Am J Roentgenol 1997;169:369–374. 25. Mathieson JR, Mayo JR, Staples CA, et al. Chronic diffuse infiltrative lung disease: comparison of diagnostic accuracy of CT and chest radiography. Radiology 1989;171:111–116. 26. Swensen SJ, Aughenbaugh GL, Brown LR. High-resolution computed tomography of the lung. Mayo Clin Proc 1989;64:​ 1284–1294. 27. Aberle DR, Gamsu G, Ray CS. High-resolution CT of benign asbestos-related diseases: clinical and radiographic correlation. ­ AJR Am J Roentgenol 1988;151:883–891.

#152201  Cust: LWW  Au: Webb  Pg. No. 41

41

28. Chooi WK, Morcos SK. High resolution volume imaging of airways and lung parenchyma with multislice CT. Br J Radiol 2004;77(special issue 1):S98–S105. 29. Aziz ZA, Padley SP, Hansell DM. CT techniques for imaging the lung: recommendations for multislice and single slice computed tomography. Eur J Radiol 2004;52:119–136. 30. Kazerooni EA, Martinez FJ, Flint A, et al. Thin-section CT obtained at 10-mm increments versus limited three-level thin-section CT for idiopathic pulmonary fibrosis: correlation with pathologic scoring. AJR Am J Roentgenol 1997;169:977–983. 31. Gamsu G, Aberle DR, Lynch D. Computed tomography in the diagnosis of asbestos-related thoracic disease. J Thorac Imaging 1989;4:61–67. 32. Friedman AC, Fiel SB, Fisher MS, et al. Asbestos-related pleural disease and asbestosis: a comparison of CT and chest radiography. AJR Am J Roentgenol 1988;150:268–275. 33. Friedman AC, Fiel SB, Radecki PD, et al. Computed tomography of benign pleural and pulmonary parenchymal abnormalities related to asbestos exposure. Semin Ultrasound CT MR 1990;11:393–408. 34. Staples CA. Computed tomography in the evaluation of ­benign ­asbestos-related disorders. Radiol Clin North Am 1992;30:​ 1191–1207. 35. Murray K, Gamsu G, Webb WR, et al. High-resolution CT sampling for detection of asbestos-related lung disease. Acad Radiol 1995;2:111–115. 36. Henschke CI. Image selection for computed tomography of the chest: a sampling approach. Invest Radiol 1992;27:908–911. 37. Leswick DA, Webster ST, Wilcox BA, et al. Radiation cost of helical high-resolution chest CT. AJR Am J Roentgenol 2005;184:742–745. 38. Engeler CE, Tashjian JH, Engeler CM, et al. Volumetric highresolution CT in the diagnosis of interstitial lung disease and bronchiectasis: diagnostic accuracy and radiation dose. AJR Am J Roentgenol 1994;163:31–35. 39. Kalender WA, Seissler W, Klotz E, et al. Spiral volumetric CT with single-breath-hold technique, continuous transport, and continuous scanner rotation. Radiology 1990;176:181–183. 40. Vock P, Soucek M. Spiral computed tomography in the assessment of focal and diffuse lung disease. J Thorac Imaging 1993;8:283–290. 41. Paranjpe DV, Bergin CJ. Spiral CT of the lungs: optimal technique and resolution compared with conventional CT. AJR Am J Roentgenol 1994;162:561–567. 42. Heiken JP, Brink JA, Vannier MW. Spiral (helical) CT. Radiology 1993;189:647–656. 43. Polacin A, Kalender WA, Marchal G. Evaluation of section sensitivity profiles and image noise in spiral CT. Radiology 1992;185:29–35. 44. Kuriyama K, Tateishi R, Kumatani T, et al. Pleural invasion by peripheral bronchogenic carcinoma: assessment with three-­ dimensional helical CT. Radiology 1994;191:365–369. 45. Flohr TG, Schaller S, Stierstorfer K, et al. Multi-detector row CT systems and image-reconstruction techniques. Radiology 2005;235:756–773. 46. Bruzzi JF, Remy-Jardin M, Delhaye D, et al. Multi-detector row CT of hemoptysis. Radiographics 2006;26:3–22. 47. Engelke C, Schaefer-Prokop C, Schirg E, et al. High-resolution CT and CT angiography of peripheral pulmonary vascular disorders. Radiographics 2002;22:739–764. 48. Sumikawa H, Johkoh T, Koyama M, et al. Image quality of highresolution CT with 16-channel multidetector-row CT: comparison between helical scan and conventional step-shoot scan. Radiat Med 2005;23:539–544. 49. Schoepf UJ, Bruening RD, Hong C, et al. Multislice helical CT of focal and diffuse lung disease: comprehensive diagnosis with reconstruction of contiguous and high-resolution CT sections from a single thin-collimation scan. AJR Am J Roentgenol 2001;177:179–184. 50. Honda O, Johkoh T, Tomiyama N, et al. High-resolution CT using multidetector CT equipment: evaluation of image quality in 11 cadaveric lungs and a phantom. AJR Am J Roentgenol 2001;177:875–879.

DESIGN SERVICES OF

42

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

51. Kelly DM, Hasegawa I, Borders R, et al. High-resolution CT using MDCT: comparison of degree of motion artifact between volumetric and axial methods. AJR Am J Roentgenol 2004;182:757–759. 52. Studler U, Gluecker T, Bongartz G, et al. Image quality from highresolution CT of the lung: comparison of axial scans and of sections reconstructed from volumetric data acquired using MDCT. AJR Am J Roentgenol 2005;185:602–607. 53. Bendaoud S, Remy-Jardin M, Wallaert B, et al. Sequential versus volumetric computed tomography in the follow-up of chronic bronchopulmonary diseases: comparison of diagnostic information and radiation dose in 63 adults. J Thorac Imaging 2011;26:190–195. doi:10.1097/RTI.0b013e3181f3a30e 54. Winklehner A, Berger N, Maurer B, et al. Screening for interstitial lung disease in systemic sclerosis: the diagnostic accuracy of HRCT image series with high increment and reduced number of slices. Ann Rheum Dis 2012;71:549–552. 55. Dodd JD, de Jong PA, Levy RD, et al. Conventional high-­ resolution CT versus contiguous multidetector CT in the detection of bronchiolitis obliterans syndrome in lung transplant recipients. J Thorac Imaging 2008;23:235–243. doi:10.1097/ RTI.0b013e3181783384 56. Honda O, Johkoh T, Yamamoto S, et al. Comparison of quality of multiplanar reconstructions and direct coronal multidetector CT scans of the lung. AJR Am J Roentgenol 2002;179:875–879. 57. Kwan SW, Partik BL, Zinck SE, et al. Primary interpretation of thoracic MDCT images using coronal reformations. AJR Am J Roentgenol 2005;185:1500–1508. 58. Arakawa H, Sasaka K, Lu WM, et al. Comparison of axial highresolution CT and thin-section multiplanar reformation (MPR) for diagnosis of diseases of the pulmonary parenchyma: preliminary study in 49 patients. J Thorac Imaging 2004;19:24–31. 59. Remy-Jardin M, Campistron P, Amara A, et al. Usefulness of coronal reformations in the diagnostic evaluation of infiltrative lung disease. J Comput Assist Tomogr 2003;27:266–273. 60. Nishino M, Kuroki M, Boiselle PM, et al. Sagittal reformations of volumetric inspiratory and expiratory high-resolution CT of the lung. Acad Radiol 2004;11:1282–1290. 61. Bhalla M, Naidich DP, McGuinness G, et al. Diffuse lung disease: assessment with helical CT—preliminary observations of the role of maximum and minimum intensity projection images [see comments]. Radiology 1996;200:341–347. 62. Coakley FV, Cohen MD, Johnson MS, et al. Maximum intensity projection images in the detection of simulated pulmonary nodules by spiral CT. Br J Radiol 1998;71:135–140. 63. Remy-Jardin M, Remy J, Artaud D, et al. Diffuse infiltrative lung disease: clinical value of sliding-thin-slab maximum intensity projection CT scans in the detection of mild micronodular patterns [see comments]. Radiology 1996;200:333–339. 64. Sakai M, Murayama S, Gibo M, et al. Can maximum intensity projection images with multidetector-row computed tomography help to differentiate between the micronodular distribution of focal and diffuse infiltrative lung diseases? J Comput Assist Tomogr 2005;29:588–591. 65. Prasad SR, Wittram C, Shepard JA, et al. Standard-dose and 50%-reduced-dose chest CT: comparing the effect on image quality. AJR Am J Roentgenol 2002;179:461–465. 66. Zompatori M, Fasano L, Mazzoli M, et al. Spiral CT evaluation of pulmonary emphysema using a low-dose technique. Radiol Med (Torino) 2002;104:13–24. 67. Smith-Bindman R, Lipson J, Marcus R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 2009;169:2078–2086. 68. Berrington de Gonzalez A, Mahesh M, Kim KP, et al. Projected cancer risks from computed tomographic scans performed in the united states in 2007. Arch Intern Med 2009;169:2071–2077. 69. Bankier AA, Tack D. Dose reduction strategies for thoracic multidetector computed tomography: background, current issues, and recommendations. J Thorac Imaging 2010;25:278–288. doi:10.1097/RTI.0b013e3181eebc49

#152201  Cust: LWW  Au: Webb  Pg. No. 42

70. Mayo JR. CT evaluation of diffuse infiltrative lung disease: dose considerations and optimal technique. J Thorac Imaging 2009;24:252–259. doi:10.1097/RTI.0b013e3181c227b2 71. Christner JA, Kofler JM, McCollough CH. Estimating effective dose for CT using dose-length product compared with using organ doses: consequences of adopting international commission on radiological protection publication 103 or dual-energy scanning. AJR Am J Roentgenol 2010;194:881–889. 72. Mayo JR, Jackson SA, Müller NL. High-resolution CT of the chest: radiation dose. AJR Am J Roentgenol 1993;160:479–481. 73. Madani A, De Maertelaer V, Zanen J, et al. Pulmonary emphysema: radiation dose and section thickness at multidetector CT quantification—comparison with macroscopic and microscopic morphometry. Radiology 2007;243:250–257. 74. Koyama H, Ohno Y, Yamazaki Y, et al. Reduced-dose chest CT with 3D automatic exposure control vs. standard chest CT: quantitative assessment of emphysematous changes in smokers’ lung parenchyma. Eur J Radiol 2012;81:1330–1334. 75. Mayo JR, Kim KI, MacDonald SL, et al. Reduced radiation dose helical chest CT: effect on reader evaluation of structures and lung findings. Radiology 2004;232:749–756. 76. Yi CA, Lee KS, Kim TS, et al. Multidetector CT of bronchiectasis: effect of radiation dose on image quality. AJR Am J Roentgenol 2003;181:501–505. 77. Jung KJ, Lee KS, Kim SY, et al. Low-dose, volumetric helical CT: image quality, radiation dose, and usefulness for evaluation of bronchiectasis. Invest Radiol 2000;35:557–563. 78. Kalra MK, Rizzo S, Maher MM, et al. Chest CT performed with z-axis modulation: scanning protocol and radiation dose. Radiology 2005;237:303–308. 79. Angel E, Yaghmai N, Jude CM, et al. Dose to radiosensitive organs during routine chest CT: effects of tube current modulation. AJR Am J Roentgenol 2009;193:1340–1345. 80. Mets OM, Willemink MJ, de Kort FP, et al. The effect of iterative reconstruction on computed tomography assessment of emphysema, air trapping and airway dimensions. Eur Radiol 2012;22:​ 2103–2109. 81. Singh S, Kalra MK, Gilman MD, et al. Adaptive statistical iterative reconstruction technique for radiation dose reduction in chest CT: a pilot study. Radiology 2011;259:565–573. 82. Prakash P, Kalra MK, Ackman JB, et al. Diffuse lung disease: CT of the chest with adaptive statistical iterative reconstruction technique. Radiology 2010;256:261–269. 83. Kim YK, Sung YM, Choi JH, et al. Reduced radiation exposure of the female breast during low-dose chest CT using organ-based tube current modulation and a bismuth shield: comparison of image quality and radiation dose. AJR Am J Roentgenol 2013;200:537–544. 84. Colombo P, Pedroli G, Nicoloso M, et al. Evaluation of the efficacy of a bismuth shield during CT examinations. Radiol Med 2004;108:560–568. 85. Coursey C, Frush DP, Yoshizumi T, et al. Pediatric chest MDCT using tube current modulation: effect on radiation dose with breast shielding. AJR Am J Roentgenol 2008;190:W54–W61. 86. Leswick DA, Hunt MM, Webster ST, et al. Thyroid shields versus z-axis automatic tube current modulation for dose reduction at neck CT. Radiology 2008;249:572–580. 87. AAPM Board of Directors. AAPM position statement on the use of bismuth shielding for the purpose of dose reduction in CT scanning. Policy PP-26-A [serial on the Internet]. 2012. https://www​ .aapm.org/publicgeneral/BismuthShielding.pdf. 88. Zwirewich CV, Mayo JR, Müller NL. Low-dose high-resolution CT of lung parenchyma. Radiology 1991;180:413–417. 89. Lee KS, Primack SL, Staples CA, et al. Chronic infiltrative lung disease: comparison of diagnostic accuracies of radiography and low- and conventional-dose thin-section CT. Radiology 1994;191:​669–673. 90. Reuter FG, Conway BJ, McCrohan ML, et al. Average radiation exposure values for three diagnostic radiographic examinations. Radiology 1990;177:341–345. 91. Müller NL. Clinical value of high-resolution CT in chronic diffuse lung disease. AJR Am J Roentgenol 1991;157:1163–1170.

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT 92. Majurin ML, Varpula M, Kurki T, et al. High-resolution CT of the lung in asbestos-exposed subjects. Comparison of low-dose and high-dose HRCT. Acta Radiol 1994;35:473–477. 93. Majurin ML, Valavaara R, Varpula M, et al. Low-dose and ­conventional-dose high resolution CT of pulmonary changes in breast cancer patients treated by tangential field radiotherapy. Eur J Radiol 1995;20:114–119. 94. Christe A, Charimo-Torrente J, Roychoudhury K, et al. Accuracy of low-dose computed tomography (CT) for detecting and characterizing the most common CT-patterns of pulmonary disease. Eur J Radiol 2013;82:e142–e150. 95. Arakawa H, Webb WR. Expiratory high-resolution CT scan. ­Radiol Clin North Am 1998;36:189–209. 96. Webb WR. Radiology of obstructive pulmonary disease. AJR Am J Roentgenol 1997;169:637–647. 97. Chen D, Webb WR, Storto ML, et al. Assessment of air trapping using postexpiratory high-resolution computed tomography. J Thorac Imaging 1998;13:135–143. 98. Lucidarme O, Coche E, Cluzel P, et al. Expiratory CT scans for chronic airway disease: correlation with pulmonary function test results. AJR Am J Roentgenol 1998;170:301–307. 99. Arakawa H, Webb WR. Air trapping on expiratory high-resolution CT scans in the absence of inspiratory scan abnormalities: correlation with pulmonary function tests and differential diagnosis. AJR Am J Roentgenol 1998;170:1349–1353. 100. Arakawa H, Webb WR, McCowin M, et al. Inhomogeneous lung attenuation at thin-section CT: diagnostic value of expiratory scans. Radiology 1998;206:89–94. 101. Vock P, Malanowski D, Tschaeppeler H, et al. Computed tomographic lung density in children. Invest Radiol 1987;22:627–631. 102. Robinson PJ, Kreel L. Pulmonary tissue attenuation with computed tomography: comparison of inspiration and expiration scans. J Comput Assist Tomogr 1979;3:740–748. 103. Millar AB, Denison DM. Vertical gradients of lung density in healthy supine men. Thorax 1989;44:485–490. 104. Webb WR, Stern EJ, Kanth N, et al. Dynamic pulmonary CT: findings in normal adult men. Radiology 1993;186:117–124. 105. Verschakelen JA, Van Fraeyenhoven L, Laureys G, et al. Differences in CT density between dependent and nondependent portions of the lung: influence of lung volume. AJR Am J Roentgenol 1993;161:713–717. 106. Webb WR, Sarin M, Zerhouni EA, et al. Interobserver variability in CT and MR staging of lung cancer. J Comput Assist Tomogr 1993;17:841–846. 107. Knudson RJ, Standen JR, Kaltenborn WT, et al. Expiratory computed tomography for assessment of suspected pulmonary emphysema. Chest 1991;99:1357–1366. 108. Stern EJ, Webb WR, Gamsu G. Dynamic quantitative computed tomography: a predictor of pulmonary function in obstructive lung diseases. Invest Radiol 1994;29:564–569. 109. Kitahara Y, Takamoto M, Maruyama M, et al. Differential diagnosis of pulmonary emphysema using the CT index: LL%w [Article in Japanese]. Nippon Kyobu Shikkan Gakkai Zasshi 1989;27:689–695. 110. Kauczor HU, Hast J, Heussel CP, et al. CT attenuation of paired HRCT scans obtained at full inspiratory/expiratory position: comparison with pulmonary function tests. Eur Radiol 2002;12:2757–2763. 111. Newman KB, Lynch DA, Newman LS, et al. Quantitative computed tomography detects air trapping due to asthma. Chest 1994;106:105–109. 112. Park CS, Müller NL, Worthy SA, et al. Airway obstruction in asthmatic and healthy individuals: inspiratory and expiratory thinsection CT findings. Radiology 1997;203:361–367. 113. Lynch DA. Imaging of asthma and allergic bronchopulmonary mycosis. Radiol Clin North Am 1998;36:129–142. 114. Beigelman-Aubry C, Capderou A, Grenier PA, et al. Mild in termittent asthma: CT assessment of bronchial cross-sectional area and lung attenuation at controlled lung volume. Radiology 2002;223:181–187.

#152201  Cust: LWW  Au: Webb  Pg. No. 43

43

115. Silva CI, Colby TV, Muller NL. Asthma and associated conditions: high-resolution CT and pathologic findings. AJR Am J Roentgenol 2004;183:817–824. 116. Dodd JD, Barry SC, Barry RB, et al. Thin-section CT in patients with cystic fibrosis: correlation with peak exercise capacity and body mass index. Radiology 2006;240:236–245. 117. Garg K, Lynch DA, Newell JD, et al. Proliferative and constrictive bronchiolitis: classification and radiologic features. AJR Am J Roentgenol 1994;162:803–808. 118. Padley SPG, Adler BD, Hansell DM, et al. Bronchiolitis obliterans: high-resolution CT findings and correlation with pulmonary function tests. Clin Radiol 1993;47:236–240. 119. Moore ADA, Godwin JD, Dietrich PA, et al. Swyer-James syndrome: CT findings in eight patients. AJR Am J Roentgenol 1992;158:1211–1215. 120. Marti-Bonmati L, Ruiz PF, Catala F, et al. CT findings in SwyerJames syndrome. Radiology 1989;172:477–480. 121. Sweatman MC, Millar AB, Strickland B, et al. Computed tomography in adult obliterative bronchiolitis. Clin Radiol 1990;41:116–119. 122. Aquino SL, Webb WR, Golden J. Bronchiolitis obliterans associated with rheumatoid arthritis: findings on HRCT and dynamic expiratory CT. J Comput Assist Tomogr 1994;18:555–558. 123. Leung AN, Fisher K, Valentine V, et al. Bronchiolitis obliterans after lung transplantation: detection using expiratory HRCT. Chest 1998;113:365–370. 124. Yang CF, Wu MT, Chiang AA, et al. Correlation of high-resolution CT and pulmonary function in bronchiolitis obliterans: a study based on 24 patients associated with consumption of Sauropus androgynus. AJR Am J Roentgenol 1997;168:1045–1050. 125. Konen E, Gutierrez C, Chaparro C, et al. Bronchiolitis obliterans syndrome in lung transplant recipients: can thin-section CT findings predict disease before its clinical appearance? Radiology 2004;231:467–473. 126. Knollmann FD, Ewert R, Wundrich T, et al. Bronchiolitis obliterans syndrome in lung transplant recipients: use of spirometrically gated CT. Radiology 2002;225:655–662. 127. Bankier AA, Van Muylem A, Scillia P, et al. Air trapping in heart-lung transplant recipients: variability of anatomic distribution and extent at sequential expiratory thin-section CT. Radiology 2003;229:737–742. 128. Stern EJ, Webb WR, Golden JA, et al. Cystic lung disease associated with eosinophilic granuloma and tuberous sclerosis: air trapping at dynamic ultrafast high-resolution CT. Radiology 1992;182:325–329. 129. Hansell DM, Wells AU, Rubens MB, et al. Bronchiectasis: functional significance of areas of decreased attenuation at expiratory CT. Radiology 1994;193:369–374. 130. Gelman M, King MA, Neal DE, et al. Focal air trapping in patients with HIV infection: CT evaluation and correlation with pulmonary function test results. AJR Am J Roentgenol 1999;172:1033–1038. 131. Khong PL, Chan GC, Lee SL, et al. Beta-thalassemia major: thinsection CT features and correlation with pulmonary function and iron overload. Radiology 2003;229:507–512. 132. Hansell DM, Wells AU, Padley SP, et al. Hypersensitivity pneumonitis: correlation of individual CT patterns with functional abnormalities. Radiology 1996;199:123–128. 133. Small JH, Flower CD, Traill ZC, et al. Air-trapping in extrin sic allergic alveolitis on computed tomography. Clin Radiol 1996;51:684–688. 134. Hansell DM, Milne DG, Wilsher ML, et al. Pulmonary sarcoidosis: morphologic associations of airflow obstruction at thin-section CT. Radiology 1998;209:697–704. 135. Gleeson FV, Traill ZC, Hansell DM. Evidence of expiratory CT scans of small-airway obstruction in sarcoidosis. AJR Am J Roentgenol 1996;166:1052–1054. 136. Silva CIS, Müller NL, Lynch DA, et al. Chronic hypersensitivity pneumonitis: differentiation from idiopathic pulmonary fibrosis and nonspecific interstitial pneumonia by using thin-section CT. Radiology 2008;246:288–297. 137. Silva CIS, Churg A, Müller NL. Hypersensitivity pneumonitis: spectrum of high-resolution CT and pathologic findings. AJR Am J Roentgenol 2007;188:334–344.

DESIGN SERVICES OF

44

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

138. Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med 2011;183:788–824. 139. Worthy SA, Park CS, Kim JS, et al. Bronchiolitis obliterans after lung transplantation: high resolution CT findings in 15 patients. AJR Am J Roentgenol 1997;169:673–677. 140. Lee ES, Gotway MB, Reddy GP, et al. Early bronchiolitis obliterans following lung transplantation: accuracy of expiratory thinsection CT for diagnosis. Radiology 2000;216:472–477. 141. Siegel MJ, Bhalla S, Gutierrez FR, et al. Post-lung transplantation bronchiolitis obliterans syndrome: usefulness of expiratory thinsection CT for diagnosis. Radiology 2001;220:455–462. 142. Wittram C, Rappaport DC. Bronchiolitis obliterans after lung transplantation: appearance on expiratory minimum intensity projection images. Can Assoc Radiol J 2000;51:103–106. 143. Arakawa H, Niimi H, Kurihara Y, et al. Expiratory high-resolution CT: diagnostic value in diffuse lung diseases. AJR Am J Roentgenol 2000;175:1537–1543. 144. Kalender WA, Rienmuller R, Seissler W, et al. Measurement of pulmonary parenchymal attenuation: use of spirometric gating with quantitative CT. Radiology 1990;175:265–268. 145. Kalender WA, Fichte H, Bautz W, et al. Semiautomatic evaluation procedures for quantitative CT of the lung. J Comput Assist ­Tomogr 1991;15:248–255. 146. Lamers RJ, Thelissen GR, Kessels AG, et al. Chronic obstructive pulmonary diseases. Evaluation with spirometrically controlled CT lung densitometry. Radiology 1994;193:109–113. 147. Robinson TE, Leung AN, Moss RB, et al. Standardized high-­ ­ resolution CT of the lung using a spirometer-triggered electron beam CT scanner. AJR Am J Roentgenol 1999;172:​ 1636–1638. 148. Long FR, Castile RG, Brody AS, et al. Lungs in infants and young children: improved thin-section CT with a noninvasive controlled-­ ventilation technique—initial experience. Radiology 1999;212:​ 588–593. 149. Beinert T, Behr J, Mehnert F, et al. Spirometrically controlled quantitative CT for assessing diffuse parenchymal lung disease. J Comput Assist Tomogr 1995;19:924–931. 150. Kauczor HU, Heussel CP, Fischer B, et al. Assessment of lung volumes using helical CT at inspiration and expiration: comparison with pulmonary function tests. AJR Am J Roentgenol 1998;171:​ 1091–1095. 151. Mergo PJ, Williams WF, Gonzalez-Rothi R, et al. Three-­dimensional volumetric assessment of abnormally low attenuation of the lung from routine helical CT: inspiratory and expiratory quantification. AJR Am J Roentgenol 1998;170:1355–1360. 152. Lee KW, Chung SY, Yang I, et al. Correlation of aging and smoking with air trapping at thin-section CT of the lung in asymptomatic subjects. Radiology 2000;214:831–836. 153. Stern EJ, Webb WR. Dynamic imaging of lung morphology with ultrafast high-resolution computed tomography. J Thorac Imaging 1993;8:273–282. 154. Stern EJ, Webb WR, Warnock ML, et al. Bronchopulmonary sequestration: dynamic, ultrafast, high-resolution CT evidence of air trapping. AJR Am J Roentgenol 1991;157:947–949. 155. Lynch DA, Brasch RC, Hardy KA, et al. Pediatric pulmonary disease: assessment with high-resolution ultrafast CT. Radiology 1990;176:243–248. 156. Gotway MB, Lee ES, Reddy GP, et al. Low-dose, dynamic, expiratory thin-section CT of the lungs using a spiral CT scanner. J ­Thorac Imaging 2000;15:168–172. 157. Lucidarme O, Grenier PA, Cadi M, et al. Evaluation of air trapping at CT: comparison of continuous-versus suspended-expiration CT techniques. Radiology 2000;216:768–772. 158. Camiciottoli G, Bartolucci M, Maluccio NM, et al. Spirometrically gated high-resolution CT findings in COPD: lung attenuation vs lung function and dyspnea severity. Chest 2006;129:558–564. 159. Nishino M, Hatabu H. Volumetric expiratory HRCT imaging with MSCT. J Thorac Imaging 2005;20:176–185.

#152201  Cust: LWW  Au: Webb  Pg. No. 44

160. Nishino M, Hatabu H. Volumetric expiratory high-resolution CT of the lung. Eur J Radiol 2004;52:180–184. 161. Nishino M, Boiselle PM, Copeland JF, et al. Value of volumetric data acquisition in expiratory high-resolution computed tomography of the lung. J Comput Assist Tomogr 2004;28:209–214. 162. Nishino M, Kuroki M, Boiselle PM, et al. Coronal reformations of volumetric expiratory high-resolution CT of the lung. AJR Am J Roentgenol 2004;182:979–982. 163. Wittram C, Batt J, Rappaport DC, et al. Inspiratory and expiratory helical CT of normal adults: comparison of thin section scans and minimum intensity projection images. J Thorac Imaging 2002;17:47–52. 164. Bankier AA, Schaefer-Prokop C, De Maertelaer V, et al. Air trapping: comparison of standard-dose and simulated low-dose thinsection CT techniques. Radiology 2007;242:898–906. 165. Akira M, Toyokawa K, Inoue Y, et al. Quantitative CT in chronic obstructive pulmonary disease: inspiratory and expiratory assessment. AJR Am J Roentgenol 2009;192:267–272. 166. Bankier AA, Gevenois PA, Hackx M, et al. Expert opinion: quantitative computed tomography analysis of chronic obstructive pulmonary disease. J Thorac Imaging 2011;26:248. doi:10.1097/ RTI.0b013e3182343906 167. Bastarrika G, Wisnivesky JP, Pueyo JC, et al. Low-dose volumetric computed tomography for quantification of emphysema in asymptomatic smokers participating in an early lung cancer detection trial. J Thorac Imaging 2009;24:206–211. doi:10.1097/ RTI.0b013e3181a65263 168. Chae EJ, Seo JB, Song J-W, et al. Slope of emphysema index: an objective descriptor of regional heterogeneity of emphysema and an independent determinant of pulmonary function. AJR Am J Roentgenol 2010;194:W248–W255. 169. Coxson HO, Rogers RM. Quantitative computed tomogra phy of chronic obstructive pulmonary disease. Acad Radiol 2005;12:1457–1463. 170. Gierada DS, Guniganti P, Newman BJ, et al. Quantitative CT assessment of emphysema and airways in relation to lung cancer risk. Radiology 2011;261:950–959. 171. Gierada DS, Pilgram TK, Whiting BR, et al. Comparison of ­standard- and low-radiation-dose CT for quantification of emphysema. AJR Am J Roentgenol 2007;188:42–47. 172. Gietema HA, Schilham AM, van Ginneken B, et al. Monitoring of smoking-induced emphysema with CT in a lung cancer screening setting: detection of real increase in extent of emphysema. Radiology 2007;244:890–897. 173. Goldin JG. Imaging the lungs in patients with pulmonary emphysema. J Thorac Imaging 2009;24:163–170. doi:10.1097/ RTI.0b013e3181b41b53 174. Han MK, Kazerooni EA, Lynch DA, et al. Chronic obstructive pulmonary disease exacerbations in the COPD Gene study: associated radiologic phenotypes. Radiology 2011;261:274–282. 175. Irion KL, Marchiori E, Hochhegger B, et al. CT quantification of emphysema in young subjects with no recognizable chest disease. AJR Am J Roentgenol 2009;192:W90–W96. 176. Madani A, Van Muylem A, de Maertelaer V, et al. Pulmonary emphysema: size distribution of emphysematous spaces on multidetector CT images: comparison with macroscopic and microscopic morphometry. Radiology 2008;248:1036–1041. 177. Madani A, Van Muylem A, Gevenois PA. Pulmonary emphysema: effect of lung volume on objective quantification at thin-section CT. Radiology 2010;257:260–268. 178. Matsuoka S, Kurihara Y, Yagihashi K, et al. Quantitative assessment of air trapping in chronic obstructive pulmonary disease using inspiratory and expiratory volumetric MDCT. AJR Am J Roentgenol 2008;190:762–769. 179. Matsuoka S, Yamashiro T, Washko GR, et al. Quantitative CT assessment of chronic obstructive pulmonary disease. Radiographics 2010;30:55–66. 180. Mets OM, Buckens CF, Zanen P, et al. Identification of chronic obstructive pulmonary disease in lung cancer screening computed tomographic scans. JAMA 2011;306:1775–1781.

DESIGN SERVICES OF

C hapter 1 Technical Aspects of High-Resolution CT 181. Mets OM, Murphy K, Zanen P, et al. The relationship between lung function impairment and quantitative computed tomography in chronic obstructive pulmonary disease. Eur Radiol 2012;22:120–128. 182. Montaudon M, Berger P, Cangini-Sacher A, et al. Bronchial measurement with three-dimensional quantitative thin-section CT in patients with cystic fibrosis. Radiology 2007;242:573–581. 183. Montaudon M, Berger P, de Dietrich G, et al. Assessment of airways with three-dimensional quantitative thin-section CT: in vitro and in vivo validation. Radiology 2007;242:563–572. 184. Pilgram TK, Quirk JD, Bierhals AJ, et al. Accuracy of emphysema quantification performed with reduced numbers of CT sections. AJR Am J Roentgenol 2010;194:585–591. 185. Martinez CH, Chen YH, Westgate PM, et al. Relationship between quantitative CT metrics and health status and bode in chronic obstructive pulmonary disease. Thorax 2012;67:399–406. 186. Schroeder JD, McKenzie AS, Zach JA, et al. Relationships between airflow obstruction and quantitative CT measurements of emphysema, air trapping, and airways in subjects with and without chronic obstructive pulmonary disease. AJR Am J Roentgenol 2013;201:W460–W470. 187. Madani A, Zanen J, de Maertelaer V, et al. Pulmonary emphysema: objective quantification at multi-detector row CT—­comparison with macroscopic and microscopic morphometry. Radiology 2006;238:1036–1043. 188. Barr RG, Berkowitz EA, Bigazzi F, et al. A combined pulmonaryradiology workshop for visual evaluation of COPD: study design, chest CT findings and concordance with quantitative evaluation. COPD 2012;9:151–159. 189. Bankier AA, De Maertelaer V, Keyzer C, et al. Pulmonary emphysema: subjective visual grading versus objective quantification with macroscopic morphometry and thin-section CT densitometry. Radiology 1999;211:851–858. 190. Kim SS, Seo JB, Lee HY, et al. Chronic obstructive pulmonary disease: lobe-based visual assessment of volumetric CT by using standard images—comparison with quantitative CT and pulmonary function test in the COPD Gene study. Radiology 2013;266:626–635. 191. Nakano Y, Muro S, Sakai H, et al. Computed tomographic measurements of airway dimensions and emphysema in smokers. Correlation with lung function. Am J Respir Crit Care Med 2000;162:1102–1108. 192. Montaudon M, Lederlin M, Reich S, et al. Bronchial measurements in patients with asthma: comparison of quantitative thinsection CT findings with those in healthy subjects and correlation with pathologic findings. Radiology 2009;253:844–853. 193. Gazourian L, Coronata AM, Rogers AJ, et al. Airway dilation in bronchiolitis obliterans after allogeneic hematopoietic stem cell transplantation. Respir Med 2013;107:276–283. 194. Rosas IO, Yao J, Avila NA, et al. Automated quantification of highresolution CT scan findings in individuals at risk for pulmonary fibrosis. Chest 2011;140:1590–1597. 195. Yoon RG, Seo JB, Kim N, et al. Quantitative assessment of change in regional disease patterns on serial HRCT of fibrotic interstitial pneumonia with texture-based automated quantification system. Eur Radiol 2013;23:692–701. 196. Shin KE, Chung MJ, Jung MP, et al. Quantitative computed tomographic indexes in diffuse interstitial lung disease: correlation with physiologic tests and computed tomography visual scores. J ­Comput Assist Tomogr 2011;35:266–271. 197. Boedeker KL, McNitt-Gray MF, Rogers SR, et al. Emphysema: effect of reconstruction algorithm on CT imaging measures. Radiology 2004;232:295–301. 198. Ashraf H, Lo P, Shaker SB, et al. Short-term effect of changes in smoking behaviour on emphysema quantification by CT. Thorax 2011;66:55–60. 199. Schoepf UJ, Becker CR, Bruening RD, et al. Electrocardiographically gated thin-section CT of the lung. Radiology 1999;212:​ 649–654. 200. Montaudon M, Berger P, Blachere H, et al. Thin-section CT of the lung: influence of 0.5-s gantry rotation and ECG triggering on image quality. Eur Radiol 2001;11:1681–1687.

#152201  Cust: LWW  Au: Webb  Pg. No. 45

45

201. Boehm T, Willmann JK, Hilfiker PR, et al. Thin-section CT of the lung: does electrocardiographic triggering influence diagnosis? Radiology 2003;229:483–491. 202. Arac M, Oner AY, Celik H, et al. Lung at thin-section CT: influence of multiple-segment reconstruction on image quality. Radiology 2003;229:195–199. 203. Ha HI, Goo HW, Seo JB, et al. Effects of high-resolution CT of the lung using partial versus full reconstruction on motion artifacts and image noise. AJR Am J Roentgenol 2006;187:618–622. 204. Remy-Jardin M, Remy J. Comparison of vertical and oblique CT in evaluation of the bronchial tree. J Comput Assist Tomogr 1988;12:956–962. 205. Grenier P, Cordeau MP, Beigelman C. High-resolution computed tomography of the airways. J Thorac Imaging 1993;8:213–229. 206. Swensen SJ, Brown LR, Colby TV, et al. Lung nodule enhancement at CT: prospective findings. Radiology 1996;201:447–455. 207. Ghaye B, Szapiro D, Mastora I, et al. Peripheral pulmonary arteries: how far in the lung does multi-detector row spiral CT allow analysis? Radiology 2001;219:629–636. 208. Ko JP, Brandman S, Stember J, et al. Dual-energy computed tomography: concepts, performance, and thoracic applications. J Thorac Imaging 2012;27:7–22. doi:10.1097/RTI.0b013e31823fe0e9 209. Thieme SF, Johnson TRC, Lee C, et al. Dual-energy CT for the assessment of contrast material distribution in the pulmonary parenchyma. AJR Am J Roentgenol 2009;193:144–149. 210. Kang M-J, Park CM, Lee C-H, et al. Dual-energy CT: clini cal applications in various pulmonary diseases. Radiographics 2010;30:685–698. 211. Chae EJ, Seo JB, Goo HW, et al. Xenon ventilation CT with a dualenergy technique of dual-source CT: initial experience. Radiology 2008;248:615–624. 212. Ringl H, Schernthaner RE, Bankier AA, et al. JPEG2000 compression of thin-section CT images of the lung: effect of compression ratio on image quality. Radiology 2006;240:869–877. 213. Bankier AA, Fleischmann D, Mallek R, et al. Bronchial wall thickness: appropriate window settings for thin-section CT and ­radiologic-anatomic correlation. Radiology 1996;199:831–836. 214. Primack SL, Remy-Jardin M, Remy J, et al. High-resolution CT of the lung: pitfalls in the diagnosis of infiltrative lung disease. AJR Am J Roentgenol 1996;167:413–418. 215. Maguire WM, Herman PG, Khan A, et al. Comparison of fixed and adjustable window width and level settings in the CT evaluation of diffuse lung disease. J Comput Assist Tomogr 1993;17:847–852. 216. Remy-Jardin M, Remy J, Giraud F, et al. Computed tomography assessment of ground-glass opacity: semiology and significance. J Thorac Imaging 1993;8:249–264. 217. Grenier P, Maurice F, Musset D, et al. Bronchiectasis: assessment by thin-section CT. Radiology 1986;161:95–99. 218. Kazerooni EA, Chow LC, Whyte RI, et al. Preoperative examination of lung transplant candidates: value of chest CT compared with chest radiography. AJR Am J Roentgenol 1995;165:1343–1348. 219. Chung MH, Edinburgh KJ, Webb EM, et al. Mixed infiltrative and obstructive disease on high-resolution CT: differential diagnosis and functional correlates in a consecutive series. J Thorac Imaging 2001;16:69–75. 220. Naidich DP, Funt S, Ettenger NA, et al. ­Hemoptysis: CT-­bronchoscopic correlations in 58 cases. Radiology 1990;177:357–362. 221. Set PA, Flower CD, Smith IE, et al. Hemoptysis: comparative study of the role of CT and fiberoptic bronchoscopy. Radiology 1993;189:677–680. 222. Bergin CJ, Rios G, King MA, et al. Accuracy of high-resolution CT in identifying chronic pulmonary thromboembolic disease. AJR Am J Roentgenol 1996;166:1371–1377. 223. King MA, Ysrael M, Bergin CJ. Chronic thromboembolic pulmonary hypertension: CT findings. AJR Am J Roentgenol 1998;170:955–960. 224. Lee KN, Lee HJ, Shin WW, et al. Hypoxemia and liver cirrhosis (hepatopulmonary syndrome) in eight patients: comparison of the central and peripheral pulmonary vasculature. Radiology 1999;211:549–553.

DESIGN SERVICES OF

46

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

225. Sherrick AD, Swensen SJ, Hartman TE. Mosaic pattern of lung attenuation on CT scans: frequency among patients with pulmonary artery hypertension of different causes. AJR Am J Roentgenol 1997;169:79–82. 226. Connolly B, Manson D, Eberhard A, et al. CT appearance of pulmonary vasculitis in children. AJR Am J Roentgenol 1996;167:​901–904. 227. Drucker EA, Rivitz SM, Shepard JA, et al. Acute pulmonary embolism: assessment of helical CT for diagnosis. Radiology 1998;209:235–241. 228. Mayo JR, Remy-Jardin M, Müller NL, et al. Pulmonary embolism: prospective comparison of spiral CT with ventilation-perfusion scintigraphy. Radiology 1997;205:447–452. 229. Remy-Jardin M, Remy J, Artaud D, et al. Spiral CT of pulmonary embolism: technical considerations and interpretive pitfalls. J Thorac Imaging 1997;12:103–117. 230. Rubin GD, Goodman LR, Lipchik RJ, et al. Helical CT for the detection of acute pulmonary embolism: experts debate. J Thorac Imaging 1997;12:81–102. 231. Belden CJ, Weg N, Minor LB, et al. CT evaluation of bone dehiscence of the superior semicircular canal as a cause of

#152201  Cust: LWW  Au: Webb  Pg. No. 46

sound- and/or pressure-induced vertigo. Radiology 2003;226:​ 337–343. 232. Jaffe TA, Nelson RC, Johnson GA, et al. Optimization of multiplanar reformations from isotropic data sets acquired with 16-­detector row helical CT scanner. Radiology 2006;238:292–299. 233. Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased isolated lungs: high-resolution CT. Radiology 1988;166:81–87. 234. Murata K, Itoh H, Todo G, et al. Centrilobular lesions of the lung: demonstration by high-resolution CT and pathologic correlation. Radiology 1986;161:641–645. 235. Takahashi K, Thompson B, Stanford W, et al. Visualization of normal pulmonary fissures on sagittal multiplanar reconstruction MDCT. AJR Am J Roentgenol 2006;187:389–397. 236. Mayo JR, Müller NL, Henkelman RM. The double-fissure sign: a motion artifact on thin-section CT scans. Radiology 1987;165:580–581. 237. Tarver RD, Conces DJ, Godwin JD. Motion artifacts on CT simulate bronchiectasis. AJR Am J Roentgenol 1988;151:1117–1119. 238. Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. ­Radiographics 2004;24:1679–1691.

DESIGN SERVICES OF

2

Normal Lung Anatomy I M P O R T A N T

T O P I C S

THE LUNG INTERSTITIUM  47

SUBPLEURAL INTERSTITIUM AND PLEURAL SURFACES 60

LARGE BRONCHI AND ARTERIES  48

NORMAL LUNG ATTENUATION  61

THE SECONDARY PULMONARY LOBULE AND LUNG ACINUS  51

NORMAL EXPIRATORY HIGH-RESOLUTION COMPUTED TOMOGRAPHY  62

ANATOMY OF THE SECONDARY LOBULE AND ITS COMPONENTS  54 THE CONCEPT OF CORTICAL AND MEDULLARY LUNG  59

Abbreviations Used in This Chapter T/D B/A BI HU LMB MinIP RMB SD

bronchial wall thickness/diameter bronchoarterial bronchus intermedius Hounsfield units left main bronchus minimum-intensity projection right main bronchus standard deviation

The accurate interpretation of high-resolution computed tomography (HRCT) images requires a detailed understanding of normal lung anatomy, normal variations, normal findings that may mimic an abnormality (1,2) and the pathologic alterations in normal lung anatomy occurring in the presence of disease (1–7). In this chapter, those aspects of lung anatomy that are important in using and interpreting HRCT are reviewed.

THE LUNG INTERSTITIUM The lung is supported by a network of connective tissue fibers called the lung interstitium. Although the lung interstitium is not generally visible on HRCT in normal patients, interstitial thickening is often recognizable. For the purpose of interpretation of HRCT and identification of abnormal findings, the interstitium can be thought of as having several components (Fig. 2-1) (8). The peribronchovascular interstitium is a system of fibers that invests bronchi and pulmonary arteries (Fig.  2-1). In the perihilar regions, the peribronchovascular interstitium forms a strong connective tissue sheath

that surrounds large bronchi and arteries (9). The more peripheral continuum of this interstitial fiber system, which is associated with small centrilobular bronchioles and arteries, may be termed the centrilobular interstitium (Fig. 2-1). Taken together, the peribronchovascular interstitium and centrilobular interstitium correspond to the “axial fiber system” described by Weibel, which extends peripherally from the pulmonary hila to the level of the alveolar ducts and sacs (8). The subpleural interstitium is located beneath the visceral pleura; it envelops the lung in a fibrous sac from which connective tissue septa penetrate into the lung parenchyma (Fig. 2-1). These septa include the interlobular septa, which are described in detail later in

Intralobular interstitium

Centrilobular interstitium

Interlobular septa Peribronchovascular interstitium Subpleural interstitium

Secondary pulmonary lobule FIGU RE 2-1  Components of the lung interstitium. Taken together, the peribronchovascular interstitium and centrilobular interstitium correspond to the “axial fiber system” described by Weibel (8). The subpleural interstitium and interlobular septa correspond to Weibel’s “peripheral fiber system.” The intralobular interstitium is roughly equivalent to the “septal fibers” described by Weibel.

47

#152201  Cust: LWW  Au: Webb  Pg. No. 47

DESIGN SERVICES OF

48

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

this chapter. The subpleural interstitium and interlobular septa are parts of the “peripheral fiber system” described by Weibel (8). The intralobular interstitium is a network of thin fibers that forms a fine connective tissue mesh in the walls of alveoli and thus bridges the gap between the centrilobular interstitium in the center of lobules and the interlobular septa and subpleural interstitium in the lobular periphery (Fig. 2-1). Together, the intralobular interstitium, peribronchovascular interstitium, centrilobular interstitium, subpleural interstitium, and interlobular septa form a continuous fiber skeleton for the lung (Fig. 2-1). The intralobular interstitium corresponds to the “septal fibers” described by Weibel (8).

LARGE BRONCHI AND ARTERIES Within the lung parenchyma, bronchi and pulmonary arteries are closely associated and branch in parallel. When imaged at an angle to their longitudinal axis, central pulmonary arteries normally appear as rounded or elliptical opacities on HRCT, accompanied by uniformly thinwalled bronchi of similar size and shape (Figs. 2-2 and 2-3). When imaged along their axis, bronchi and vessels should appear roughly cylindrical or show slight tapering as they branch, depending on the length of the segment that is visible; tapering of a vessel or bronchus is most easily seen when a long segment is visible. The outer walls of pulmonary arteries form a smooth and sharply defined interface with the surrounding lung,

whether they are seen in cross section or along their length. The walls of large bronchi, outlined by lung on one side and air in the bronchial lumen on the other, should appear smooth and of uniform thickness. As indicated in the previous section, bronchi and arteries are encased by  ­the  peribronchovascular interstitium, which extends from the pulmonary hila into the peripheral lung. Thickening of the peribronchial and perivascular interstitium can result in irregularity of the interface between arteries and bronchi and the adjacent lung (6,9,10).

Bronchial Diameter and Bronchoarterial Ratio In most normal subjects, the bronchi and adjacent pulmonary arteries are similar in diameter. Their diameters may be compared by using the so-called bronchoarterial (B/A) ratio, defined as the internal diameter (i.e., luminal diameter) of the bronchus divided by the diameter of the adjacent pulmonary artery. To avoid an exaggeration of diameters caused by obliquity of the bronchus and artery relative to the scan plane, the least diameter of the bronchus and artery are used for measurement. The B/A ratio in normal subjects generally averages 0.65 to 0.70 (Figs. 2-4 and 2-5) (11,12). Three-dimensional reconstruction has also been used to accurately measure bronchial lumen area and diameter perpendicular to the bronchial axis (13). A B/A ratio greater than 1 is usually believed to indicate bronchiectasis, although this may be seen in some normal subjects. For example, in an HRCT evaluation of

B

A

FIGU RE 2-2  Axial HRCT appearances of large bronchi and arteries in the upper (A) and lower (B) lobes of a normal subject, imaged with window settings of –700/1,000 HU. The diameters of vessels and their neighboring bronchi are approximately equal. The outer walls of bronchi and pulmonary vessels are smooth and sharply defined. Bronchi are usually invisible within the peripheral 2 cm of lung, despite the fact that vessels are well seen in this region.

#152201  Cust: LWW  Au: Webb  Pg. No. 48

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

49

FIGU RE 2-5  B/A ratio in a normal subject. A detailed view of the right lower lobe (same subject as shown in Fig. 2-2) shows a bronchus (arrow) having an internal diameter less than the diameter of the accompanying pulmonary artery. Other visible bronchi show a similar B/A ratio.

FIGU RE 2-3  Normal appearances of large bronchi and arteries. In an isolated inflated lung, the smallest bronchi visible (arrows) measure 2 to 3 mm in diameter. Bronchi and bronchioles are not visible within the peripheral 1 cm of lung in this preparation, although the artery branches in the peripheral lung are sharply seen. Note: The “isolated” lungs illustrated in this book are fresh lungs obtained at autopsy and scanned while inflated with air at a pressure of approximately 30 cm of water. (From Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased isolated lungs: high-resolution CT. Radiology 1988;166(1, pt 1):81–87, with permission.)

14 normal subjects (12), although the B/A ratio averaged 0.65 ± 0.16, 7% of scan interpretations were believed to show some evidence of bronchial dilatation. The presence of a B/A ratio greater than 1 in normal subjects has been associated with increasing age. In a study by Matsuoka et al. (11), B/A ratios were measured at the segmental and subsegmental levels in the apical and posterior basal segments in 85 normal subjects. A significant correlation was found between the B/A ratio and age (r = 0.768, p < 0.0001). When the subjects were considered in three age groups, the mean B/A ratios were 0.609 ± 0.05 in subjects 21 to 40 years old, 0.699 ± 0.067 in subjects 41 to 64 years old, and 0.782 ± 0.078

FIGU RE 2-4  B/A ratio. The B/A ratio is calculated by dividing the internal diameter (i.e., luminal diameter) of the bronchus (B) by the diameter of the adjacent pulmonary artery (A). It averages 0.65 to 0.70 in normal subjects.

#152201  Cust: LWW  Au: Webb  Pg. No. 49

in subjects 65 years and older (p < 0.0001). At least one bronchus with a B/A ratio greater than 1 was seen in 41% of p ­ atients older than 65 years, and in this group, 19% of measured bronchi had a B/A ratio greater than 1 (Fig. 2-6). Seven percent of subjects 41 to 64 years of age had at least one bronchus (5% of measured bronchi) with a B/A ratio greater than 1. None of the subjects 21 to 40 years of age showed this finding. In a study by Copley et al. (2), prone HRCT scans in 40 subjects over 75 years of age were compared to scans obtained in a group less than 55 years of age; none had known pulmonary disease. Bronchial dilatation was seen more frequently (p < 0.001) in the older group (60%) than in the younger group (6%). This finding was independent of smoking history.

FIGU RE 2-6  Increased B/A ratio in a normal 78-year-old man. The internal diameter of a bronchus (arrow) in the right lower lobe exceeds the diameter of the adjacent pulmonary artery. The B/A ratio increases with age and may exceed 1 in normal patients older than 40 years.

DESIGN SERVICES OF

50

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

An increase in B/A ratio may also be seen in normal subjects living at high altitude (14–16). It has been suggested that this results from mild hypoxemia resulting in bronchial dilatation and vasoconstriction. Kim et al. (14) found that 9 of 17 (53%) normal subjects living at an altitude of 1,600 m had evidence of at least one bronchus equal to or greater in size than the adjacent pulmonary artery; these authors found that only 2 of 16 (12.5%) individuals living at sea level showed a similar finding. In this study, the mean B/A ratio was 0.76 at an altitude of 1,600 m. Similarly, Lynch et al. (16) compared the internal diameters of lobar, segmental, subsegmental, and smaller bronchi to those of adjacent pulmonary artery branches in 27 normal subjects living in Denver, Colorado, at an a­ ltitude of about 1 mile. The authors found that 37 of 142 (26%) bronchi evaluated, and 59% of individuals, had increased B/A ratios. A convincing relationship has not been shown between the B/A ratio and the location of the bronchi being evaluated. Evaluation of the distribution bronchi with a B/A ratio greater than 1 has generally failed to reveal any relationship to lobe or anteroposterior location within the lungs (14,17). The B/A ratio may also appear to be greater than 1 if the scan traverses an undivided bronchus near its branch point and its accompanying artery has already branched. In this situation, two artery branches may be seen to lie adjacent to the “dilated” bronchus.

Bronchial Wall Thickness The thickness of a normal airway wall is related to its diameter. According to Weibel, 2nd- to 4th-generation (lobar to segmental) bronchi have a wall thickness of approximately 1.5 mm and a mean diameter between 5 and 8 mm (i.e., the bronchial wall thickness is about 20%–30% of the bronchial diameter); 6th- to 8th-generation (subsegmental) airways have a wall thickness of approximately 0.3 mm and mean diameters between 1.5 and 3 mm (i.e., the airway wall is 10%–20% of its diameter); and 11thto 13th-generation airways have diameters measuring 0.7 to 1 mm with walls of 0.1 to 0.15 mm (i.e., the airway wall is about 15% of its diameter) (Table 2-1) (18,19).

FIGU RE 2-7  Measurement of bronchial wall thickness using the T/D ratio. This ratio is defined as wall thickness (T) divided by the total diameter of the bronchus (D). In normal subjects, it averages about 0.2 or 20%.

The relationship between bronchial wall thickness and diameter may be expressed by using the T/D ratio, defined as wall thickness (T) divided by the total diameter of the bronchus (D) (Fig. 2-7). This ratio may be measured using HRCT and averages about 20% for segmental and subsegmental bronchi (Fig. 2-8), quite similar to the anatomical measurements described previously. In one study (11), HRCT was performed in 85 subjects without cardiopulmonary disease. The T/D ratio was measured at the segmental and subsegmental levels of the apical and posterior basal segments. The images were viewed at a window level of –450 H and a window width of 1,500 H, believed best for accurate bronchial measurements (20,21). Overall, the T/D ratio measured 0.200 ± 0.015 (range, 0.171–0.227). In another study, the T/D ratio measured 0.23 (± 0.04) in 14 normal subjects (22). In a study by Matsuoka et al. (11), no significant correlation was found between the T/D ratio and age. However, in a study by Copley et al. (2) of HRCT findings in normal elderly patients, bronchial wall thickening was seen more frequently (55%; p < 0.001) in a group of

Table 2-1  Relation of Airway Diameter to Wall Thickness Airway

Diameter (mm)

Wall thickness (mm)

5–8

1.5

1.5–3.0

0.2–0.3

Lobular bronchiole

1

0.15

Terminal bronchiole

0.7

0.1

Acinar bronchiole

0.5

0.05

Lobular and segmental bronchi Subsegmental bronchi/ bronchiole

Modified from Weibel ER. High resolution computed tomography of the pulmonary parenchyma: anatomical background. Paper presented at Fleischner Society symposium on chest disease; Scottsdale, AZ; 1990.

#152201  Cust: LWW  Au: Webb  Pg. No. 50

FIGU RE 2-8  T/D ratio in a normal subject. A detailed view of the right lower lobe (same subject as shown in Figs. 2-2 and 2-5). The wall thickness of the bronchus indicated by the arrow appears to be about one-fifth of its diameter.

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

subjects older than 75 years than in a group of patients less than 55 years of age (6%). The smallest airways normally visible using HRCT have a diameter of approximately 2 mm and a wall thickness of 0.2 to 0.3 mm (23). In normal subjects, airways in the peripheral 2 cm of lung are uncommonly seen because their walls are too thin (Fig. 2-3) (24). It has also been reported that airways in the peripheral 1 cm of lung are rarely seen, except adjacent to the mediastinum (15,25). In a study by Kim et al., airways were visible within 1 cm of the mediastinal pleura in 40% of normal subjects (25). It is important to keep in mind that the window settings used for the interpretation of HRCT may have a significant effect on the apparent thickness of bronchial walls (see Chapter 1). Furthermore, the bronchial wall as shown on HRCT represents not only the wall itself, but also the surrounding peribronchovascular interstitium. Thickening of this interstitium may result in what appears to be bronchial wall thickening; the term “peribronchial cuffing” has been used to describe this occurrence on plain radiographs.

THE SECONDARY PULMONARY LOBULE AND LUNG ACINUS The lung comprises numerous anatomical units smaller than a lobe or segment. The secondary pulmonary lobule and lung acinus are widely regarded to be the most important of these subsegmental lung units (26). The secondary pulmonary lobule, as defined by Miller, refers to the smallest unit of lung structure marginated by connective tissue septa (19,27) (Figs. 2-1, 2-9, and 2-10). Secondary pulmonary lobules are irregularly polyhedral in shape and vary in size, measuring from 1 to 2.5 cm in diameter in most locations (8,19,28–30). In one study, the average diameter of pulmonary lobules measured in several adults ranged from 11 to 17 mm (30). Airways, pulmonary arteries and veins, lymphatics, and the various components of the pulmonary interstitium are all represented at the level of the pulmonary lobule (Figs. 2-1, 2-9, and 2-10). Each secondary lobule is supplied by a small bronchiole and pulmonary artery branch, and is variably marginated in different lung regions by connective tissue interlobular septa containing pulmonary veins and lymphatics (31). Secondary lobular anatomy is easily visible on the surface of the lung because of these interlobular septa (19,28). The pulmonary acinus is smaller than a secondary lobule. It is defined as the portion of lung distal to a terminal bronchiole (the last purely conducting airway) and supplied by a first-order respiratory bronchiole or bronchioles (32,33). Because respiratory bronchioles are the largest airways that have alveoli in their walls, an acinus is the largest lung unit in which all airways participate in gas exchange. Acini are usually described as ranging from 6 to 10 mm in diameter and average 7 to 8 mm in adults (Figs. 2-10A and 2-11) (30,34).

#152201  Cust: LWW  Au: Webb  Pg. No. 51

A

51

Pulmonary lobules

Bronchioles and pulmonary arteries

Pulmonary veins

Interlobular septa

B FIGU RE 2-9  Pulmonary lobular anatomy. A and B: Pulmonary lobules that are irregularly polyhedral or conical in shape are often visible on the surface of the lung, as shown in this diagram of five lobules visible on the posterior surface of the right lung. B: Lobules are supplied by small bronchiolar and pulmonary artery branches, which are central in location. They are variably marginated by connective tissue interlobular septa that contain pulmonary vein and lymphatic branches. (Specimen photograph courtesy of Martha Warnock, MD.)

Secondary pulmonary lobules usually comprise a dozen or fewer acini, although the number varies considerably in different reports (18,35). In a study by Itoh et al. (36), the number of acini counted in lobules of varying sizes ranged from 3 to 24.

Historical Considerations Concepts regarding the importance of the secondary pulmonary lobule, acinus, and smaller lung units have evolved during the past 300 years in conjunction with continued progress in understanding lung anatomy, pathology, and physiology. An excellent perspective on the

DESIGN SERVICES OF

52

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

Visceral pleura Acinus

Interlobular septa

Bronchioles

Arteries

A

Pulmonary vein

1 cm

B FIGU RE 2-10  A: Anatomy of the secondary pulmonary lobule, as defined

by Miller. Two adjacent lobules are shown in this diagram. B: Radiographic anatomy of the secondary pulmonary lobule. Radiograph of a 1-mm lung slice taken from the lower lobe. Two well-defined secondary pulmonary lobules are visible. Lobules are marginated by thin interlobular septa (S) containing pulmonary vein (V) branches. Bronchioles (B) and pulmonary arteries (A) are centrilobular. Bar = 1 cm. (Reprinted from Itoh H, Murata K, Konishi J, et al. Diffuse lung disease: pathologic basis for the high-resolution computed tomography findings. J Thorac Imaging 1993;8:176, with permission.)

sequence of events and incremental discoveries made during this period has been provided by Miller (37). The earliest detailed description (1676) of the secondary pulmonary lobule was provided by Thomas Willis, who studied lung structure by the injection of mercury and other fluids into the bronchi and pulmonary vessels. He found that “little lobes” (i.e., the lobules) arose from small branches of the trachea and were separated from each other by a “membrane” (Fig. 2-12). Bronchioles

#152201  Cust: LWW  Au: Webb  Pg. No. 52

entering the little lobes were described as dividing into a large number of fine branches, which led to minute “bladders” or “vesicles.” Georg Rindfleisch (1875) first used the term acinus to indicate a sublobular lung unit. He described the secondary lobule as supplied by a bronchiole, which divided into progressively smaller bronchiolar branches, finally giving rise to arborizing “alveolengänge” (alveolar passages), which collectively formed a “lung acinus” (Fig. 2-13). The acinus, according to Rindfleisch, was a much more consistent unit of lung structure than the secondary lobule because of variation in the size of lobules. However, he regarded the secondary lobule to be more important than the acinus pathologically, in that disease processes tended to be limited by the connective tissue septa marginating the lobules. In 1881, Rudolph Kolliker, using the lung of an ­executed criminal, provided a more detailed analysis of the finer divisions of the bronchial tree, and described ­respiratory bronchioles as airways having both bronchiolar epithelium and alveoli in their walls. He distinguished respiratory bronchioles from proximal airways not having alveoli in their walls (i.e., terminal bronchioles) (Fig. 2-14) and distal airways with numerous alveoli in their walls (i.e., alveolengange, subsequently termed alveolar ducts), thus providing the basis for defining the lung acinus relative to airway anatomy. In his 1947 book, The Lung, William Snow Miller reviewed lung anatomy in detail (38). His definitions of the secondary pulmonary lobule and acinus are still in use today (see previous definitions). However, he also considered the primary pulmonary lobule to be a fundamental unit of lung structure. He defined the primary pulmonary lobule as all alveolar ducts, alveolar sacs, and alveoli distal to the last respiratory bronchiole, along with their associated blood vessels, nerves, and connective tissues (Figs. 2-15 and 2-16). However, because the term “primary pulmonary lobule” is not in common use today, “secondary pulmonary lobule,” “secondary lobule,” and “lobule” are often used interchangeably; generally, they should be considered as synonymous (26). In this book, each of these terms is used to refer to Miller’s secondary pulmonary lobule. In 1958, Reid suggested an alternate definition of the pulmonary lobule, based on the branching pattern of peripheral bronchioles identified bronchographically rather than by the presence and location of connective tissue septa (31,32,35). On bronchograms, small bronchioles can be seen to arise at intervals of 5 to 10 mm from larger airways; these small bronchioles show branching at approximately 2-mm intervals, the so-called millimeter pattern (32). Airways showing the millimeter pattern were considered by Reid to be intralobular, with each branch corresponding to a terminal bronchiole (35). She considered lobules to be the lung units supplied by three to five “millimeter pattern” bronchioles. Although Reid’s criteria delineate lung units of approximately equal size, about 1 cm in diameter and containing three to five acini,

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

53

Acinar artery and bronchioles Diameter 0.5 mm Bronchiolar wall thickness 0.05 – 0.1 mm Interlobular septa Thickness 0.1 mm

Acinus 0.6–1 cm

Visceral pleura Thickness 0.1 mm

Lobular bronchiole Diameter 1 mm Wall thickness 0.15 mm

Lobular artery Diameter 1 mm

A

Pulmonary vein Diameter 0.5 mm

1 cm Acinar artery sometimes visible 3–5 mm from septa or the pleura

Centrilobular arteries visible 5–10 mm from septa or the pleura Interlobular septa sometimes visible

Veins visible 1–2 cm from the pleura

B

Airway 3 mm in diameter visible 3 cm from the pleura

FIGU RE 2-11  Dimensions of secondary lobular structures (A) and their visibility on HRCT (B).

it should be understood that this definition does not necessarily describe lung units equivalent to the secondary lobules as defined by Miller and marginated by interlobular septa (Fig. 2-17) (35,36). Miller’s definition is most applicable to the interpretation of HRCT, and is widely

#152201  Cust: LWW  Au: Webb  Pg. No. 53

accepted by both anatomists and pathologists because interlobular septa are visible on histologic sections (36). A foundation for our current understanding of secondary lobular anatomy and its significance in radiographic interpretation was provided by Heitzman et al. in

DESIGN SERVICES OF

54

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

FIGU RE 2-14  Peripheral airway anatomy described by Kolliker (1881). The airway at the lower right (indicated as b) represents a terminal bronchiole. More peripheral airways (i.e., br and br.r) represent respiratory bronchioles. Alveolar ducts (alveolengange) are indicated as ag. The airways peripheral to the terminal bronchiole are acinar.

FIGU RE 2-12  Pulmonary lobules according to Willis (1676). Individual lobules arise from small bronchial branches.

two papers published in 1969 (28,31) and subsequently detailed in Heitzman’s book The Lung: RadiologicPathologic Correlations (39). He described the plain radiographic appearances of various lobular abnormalities, carefully correlated with inflated and fixed lung specimens. In his initial papers, he described the appearance of septal thickening associated with fibrosis or lymphatic and pulmonary venous abnormalities, and panlobular consolidation in pulmonary infarction and bronchopneumonia. In his more detailed descriptions published later, he further emphasized the radiographic appearances of “lobular core” structures, and demonstrated the radiographic and pathologic findings of peribronchiolar nodules and sublobular opacities, which are now generally referred to using the term “centrilobular.”

ANATOMY OF THE SECONDARY LOBULE AND ITS COMPONENTS An understanding of secondary lobular anatomy and the appearances of lobular structures is key to the ­interpretation of HRCT (Figs. 2-10, 2-11, and 2-18 to 2-24). HRCT can show many features of the secondary pulmonary lobule in both normal and abnormal lungs, and many lung diseases, particularly interstitial diseases, produce characteristic changes in lobular structures (5,6,10,23,24,26,40,41). For the purposes of the interpretation of HRCT, the secondary lobule is most appropriately conceptualized as having three principal parts or components: FIGU RE 2-13  A pulmonary lobule and acinus as shown by Rindfleisch (1875). The lobule is supplied by a bronchiole, which divides into smaller branches. The acinus (arrows) shows alveolengange (i.e., alveolar ducts).

#152201  Cust: LWW  Au: Webb  Pg. No. 54

1. Interlobular septa and contiguous interstitium 2. Centrilobular structures 3. Lobular parenchyma and acini

DESIGN SERVICES OF

subpleural

C hapter 2 Normal Lung Anatomy

A

55

B

FIGU RE 2-15  The secondary pulmonary lobule. A: This diagram shows a secondary pulmonary lobule in the peripheral lung, surrounded by connective tissue septa containing pulmonary vein branches. Lobular arteries are shown in black, whereas airways are shown in outline. The lobular bronchiole is traced to alveoli in the lung periphery. The red arrow shows a terminal bronchiole; the blue arrows show respiratory bronchioles with alveoli arising from their walls. B: The approximate size of an acinus is shown in blue, whereas a primary pulmonary lobule is in purple. (Adapted from Miller WS. The lung. Springfield, IL: Charles C Thomas; 1947:204; arrows and shading have been added to the original illustration.)

Interlobular Septa Anatomically, secondary lobules are marginated by connective tissue interlobular septa, which extend inward from the pleural surface (Figs. 2-10, 2-11, and 2-18 to 2-24). The interlobular septa are part of the peripheral interstitial fiber system described by Weibel (Fig. 2-1) (8), and contain pulmonary veins and lymphatics.

Pulmonary lobules in the lung periphery are relatively large in size, and are marginated by interlobular septa that are thicker and better defined than in other parts of the lung (31,42). Peripheral lobules tend to be relatively uniform in appearance, often appearing cuboidal or pyramidal in shape (31). Pulmonary lobules in the central lung are smaller and more irregular in shape than in the peripheral

Pleural surface

Alveolar duct

Miller’s lobules marginated by interlobular septa

Reid’s lobule with 3–5 acini

Respiratory bronchiole

Artery

Vein

FIGU RE 2-16  The primary pulmonary lobule. Miller defined the primary pulmonary lobule as all alveolar ducts, atria (A), alveolar sacs (SAL), and alveoli (A) distal to the last respiratory bronchiole, along with their associated blood vessels, nerves, and connective tissues. (From Miller WS. The lung. Springfield, IL: Charles C Thomas; 1947:75.)

#152201  Cust: LWW  Au: Webb  Pg. No. 55

FIGU RE 2-17  Relative size and relationships of “Miller’s lobule” and “Reid’s lobule.”

DESIGN SERVICES OF

56

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A FIGU RE 2-19  Interlobular septa in continuity in an isolated lung. On HRCT, long interlobular septa (arrows) can be seen marginating several secondary lobules. The septa in this lung are slightly thickened by fluid. Septa are well seen peripherally, but note that the septa and, therefore, secondary lobules are less well defined in the central lung. (From Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased isolated lungs: high-resolution CT. Radiology 1988;166:81, with permission.)

B F I G U R E 2-18  Interlobular septa in an isolated lung. A: Some thin, normal interlobular septa (small arrows) are faintly visible in the peripheral lung. Interlobular septa along the mediastinal pleural surface (large arrows) are slightly thickened by edema fluid and are more easily seen. Note that a very thin line is visible at the pleural surfaces and in the lung fissure, similar in appearance and thickness to the normal interlobular septa. This line represents the subpleural interstitial compartment and visceral pleura. (From Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased isolated lungs: high-resolution CT. Radiology 1988;166:81, with permission.) B: Paper-mounted lung slice at the same level as A. Lobular lung anatomy in the upper lobe is well seen because of pigment deposition in relation to interlobular septa. The same pulmonary lobule (black arrows) as shown in A is visible on the surface of the lung. The branching centrilobular bronchiole (white arrow) is visible.

lung, and are marginated by interlobular septa that are thinner and less well defined. When visible, lobules in the central lung may appear hexagonal or polygonal in shape. It should be kept in mind, however, that the size, shape, and appearance of pulmonary lobules as seen on HRCT are significantly affected by their orientation relative to the scan plane.

#152201  Cust: LWW  Au: Webb  Pg. No. 56

Interlobular septa are thickest and most numerous in the apical, anterior, and lateral aspects of the upper lobes, the anterior and lateral aspects of the middle lobe and lingula, the anterior and diaphragmatic surfaces of the lower lobes, and along the mediastinal pleural surfaces  (43); thus, secondary lobules are best defined in these regions (Fig. 2-24B). Septa measure about 100 μm (0.1 mm) in thickness in a subpleural location (5,8,23,24). As discussed in Chapter 1, the visibility of normal lobular structures on HRCT is related to their size and orientation relative to the plane of scan, although size is most important (Fig. 2-11). Generally, the smallest structures visible on HRCT range from 0.3 to 0.5 mm in thickness, although thinner structures, measuring 0.1 to 0.2 mm, are occasionally seen. Thus, interlobular septa in the peripheral lung are at the lower limit of HRCT resolution (23), but nonetheless they are often visible on HRCT scans performed in vitro (24). With in vitro HRCT, interlobular septa are often visible as very thin, straight lines of uniform thickness that are usually 1 to 2.5 cm in length and perpendicular to the pleural surface (Figs. 2-11 and 2-18). Several septa in continuity can be seen as a linear opacity measuring up to 4 cm in length (Fig. 2-19) (24). On clinical scans in normal patients, interlobular septa are less commonly seen and are seen less well than they are in studies of isolated lungs. A few septa are often visible in the lung periphery in normal subjects, but they tend to be inconspicuous (Figs. 2-20 and 2-21); normal septa are seen most often in areas where they are best developed (i.e., in the apices, anteriorly, and along the mediastinal pleural surfaces) (6,44). When visible, they are usually seen extending to the pleural surface. In the central lung, septa are thinner than they are peripherally and are infrequently seen in normal subjects (Fig. 2-19); often, interlobular septa that are clearly defined in this region are abnormally thickened

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

A

57

B

FIGU RE 2-20  Normal HRCT lobular anatomy. A: Axial HRCT of the left upper lobe in a normal subject. Thin interlobular septa (black arrows) are visible in the posterior lung, outlining a normal pulmonary lobule, but otherwise septa are inconspicuous. The centrilobular artery (white arrow) is clearly seen. B: In the same patient, a scan through the left lower lobe shows normal pulmonary vein branches (black arrows) marginating pulmonary lobules. The centrilobular artery branches (white arrows) are visible as a rounded dot between the veins.

(Fig. 2-22). Occasionally, when interlobular septa are not clearly visible, their locations can be inferred by locating septal pulmonary vein branches, approximately 0.5 mm in diameter. Veins can sometimes be seen as linear, arcuate, or branching structures (Figs. 2-20B and 2-21), or as a row or chain of dots surrounding centrilobular arteries and approximately 5 to 10 mm from them. Pulmonary veins may also be identified by their pattern of branching; it is common for small veins to arise at nearly right angles to a much larger main branch (Fig. 2-20B). Interlobular septa are more frequently visible in elderly subjects than in young patients, in smokers, and tend to

FIGURE 2-21  Normal HRCT lobular anatomy. HRCT of the right upper lobe in a normal subject. Thin interlobular septa identify a lobule (black arrows) in the lung periphery. A pulmonary vein branch (large white arrows) is visible in relation to the periphery of the lobule. The centrilobular artery is also seen as a white dot (small white arrow). Other pulmonary artery branches are visible 5 to 10 mm from the pleural surface.

#152201  Cust: LWW  Au: Webb  Pg. No. 57

increase in visibility over time (1,2,45). In a study by Copley et  al. (2), prone HRCT obtained in 40 subjects over 75  years of age was compared to HRCT obtained in a younger age group (less than 55 years of age); none had known pulmonary disease. Interlobular septal thickening was seen in 18% of the older group and none of

FIGU RE 2-22  Normal lobular anatomy in the right upper lobe. Interlobular septa are more evident than in most patients, but likely normal. Centrilobular arteries (red arrows) are visible as branching or dotlike opacities about 1 cm from the pleural surface. Veins (blue arrows) are seen in relation to interlobular septa.

DESIGN SERVICES OF

58

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A A

B FIGU RE 2-23  Centrilobular anatomy in an isolated lung. A: On a CT scan obtained with 1-cm collimation, pulmonary artery branches (arrows) with their accompanying bronchi can be identified. B: On an HRCT scan at the same level, interlobular septa can be seen marginating one or more lobules. Pulmonary artery branches (arrows) can be seen extending into the centers of pulmonary lobules, but intralobular bronchioles are not visible. The last visible branching point of pulmonary arteries is approximately 1 cm from the pleural surface. Bronchi are invisible within 2 or 3 cm of the pleural surface. (From Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased isolated lungs: high-resolution CT. Radiology 1988;166:81, with permission.)

the younger patients. The septal thickening was limited to one lobe in two subjects and was widespread (two or more lobes) and bilateral in five.

Centrilobular Region and Centrilobular Structures The central portion of the lobule, referred to as the centrilobular region or lobular core (31), contains the pulmonary artery and bronchiolar branches that supply the lobule, lymphatics, and some supporting connective tissue (5,8,19,23,24,33). It is difficult to precisely define lobules in relation to the bronchial or arterial trees; lobules do not arise at a specific branching generation or from a specific type of bronchiole or artery (19). Branching of the lobular bronchiole and artery is irregularly dichotomous (36). When they divide, they generally

#152201  Cust: LWW  Au: Webb  Pg. No. 58

B FIGU RE 2-24  HRCT appearance of pulmonary lobules in patients with interlobular septal thickening. A: Lobular anatomy in a patient with interstitial pulmonary edema. Lobules are easily identified because of interlobular septal thickening. Centrilobular arteries are visible as branching of dotlike opacities in the centrilobular region. B: Lobular anatomy on a sagittal reconstruction in a patient with unilateral pulmonary vein atresia and interlobular septal thickening. Note that lobules are best delineated in the upper lobes.

divide into two branches. Most often, they divide into two branches of different sizes (one branch being nearly the same size as the one it arose from, and the other being smaller) (Figs. 2-10B and 2-11). Thus, on bronchograms, arteriograms, or HRCT, there often appears to be a single dominant bronchiole or artery in the center of the lobule, which gives off smaller branches at intervals along its length. The HRCT appearances and visibility of centrilobular structures are determined primarily by their size (Fig. 2-11). Secondary lobules are supplied by arteries and bronchioles measuring approximately 1 mm in diameter, whereas intralobular terminal bronchioles and arteries measure approximately 0.7 mm in diameter, and acinar bronchioles and arteries range from 0.3 to 0.5 mm in

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

diameter. Arteries of this size can be easily resolved using the HRCT technique (23,24). On clinical scans, a linear, branching, or dotlike opacity frequently seen within the center of a lobule, or within 1 cm of the pleural surface, usually represents the intralobular artery branch or its divisions (Figs. 2-20 to 2-24) (5,23,24). Pulmonary artery branches supplying a secondary pulmonary lobule and having a diameter of about 1 mm contain numerous parallel elastic laminae in their walls and are termed elastic arteries (46). Arteries smaller than 1 mm and larger than 0.1 mm in diameter generally contain smooth muscle in their walls, and are termed muscular arteries. Vessels smaller than 0.1 mm in diameter are termed pulmonary arterioles (46). The smallest arteries resolved using HRCT extend to within 3 to 5 mm of the pleural surface or lobular margin and are as small as 0.2 mm in diameter (5,23,24); thus, pulmonary arterioles are not visible on HRCT. Centrilobular arteries visible on HRCT are not seen to extend to the pleural surface in the absence of atelectasis (Figs. 2-20 to 2-24). The visibility of bronchioles in normal subjects depends on their wall thickness rather than diameter. For a 1-mm bronchiole supplying a secondary lobule, the thickness of its wall measures approximately 0.15 mm; this is at the lower limit of HRCT resolution. The wall of a terminal bronchiole measures only 0.1 mm in thickness, and that of an acinar bronchiole only 0.05 mm, both of which are below the resolution of HRCT technique for a tubular structure (Fig. 2-11). In one in vitro study, only bronchioles having a diameter of 2 mm or more or having a wall thickness of more than 100 mm (0.1 mm) were visible using HRCT (23); and resolution is certainly less than this on clinical scans. It is important to remember that on clinical HRCT, intralobular bronchioles are not normally visible, and bronchi or bronchioles are rarely seen within 1 cm of the pleural surface in most locations (Figs. 2-20 and 2-24) (15,25).

Lobular (Lung) Parenchyma and Lung Acini The substance of the secondary lobule, surrounding the centrilobular region and contained within the interlobular septa, consists of functioning lung parenchyma, namely, alveoli and the associated pulmonary capillary bed, supplied by small airways and branches of the pulmonary arteries and veins. This parenchyma is supported by a connective tissue stroma, a fine network of very thin fibers within the alveolar septa termed the intralobular interstitium (Fig. 2-1) (8,19), which is normally invisible. On HRCT, the lobular parenchyma should be of greater opacity than air, but this difference may vary with window settings (see Chapter 1). Some small intralobular vascular branches are often visible. It should be emphasized that all three interstitial fiber systems described by Weibel (axial, peripheral, and septal) are represented at the level of the pulmonary lobule (Fig. 2-1), and abnormalities in any can produce recognizable lobular abnormalities on HRCT (24). Axial

#152201  Cust: LWW  Au: Webb  Pg. No. 59

59

(centrilobular) fibers surround the artery and bronchiole in the lobular core, peripheral fibers comprising the interlobular septa marginate the lobule, and septal fibers (the intralobular interstitium) extend throughout the substance of the lobule in relation to the alveolar walls. Pulmonary acini are not themselves recognizable on HRCT. However, artery branches supplying pulmonary acini measure approximately 0.5 mm and are large enough to be seen (Fig. 2-11); Murata et al. (23) showed that pulmonary artery branches as small as 0.2 mm, associated with a respiratory bronchiole and thus acinar in location, are visible on HRCT and extend to within 3 to 5 mm of the lobular margins or pleural surface. As with the lobular bronchiole, first-order respiratory bronchioles supplying an acinus are too small to be seen.

THE CONCEPT OF CORTICAL AND MEDULLARY LUNG At least partially based on differences in lobular anatomy, it has been suggested that the lung can be divided into a peripheral cortex and a central medulla (31,42). Although these terms are not in general use, the concept of cortical and medullary lung regions is useful in highlighting differences in lung anatomy, as well as the varying appearances of secondary pulmonary lobules in the peripheral and central lung regions (47). It also serves to emphasize some anatomical (and perhaps physiologic) differences between the peripheral and central lung that are useful in predicting the HRCT distribution of some lung diseases (48).

Peripheral or Cortical Lung Cortical lung can be conceived of as consisting of two or three rows or tiers of well-organized and well-defined secondary pulmonary lobules, which together form a ­ layer 3 to 4 cm in thickness at the lung periphery and along the lung surfaces adjacent to the interlobar fissures (Fig. 2-25) (31,42). The pulmonary lobules in the lung cortex are relatively large in size and are marginated by interlobular septa that are thicker and better defined than in other parts of the lung; thus, cortical lobules tend to be better defined than those in the central or medullary lung. Bronchi and pulmonary vessels in the lung cortex are relatively small; although cortical arteries and veins are visible on HRCT, bronchi and bronchioles are uncommonly visible. This contrasts with the anatomy of medullary lung, in which large vessels and bronchi are visible. Lobules in the lung cortex tend to be relatively uniform in appearance and can be conceived of as being similar to the stones in a Roman arch: all of similar size and shape (Fig. 2-25) (42). They can appear cuboidal or be shaped like a truncated cone or pyramid (31). However, it should be remembered that the size, shape, and appearance of pulmonary lobules as seen on HRCT are significantly affected by the orientation of the scan plane relative to the central and longitudinal axes of the lobules. A single scan typically traverses different parts of

DESIGN SERVICES OF

60

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

Lung “cortex”

Lung “medulla”

adjacent lobules (Fig. 2-22), resulting in widely varying appearances of the lobules, despite the fact that they are all of similar size and shape.

Central or Medullary Lung Pulmonary lobules in the central or medullary lung are smaller and more irregular in shape than in the cortical lung and are marginated by interlobular septa that are thinner and less well defined. When visible, medullary lobules may appear hexagonal or polygonal in shape, but well-defined lobules are uncommonly seen in normal subjects. In contrast with the peripheral lung, perihilar vessels and bronchi in the lung medulla are large and easily seen on HRCT.

SUBPLEURAL INTERSTITIUM AND PLEURAL SURFACES Diffuse infiltrative lung diseases involving the subpleural interstitium or pleura can result in abnormalities visible at the pleural surfaces.

Subpleural Interstitium and Visceral Pleura The visceral pleura consists of a single layer of flattened mesothelial cells that is subtended by layers of fibroelastic connective tissue; it measures 0.1 to 0.2 mm in thickness (49,50). The connective tissue component of the visceral pleura is generally referred to on HRCT as the subpleural interstitium and is part of the “peripheral” interstitial fiber network described by Weibel (Fig. 2-1) (8). The subpleural interstitium contains small vessels, which are involved in the formation of pleural fluid, and lymphatic branches. Interstitial lung diseases that affect the interlobular septa or result in lung fibrosis often result in abnormalities of the subpleural interstitium. Abnormalities of the subpleural interstitium can be recognized over the costal surfaces of the lung, but are more

#152201  Cust: LWW  Au: Webb  Pg. No. 60

FIGU RE 2-25  Corticomedullary differentiation in the lung. The lung cortex is composed of one or two rows or tiers of wellorganized and well-defined secondary pulmonary lobules 3 to 4 cm in thickness. The pulmonary lobules in the lung cortex tend to be well defined and relatively large, and can be conceived of as being similar to the stones in a Roman arch: all of similar size and shape. The cortical airways and vessels are small, usually less than 2 to 3 mm in diameter.

easily seen in relation to the major fissures; in this location, two layers of the visceral pleura and subpleural interstitium come in contact. The major fissures are consistently visualized on HRCT as continuous, smooth, and very thin linear opacities. Normal fissures are less than 1 mm thick, smooth in contour, uniform in thickness, and sharply defined. The visceral pleura and subpleural interstitium along the costal surfaces of lung are not visible on HRCT in normal subjects. A few small dots in relation to the fissures or at the pleural surface may be seen in normal subjects, reflecting the presence of subpleural veins or the points of intersection of interlobular septa with the pleural surface.

Parietal Pleura The parietal pleura, as with the visceral pleura, consists of a mesothelial cell membrane in association with a thin layer of connective tissue. The parietal pleura is somewhat thinner than the visceral pleura, measuring approximately 0.1 mm (49,50). External to the parietal pleura is a thin layer of loose areolar connective tissue or extrapleural fat, which separates the pleura from the fibroelastic endothoracic fascia that lines the thoracic cavity (Figs. 2-26 and 2-27); the endothoracic fascia is approximately 0.25 mm thick (50,51). External to the endothoracic fascia are the innermost intercostal muscles and ribs. The innermost intercostal muscles pass between adjacent ribs but do not extend into the paravertebral regions. As stated in Chapter 1, window-level/width settings of 50/350 HU are best for evaluating the parietal pleura and adjacent chest wall. Images at a level of –600 HU with an extended window width of 2,000 HU are also useful in evaluating the relationship of peripheral parenchymal abnormalities to the pleural surfaces (5,52). On HRCT in normal patients, the innermost intercostal muscle is often visible as a 1- to 2-mm-thick stripe of soft-tissue opacity (i.e., the intercostal stripe) at the lung–chest wall interface, passing between adjacent rib segments in the anterolateral, lateral, and posterolateral thorax (Fig. 2-28). The parietal pleura is too thin to see on

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

61

FIGU RE 2-26  Anatomy of the pleural surfaces and chest wall.

FIGU RE 2-27  Anatomy of the parietal pleura and chest wall in a section of a cadaver. The parietal pleura and endothoracic fascia are visible as a thin white layer, lining the thoracic cavity. Little extra thoracic fat is present in this example. The innermost intercostal muscle is visible external to the parietal pleura, measuring 1 to 2 mm in thickness. External to this is a layer of fat containing the intercostal vessels and nerve. The intercostal muscles are absent in the paravertebral regions; only parietal pleura, endothoracic fascia, and paravertebral fat are visible.

FIGU RE 2-28  Normal intercostal stripe and paravertebral line. On HRCT in a normal subject, the intercostal stripe is visible as a thin white line. Although it represents the combined thickness of visceral and parietal pleurae, the fluid-filled pleural space, endothoracic fascia, and innermost intercostal muscle, it primarily represents the innermost intercostal muscle. The intercostal stripe is seen as separate from the more external layers of the intercostal muscles because of a layer of intercostal fat. Posteriorly, the intercostal stripe is visible anterior to the lower edge of a rib. Only a very thin line (i.e., the paravertebral line) is visible in the paravertebral region.

HRCT along the costal pleural surfaces, even in combination with the visceral pleura and endothoracic fascia (53). However, in the paravertebral regions, the innermost intercostal muscle is anatomically absent, and a very thin line (i.e., the paravertebral line) is sometimes visible at the interface between the lung and paravertebral fat or rib (Figs. 2-28 and 2-29) (53). This line probably represents the combined thickness (0.2–0.4 mm) of the normal pleural layers and endothoracic fascia.

NORMAL LUNG ATTENUATION Generally speaking, the lung appears homogeneous in attenuation on HRCT scans obtained at full inspiration. Measurements of lung attenuation in normal subjects usually range from –700 to –900 HU, corresponding to lung densities of approximately 0.300 to 0.100 g/mL, respectively (54,55). In most patients, normal mean lung attenuation ranges from 750 to 860 HU, although it may measure more or less in individual lung regions. In a study

#152201  Cust: LWW  Au: Webb  Pg. No. 61

FIGU RE 2-29  The paravertebral line. In the paravertebral regions (arrows), the innermost intercostal muscle is absent, and, at most, a very thin line (the paravertebral line) is present at the lung–chest wall interface. As in this case, a distinct line may not be seen.

DESIGN SERVICES OF

62

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

by Lamers et al. (56), with HRCT obtained using spirometric control of lung volume, the mean lung attenuation measured in 20 healthy subjects at 90% of vital capacity was –859 HU (standard deviation [SD], 39) in the upper lung zones and –847 (SD, 34) in the lower lung zones. A study by Chen et al. (57) of 13 patients with normal pulmonary function tests showed an average lung attenuation of –814 ± 24 HU on HRCT when the entire cross-section of lung was used for measurement and an attenuation of –829 ± 21 HU (range, –858 to –770 HU) using three small regions of interest placed in anterior, middle, and posterior lung regions. A mean lung density of –866 ± 16 HU was found by Gevenois et al. (58) in a study of 42 healthy subjects (21 men, 21 women) from 23 to 71 years of age. In this study, no significant correlation between mean lung density and age was found, but there was a significant correlation between the total lung capacity, expressed as absolute values and mean lung density. However, these authors found a significant correlation between the relative area of pixels less than –950 HU (usually considered emphysema) and age (r = 0.328; p = 0.034). Others (59) have found a significant correlation between both mean lung density and the percentage of pixels with lung density less than thresholds of –910 and –950 HU, with age and sex. Although the absolute attenuation differences found in this study were small, significantly lower mean lung attenuation was found for nonsmokers older than 75 years (901.7 HU ± 2.5; p = 0.006), compared to those less than 55 years of age (889.7 HU ± 3.4), and men (905.6 HU ± 2.7; p = 0.003) as compared to women (894.3 HU ± 2.6) (59). Irion et al. (60) also found that young (ages 19–41), healthy nonsmokers, with no recognizable lung disease, can have a small proportion of lung on inspiratory HRCT (mean value 0.19%) with an attenuation of less than –950 HU. On expiration, this percentage decreased to 0.04%. Mets et al. found this value to be 0.97% in a study of 70 healthy young men with normal spirometry, most of whom were nonsmokers (61). An attenuation gradient is normally present, with the most dependent lung regions being the densest and the least dependent lung regions being the least dense. This gradient is largely caused by regional differences in blood and gas volume that, in turn, are determined by gravity, mechanical stresses on the lung, and intrapleural pressure (52,54). Differences in attenuation between anterior and posterior lung have been measured in supine patients, and values generally range from 50 to 100 HU (54,62,63), although gradients of more than 200 HU have been reported (62). The anteroposterior attenuation gradient was found to be nearly linear and was present regardless of whether the subject was supine or prone (62). Genereux measured anteroposterior attenuation gradients at three levels (aortic arch, carina, and above the right hemidiaphragm) in normal subjects (63). An anteroposterior attenuation gradient was found at all levels,

#152201  Cust: LWW  Au: Webb  Pg. No. 62

although the gradient was greater at the lung bases than in the upper lung; the anteroposterior gradient averaged 36 HU at the aortic arch, 65 HU at the carina, and 88 HU at the lung bases. The attenuation gradient was even larger if only cortical lung was considered. Within cortical lung, the attenuation differences at the three levels studied were 45, 81, and 113 HU, respectively. Vock et al. (54) analyzed CT-measured pulmonary attenuation in children. In general, lung attenuation in children is greater than in adults (54,62), but anteroposterior attenuation gradients were similar to those found in adults, averaging 56 HU at the subcarinal level. Although most authors have reported that normal anteroposterior lung attenuation gradients are linear, with attenuation increasing gradually from anterior to posterior lung, the lingula and superior segments of the lower lobes can appear relatively lucent in many normal subjects (64); focal lucency in these segments should be considered a normal finding. Although the reason is unclear, these slender segments may be less well ventilated than adjacent lung and therefore less well perfused, or some air trapping may be present.

NORMAL EXPIRATORY HIGH-RESOLUTION COMPUTED TOMOGRAPHY Expiratory HRCT is generally performed to detect air trapping in patients with small airway obstruction or emphysema. On expiratory scans in normal subjects, HRCT findings include an increase in lung attenuation, decrease in cross-sectional lung area (65), and reduction in airway size (66). Air trapping of limited extent may also be seen in normal subjects.

Lung Attenuation Changes with Expiration Lung parenchyma normally increases in CT attenuation as lung volume is reduced during expiration. This change can generally be recognized on HRCT as an increase in  lung opacity (Figs. 2-30 to 2-32; see Figs. 1-27 and 1-30) (10,54,62,67–70). Robinson and Kreel (68) found a significant inverse correlation between lung volume determined spirometrically and CT-measured lung attenuation, for the whole lung (r = –0.680, p > 0.0005) and for anterior, middle, and posterior lung zones considered individually. The mean lung attenuation increase between full inspiration and expiration ranges from 80 to more than 300 HU, with the largest changes being found (a) in dependent lung regions, (b) at the lung bases, (c) with dynamic scanning during forced expiration, and (d) when measurements are made using small (e.g., 2–4 cm) regions of interest rather than the entire cross-section of lung (10,54,56,64,69–73). In a study of young, normal volunteers, an increase in lung attenuation averaging 200

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

63

A

B

C

FIGU RE 2-30  Normal dynamic expiratory HRCT. Inspiratory (A) and expiratory (B) images from a sequence of 10 scans obtained during forced expiration in a normal subject. Lung attenuation increases and cross-sectional lung area decreases on the expiratory scan. C: A region of interest has been positioned in the posterior lung, and a timeattenuation curve calculated for this region of interest shows an increase in attenuation from –850 to –625 from maximal inspiration (I) to maximal expiration (E). Each point on the time-density curve represents one image from the dynamic sequence.

A

B

FIGU RE 2-31  Dynamic inspiratory (A) and expiratory (B and C) HRCT

in a normal subject, obtained with low (40) mA. On the inspiratory scan (A), lungs appear homogeneous in attenuation. Lung attenuation measured –875 HU in the posterior right lung. During rapid expiration (B), image quality is degraded by respiratory motion. On a scan at maximum expiration (C), lung decreases in volume and increases in attenuation. Posterior dependent lung increases in attenuation to a greater degree than anterior nondependent lung, now measuring –750 HU. Note some anterior bowing of the posterior tracheal membrane, typical of expiratory images.

C

#152201  Cust: LWW  Au: Webb  Pg. No. 63

DESIGN SERVICES OF

64

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

A

B

FIGU RE 2-32  Inspiratory (A) and postexpiratory (B) HRCT in a normal subject. On the expiratory scan, lung increases in attenuation. Posterior dependent lung increases in attenuation to a greater degree than anterior nondependent lung.

HU ± 29.7 was found during dynamic forced expiration, using 2-cm regions of interest, but the increase was variable and ranged from 84 to 372 HU in different patients and considering dependent and nondependent lung regions separately (64). In a study of 10 nonsmokers with normal pulmonary function, using 4-cm regions of interest, the attenuation increase following expiration ranged from 35 to 139 HU (mean, 90 HU) in the anterior middle lobe, 64 to 147 HU (mean, 122 HU) in the anterior lingua, and 100 to 363 HU (mean, 237 HU) in the posterior lower lobes (73). In a study by Chen et al. (57) of patients with normal pulmonary function tests, the average lung attenuation increase on postexpiratory HRCT was 144 ± 47 HU (range, 85–235 HU) when three small regions of interest placed in different lung regions were used for measurement and 149 ± 54 HU when the entire crosssection of lung was used for measurement. Average lung attenuation on postexpiratory HRCT was –685 HU ± 51 (range, –763 to –580 HU) using three regions of interest and –665 ± 80 HU for the entire cross-section of lung (57). Other studies have shown an increase in average postexpiratory lung attenuation increase of about 100 to 130 HU in normal subjects or patients with normal pulmonary function (70,71,74). Millar and Denison (67) calculated the physical density of lung at full inspiration and expiration, based on the assumption that physical density had linear relation to radiographic density (physical density = 1 – CT attenuation in HU/1,000) (75). Using this method, peripheral lung tissue density was measured as

#152201  Cust: LWW  Au: Webb  Pg. No. 64

0.0715 g/cm3 (SD, 0.017) at full inspiration and 0.272 g/ cm3 (SD, 0.067) at end expiration. Using dynamic expiratory HRCT, a greater increase in lung attenuation may be seen than with static imaging. CT obtained with spirometric control has been used to determine lung attenuation at specific lung volumes. According to Kalender et al., using spirometrically triggered CT (76), a 10% change in vital capacity resulted, on average, in a change of approximately 16 HU, and estimates of lung attenuation at 0% and 100% of vital capacity were –730 and –895 HU, respectively. In a study by Lamers et al. (56), with HRCT obtained using spirometric control of lung volume, the mean lung attenuation in 20 healthy subjects measured in the upper lung zones at 90% of vital capacity was –859 HU (SD, 39), whereas at 10% of vital capacity, it was –786 HU (SD, 39). In the lower lung zones, lung attenuation increased from –847 HU (SD, 34) at 90% of vital capacity to –767 HU (SD, 56) at 10% of vital capacity. In a study of spirometrically gated HRCT (77) at 20%, 50%, and 80% of vital capacity, mean lung attenuation measured –747, –816, and –855 HU, respectively. In children, the CT attenuation of lung parenchyma is higher than in adults and decreases with age (54,62). Attenuation increases seen with expiration are similar to those found in adults. Ringertz et al. (78), using ultrafast CT, measured the CT attenuation of children younger than 2.5 years during quiet respiration; the average CT lung attenuation was –551 HU (SD, 106) on inspiration

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

and –435 HU (SD, 103) on expiration. Vock et al. (54) measured the lung attenuation changes in children ranging in age from 9 to 18 years. Mean lung attenuation at full inspiration and full expiration measured –804 and –646 HU, respectively. The anteroposterior attenuation differences were similar to those seen in adults, averaging 56 HU at the subcarinal level, and increased with maximal expiration and increased during expiration (54). Long et al. (79) described normal regional CT lung attenuation measured using controlled and suspended ventilation at end inspiration and end expiration in sedated children aged 3 months to 4.5 years. This study used a positive pressure of 25 cm of water to simulate inspiration, resulting in approximately 95% of total lung capacity, and a pressure of 0 for expiration. The average lung attenuation of all children studied averaged –834 HU (± 44 HU) at end inspiration and –633 HU (±102 HU) at end expiration. Lung density declined linearly in the first few years of life and thereafter approximated adult values. Usually, dependent lung regions show a greater increase in lung attenuation during expiration than do nondependent lung regions irrespective of the patient’s position (10,62,64,68,69,80). As a result, the anteroposterior attenuation gradients normally seen on inspiration are significantly greater on expiratory scans (Fig. 2-32) (54,68,69); the increase in the anterior-to-­posterior attenuation gradient after expiration has been reported to range from 47 to 130 HU in different studies (10,54,64,69). Furthermore, the expiratory lung attenuation increase in dependent lung regions is greater in the lower lung zones than in the middle and upper zones, probably due to greater diaphragmatic movement or greater basal blood volume (64). In the study by Tanaka et al. of asymptomatic nonsmokers with normal pulmonary function, the postexpiratory attenuation increase at the lung bases averaged 131.4 ± 52.1 HU, compared to 102 ± 55.3 and 100 ± 40.2 in the upper and midlungs, respectively (72); similar differences were also seen in smokers and exsmokers. The sum of these changes may be recognizable

as increased attenuation or dependent density on supine scans at low lung volume. Although using measurements of attenuation gradients on inspiration and expiration has been investigated as a method of diagnosing lung disease (10,67,81), this technique has yet not assumed a significant clinical role.

Air Trapping on Expiratory High-Resolution Computed Tomography Abnormal retention of gas (i.e., air) within a lung or part of a lung, as a result of airway obstruction or abnormalities in lung compliance, is termed air trapping. Air trapping is present if lung parenchyma remains lucent on expiratory scans or shows a less than normal increase in attenuation after expiration. In as many as 60% of normal subjects, areas of air trapping are visible on expiratory HRCT scans (Figs. 2-33 and 2-34). This appearance is most common in the dependent lung, at the lung bases, and in the superior segments of the lower lobes (82). Air trapping may involve individual pulmonary lobules or groups of lobules (64,72,83,84). Normal air trapping is generally limited to a small proportion of lung volume, although the range of reported values varies from about 5% to about 25% if smokers are included in the analysis (57,72,73,84). Tanaka et al. (72) found findings of air trapping in 64% of asymptomatic patients with normal pulmonary function. In a study by Chen et al. (57), focal areas of air trapping, including the superior segments of the lower lobes, were visible in 61% of patients with normal pulmonary function tests; air trapping involved up to 25% of lung when the superior segments were included in analysis. In a study by Lee et al. (83), air trapping was seen in 52% of 82 asymptomatic subjects with normal pulmonary function tests. Lee et al. (83) also found that the frequency of air trapping increased with age (p < 0.05), being seen in 23% of patients aged 21 to 30 years, 41% of those aged 31 to 40 years, 50% of those aged 41 to 50 years, 65% of

B

A

FIGU RE 2-33  Inspiratory (A) and postexpiratory (B) HRCT in a normal subject. On the expiratory scan, there is relative lucency in the superior segments of the lower lobes, posterior to the major fissures. This appearance is normal. Also, focal air trapping is present in a single lobule (arrow) in the posterior right lung. Note the slight anterior bowing of the posterior right BI. This may be seen in some patients on expiration.

#152201  Cust: LWW  Au: Webb  Pg. No. 65

65

DESIGN SERVICES OF

66

s e c t i o n I High-Resolution CT Techniques and Normal Anatomy

FIGU RE 2-34  Dynamic expiratory HRCT in a normal subject showing air trapping in the anterior lingula (arrows) and relative lucency posterior to the left major fissure. Pulmonary lobules in the lung medulla are smaller and less well defined than in the periphery. However, vessels and bronchi in the lung medulla are large and easily seen on HRCT. Note the anterior bowing of the posterior wall of the right bronchus.

those aged 51 to 60 years, and 76% of those older than 61 years. In a study of 10 young, normal subjects, Webb et al. (64) found that although air trapping was present in four patients (40%), it was limited in extent. In another study, discounting the superior segments and air trapping involving less than two contiguous or five noncontiguous pulmonary lobules, air trapping was not seen on expiratory scans in 10 healthy nonsmokers, although it was visible in 40% of patients with suspected chronic airways disease who had normal pulmonary function tests (85). Mastora et al. (84) assessed inspiratory and postexpiratory HRCT in 250 volunteers, including 144 smokers, 47 ex-smokers, and 59 nonsmokers. Air trapping was seen in 62% of the subjects. Lobular air trapping (fewer than three adjacent lobules) was seen in 47%, without significant differences among smokers, ex-smokers, and nonsmokers. Segmental (ranging from three adjacent lobules to a segment) air trapping (seen in 14%) and lobar (larger than a segment) air trapping (seen in 1%) were more frequent among smokers and ex-smokers (p < 0.001). Air trapping was limited to less than 25% of lung area in 72.5% of subjects with air trapping. Tanaka et al. (72) studied 50 subjects with normal pulmonary function, including 26 nonsmokers and 24 smokers (14 current and 10 ex-smokers). All 50 subjects who underwent HRCT with images were obtained during deep inspiration and end expiration at three levels. Air trapping was visually classified into four degrees (none, lobular, mosaic, or extensive), and the extent of air trapping was also calculated. The mean increase in lung attenuation in the three levels at expiration was 111.9 HU ± 46.3. The overall frequency of air trapping was 64%. Lobular (one or two adjacent lobules), mosaic (three or more regions of lobular air trapping), and extensive (larger than three adjacent lobules and subsegmental, segmental, or lobar in distribution) air trapping were seen in 10 (20%), 14 (28%), and 8 (16%) patients, respectively. There was no

#152201  Cust: LWW  Au: Webb  Pg. No. 66

significant difference in the visual grade and extent of air trapping among the nonsmokers, smokers, and ex-smokers (72). The extent of air trapping relative to cross-sectional lung area averaged 5.6 ± 6.4% in nonsmokers (range, 0%–20.4%) and 5.9 ± 4.2% and 6.6 ± 4.5% (range, 0%–13.8%) in smokers and ex-smokers, respectively. Postexpiratory minimum-intensity projection (MinIP) images may be useful in detecting air trapping and can increase the conspicuity of this finding (73). Wittram et al. (73) reviewed inspiratory and postexpiratory HRCT obtained using helical technique and 1-mm collimation to 10-mm-thick MinIP images in 10 healthy nonsmokers with normal pulmonary function tests. HRCT and MinIPs demonstrated a smooth anterior-to-posterior attenuation gradient within the lung parenchyma, which was accentuated by expiration. Expiratory HRCT and MinIPs demonstrated air trapping in 8 of 10 subjects and in 31 of the 40 regions assessed, although the extent of air trapping was limited, averaging 7.2% of lung area.

Changes in Cross-Sectional Lung Area The reduction in cross-sectional lung area that occurs with expiration has been assessed in several studies and usually ranges from 40% to 50%. In a study of dynamic expiratory HRCT, Webb et al. (64) determined the percent decrease in lung cross-sectional area from full inspiration to full expiration in 10 young, normal volunteers. The area change ranged from 14.8% to 61.3% for all subjects, subject positions, and lung regions. The greatest percentage decrease in cross-sectional area during exhalation occurred in the upper lung zones. This value averaged 51.3% (SD, 6.7) in the supine position and 43.1% (SD, 10.2) in the prone position. The percentage decrease in lung cross-sectional area was least at the lung bases, averaging 30.9% (SD, 7.5) in the supine position and 25.2% (SD, 5) in the prone position. The average area changes for the midlung regions were intermediate between those of upper and lower lung zones, measuring 38.9% (SD, 7.4) in the supine position and 36.7% (SD, 5.3) in the prone position. Similarly, in a study by Lucidarme et al. (85), cross-sectional lung area decreased by an average of 43% (range, 34%–57%) in a group of 10 normal volunteers. Mitchell et al. (65) measured lung area on inspiratory and end expiratory scans at the level of the carina in 78 normal subjects. The percentage change in area from inspiration to expiration averaged 55% (SD, 8.7%). Ederle et al. (71) found a decrease in cross-sectional lung area of 24% in 47 patients with normal pulmonary function. Changes in cross-sectional lung area during expiration can be related to changes in lung attenuation as shown on HRCT. Simply stated, attenuation increases at the same time that cross-sectional lung area decreases during expiration (Fig. 2-30). For example, Robinson and Kreel (68) found a significant inverse correlation between the expiratory change in cross-sectional lung area measured on CT and changes in CT-measured lung attenuation (r = –0.793,

DESIGN SERVICES OF

C hapter 2 Normal Lung Anatomy

p > 0.0005). In a study using dynamic expiratory HRCT (64), a correlation between cross-sectional lung area and lung attenuation was found for each of three lung regions evaluated (upper lung, r = 0.51, p = 0.03; midlung, r = 0.58, p = 0.01; lower lung, r = 0.51, p = 0.05). The lower lung zone showed a greater attenuation increase for a given area change; this phenomenon likely reflects the much greater effect of diaphragmatic elevation on basal lung attenuation than occurs in the upper lungs. In a study by Ederle et al. (71), inspiratory and expiratory HRCT were compared in 47 patients with normal pulmonary function. Mean lung attenuation correlated with cross-sectional lung area on both inspiratory (r = –0.66, p < 0.0005) and expiratory scans (r = –0.63, p 1 cm

Usually 10 mm, findings of emphysema (e.g., centrilobular emphysema)

FIGU RE 6-36  Algorithmic approach to cystic airspaces in the subpleural lung.

184

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

Cystic airspaces

Subpleural

Intraparenchymal

No visible walls

Visible walls

Spotty, holes small, UL predominant

Patchy, central, branching, signet ring sign, air-fluid levels

Numerous, symmetrical, intervening lung normal

Few or scattered cysts, with ground-glass opacity or other findings of lung disease

Centrilobular emphysema

Cystic bronchiectasis

Cystic lung disease

Pneumatoceles, isolated cysts e.g. DIP, PJP, HP

Some cysts UL subpleural, large, predominant, LL predominance, cysts skin lesions, irregular shapes, renal tumors smoker Birt-Hogg-Dube syndrome

LCH

Diffuse, cysts round, woman

Limited number of cysts, associated with vessels, Sjögren syndrome, AIDS

LAM

LIP amyloidosis

FIGU RE 6-37  Algorithmic approach to intraparenchymal cystic airspaces.

The term lung cyst is used to describe a thin-walled, welldefined, and well-circumscribed air-­ containing lesion. Although there are a number of causes of cystic lung disease, the algorithm in Fig. 6-37 reviews only the diseases most frequently seen; the appearances of other causes are listed in Table 6-1. LCH and LAM result in multiple lung cysts (21–27). The cysts have a thin but easily discernible wall, ranging up to a few millimeters in thickness. Associated findings of fibrosis are usually absent or much less conspicuous than they are in patients with honeycombing. In these diseases, the cysts are usually interspersed within areas of normal-appearing lung. In patients with LCH, the cysts can have bizarre shapes, typically have upper lobe predominance, and may occur in men. LAM is associated with rounder and more uniformly shaped cysts, is diffusely distributed from apex to base, and exclusively occurs in women. Cysts are sometimes seen in patients with LIP associated with Sjögren syndrome, AIDS, or other systemic diseases; cystic airspaces in LIP have thin walls, measure 1 to 30 mm in diameter, and are typically less numerous than those in LCH and LAM (15,28,76). They may be associated with vessels, and cysts in amyloidosis or light-chain disease, also associated with Sjögren syndrome, may have a similar appearance. Cysts representing pneumatoceles can be seen in patients with infection, particularly P.  ­jirovecii pneumonia; pneumatoceles are

often scattered and patchy in distribution and limited in ­number, and findings of pneumonia or a history of pneumonia may be present. Scattered cysts are also seen in association with lung disease in patients with HP and DIP. Birt-Hogg-Dube syndrome is associated with large subpleural cysts, renal tumor, and skin lesions.

REFERENCES 1. Naidich DP. High-resolution computed tomography of cystic lung disease. Semin Roentgenol 1991;26:151–174. 2. Hansell DM, Bankier AA, MacMahon H, et al. Fleischner S­ ociety: glossary of terms for thoracic imaging. Radiology 2008;246: 697–722. 3. Austin JH, Müller NL, Friedman PJ, et al. Glossary of terms for CT of the lungs: recommendations of the Nomenclature Committee of the Fleischner Society. Radiology 1996;200:327–331. 4. Tuddenham WJ. Glossary of terms for thoracic radiology: recommendations of the Nomenclature Committee of the Fleischner Society. AJR Am J Roentgenol 1984;143:509–517. 5. Müller NL, Miller RR, Webb WR, et al. Fibrosing alveolitis: CT-pathologic correlation. Radiology 1986;160:585–588. 6. Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased isolated lungs: high-resolution CT. Radiology 1988;166:81–87. 7. Nishimura K, Kitaichi M, Izumi T, et al. Usual interstitial pneumonia: histologic correlation with high-resolution CT. Radiology 1992;182:337–342. 8. Hogg JC. Benjamin Felson lecture. Chronic interstitial lung disease of unknown cause: a new classification based on pathogenesis. AJR Am J Roentgenol 1991;156:225–233.

C hapter 6 HRCT Findings: Air-Filled Cystic Lesions 9. Galvin JR, Frazier AA, Franks TJ. Collaborative radiologic and histopathologic assessment of fibrotic lung disease. Radiology 2010;255:692–706. 10. Aquino SL, Webb WR, Zaloudek CJ, et al. Lung cysts associated with honeycombing: change in size on expiratory CT scans. AJR Am J Roentgenol 1994;162:583–584. 11. Worthy SA, Brown MJ, Müller NL. Technical report: cystic air spaces in the lung: change in size on expiratory high-resolution CT in 23 patients. Clin Radiol 1998;53:515–519. 12. Lee KN, Yoon SK, Choi SJ, et al. Cystic lung disease: a comparison of cystic size, as seen on expiratory and inspiratory HRCT scans. Korean J Radiol 2000;1:84–90. 13. Souza CA, Finley R, Müller NL. Birt-Hogg-Dubé syndrome: a rare cause of pulmonary cysts. AJR Am J Roentgenol 2005;185: 1237–1239. 14. Jeong YJ, Lee KS, Chung MP, et al. Amyloidosis and lymphoproliferative disease in Sjögren syndrome: thin-section computed tomography findings and histopathologic comparisons. J Comput Assist Tomogr 2004;28:776–781. 15. Desai SR, Nicholson AG, Stewart S, et al. Benign pulmonary lymphocytic infiltration and amyloidosis: computed tomographic and pathologic features in three cases. J Thorac Imaging 1997;12: 215–220. 16. Colombat M, Mal H, Copie-Bergman C, et al. Primary cystic lung light chain deposition disease: a clinicopathologic entity derived from unmutated B cells with a stereotyped IGHV4-34/IGKV1 receptor. Blood 2008;112:2004–2012. 17. Colombat M, Stern M, Groussard O, et al. Pulmonary cystic disorder related to light chain deposition disease. Am J Respir Crit Care Med 2006;173:777–780. 18. Wongsripuemtet J, Ruangchira-urai R, Stern EJ, et al. Benign metastasizing leiomyoma. J Thorac Imaging 2012;27:W41–W43. 19. Seaman DM, Meyer CA, Gilman MD, et al. Diffuse cystic lung disease at high-resolution CT. AJR Am J Roentgenol 2011;196:1305–1311. 20. Irion KL, Hocchegger B, Marchiori E, et al. Proteus syndrome: highresolution CT and CT pulmonary densitovolumetry findings. J Thorac Imaging 2009;24:45–48. doi:10.1097/RTI.0b013e31818b20cd 21. Moore AD, Godwin JD, Müller NL, et al. Pulmonary histiocytosis X: comparison of radiographic and CT findings. Radiology 1989;172:249–254. 22. Brauner MW, Grenier P, Mouelhi MM, et al. Pulmonary histiocytosis X: evaluation with high-resolution CT. Radiology 1989;172: 255–258. 23. Lenoir S, Grenier P, Brauner MW, et al. Pulmonary lymphangiomyomatosis and tuberous sclerosis: comparison of radiographic and thin-section CT findings. Radiology 1990;175:329–334. 24. Templeton PA, McLoud TC, Müller NL, et al. Pulmonary lymphangioleiomyomatosis: CT and pathologic findings. J Comput Assist Tomogr 1989;13:54–57. 25. Sherrier RH, Chiles C, Roggli V. Pulmonary lymphangioleiomyomatosis: CT findings. AJR Am J Roentgenol 1989;153:937–940. 26. Rappaport DC, Weisbrod GL, Herman SJ, et al. Pulmonary lymphangioleiomyomatosis: high-resolution CT findings in four cases. AJR Am J Roentgenol 1989;152:961–964. 27. Müller NL, Chiles C, Kullnig P. Pulmonary lymphangiomyomatosis: correlation of CT with radiographic and functional findings. Radiology 1990;175:335–339. 28. Johkoh T, Müller NL, Pickford HA, et al. Lymphocytic interstitial pneumonia: thin-section CT findings in 22 patients. Radiology 1999;212:567–572. 29. Ichikawa Y, Kinoshita M, Koga T, et al. Lung cyst formation in Lymphocytic interstitial pneumonia: CT features. J Comput Assist Tomogr 1994;18:745–748. 30. Abbott GF, Rosado-de-Christenson ML, Frazier AA, et al. From the archives of the AFIP: lymphangioleiomyomatosis: radiologicpathologic correlation. Radiographics 2005;25:803–828. 31. Avila NA, Kelly JA, Dwyer AJ, et al. Lymphangioleiomyomatosis: correlation of qualitative and quantitative thin-section CT with pulmonary function tests and assessment of dependence on pleurodesis. Radiology 2002;223:189–197.

185

32. Pallisa E, Sanz P, Roman A, et al. Lymphangioleiomyomatosis: pulmonary and abdominal findings with pathologic correlation. Radiographics 2002;22 Spec No:S185–S198. 33. Koyama M, Johkoh T, Honda O, et al. Chronic cystic lung disease: diagnostic accuracy of high-resolution CT in 92 patients. AJR Am J Roentgenol 2003;180:827–835. 34. Abbott GF, Rosado-de-Christenson ML, Franks TJ, et al. From the archives of the AFIP: pulmonary Langerhans cell histiocytosis. Radiographics 2004;24:821–841. 35. Abbritti M, Mazzei MA, Bargagli E, et al. Utility of spiral CAT scan in the follow-up of patients with pulmonary Langerhans cell histiocytosis. Eur J Radiol 2012;81:1907–1912. 36. Attili AK, Kazerooni EA, Gross BH, et al. Smoking-related interstitial lung disease: radiologic-clinical-pathologic correlation. Radiographics 2008;28:1383–1396. 37. Galvin JR, Franks TJ. Smoking-related lung disease. J Thorac Imaging 2009;24:274–284. doi:10.1097/RTI.0b013e3181c1abb7 38. Seely JM, Salahudeen SS, Cadaval-Goncalves AT, et al. Pulmonary Langerhans cell histiocytosis: a comparative study of computed tomography in children and adults. J Thorac Imaging 2012;27: 65–70. doi:10.1097/RTI.0b013e3181f49eb6 39. Avila NA, Dwyer AJ, Rabel A, et al. Sporadic lymphangioleiomyomatosis and tuberous sclerosis complex with lymphangioleiomyomatosis: comparison of CT features. Radiology 2007;242:​ 277–285. 40. Schmithorst VJ, Altes TA, Young LR, et al. Automated algorithm for quantifying the extent of cystic change on volumetric chest CT: initial results in lymphangioleiomyomatosis. AJR Am J Roentgenol 2009;192:1037–1044. 41. Koyama M, Johkoh T, Honda O, et al. Pulmonary involvement in primary Sjogren’s syndrome: spectrum of pulmonary abnormalities and computed tomography findings in 60 patients. J Thorac Imaging 2001;16:290–296. 42. Tobino K, Hirai T, Johkoh T, et al. Differentiation between BirtHogg-Dube syndrome and lymphangioleiomyomatosis: quantitative analysis of pulmonary cysts on computed tomography of the chest in 66 females. Eur J Radiol 2012;81:1340–1346. 43. Agarwal PP, Gross BH, Holloway BJ, et al. Thoracic CT findings in Birt-Hogg-Dube syndrome. Am J Roentgenol 2011;196:349–352. 44. Copley SJ, Wells AU, Hawtin KE, et al. Lung morphology in the elderly: comparative CT study of subjects over 75 years old versus those under 55 years old. Radiology 2009;251:566–573. 45. Pipavath SNJ, Schmidt RA, Takasugi JE, et al. Chronic obstructive pulmonary disease: radiology-pathology correlation. J Thorac Imaging 2009;24:171–180. doi:10.1097/RTI.0b013e3181b32676 46. Sanders C, Nath PH, Bailey WC. Detection of emphysema with computed tomography: correlation with pulmonary function tests and chest radiography. Invest Radiol 1988;23:262–266. 47. Müller NL, Staples CA, Miller RR, et al. “Density mask”: an objective method to quantitate emphysema using computed tomography. Chest 1988;94:782–787. 48. Murata K, Itoh H, Todo G, et al. Centrilobular lesions of the lung: demonstration by high-resolution CT and pathologic correlation. Radiology 1986;161:641–645. 49. Miller RR, Müller NL, Vedal S, et al. Limitations of computed tomography in the assessment of emphysema. Am Rev Respir Dis 1989;139:980–983. 50. Hruban RH, Meziane MA, Zerhouni EA, et al. High resolution computed tomography of inflation fixed lungs: pathologicradiologic correlation of centrilobular emphysema. Am Rev Respir Dis 1987;136:935–940. 51. Goldin JG. Imaging the lungs in patients with pulmonary ­emphysema. J Thorac Imaging 2009;24:163–170. doi:10.1097/ RTI.0b013e3181b41b53 52. Kinsella N, Müller NL, Vedal S, et al. Emphysema in silicosis: a comparison of smokers with nonsmokers using pulmonary function testing and computed tomography. Am Rev Respir Dis 1990;141:1497–1500. 53. Akira M, Higashihara T, Yokoyama K, et al. Radiographic type p pneumoconiosis: high-resolution CT. Radiology 1989;171: 117–123.

186

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

54. Stern EJ, Webb WR, Weinacker A, et al. Idiopathic giant bullous emphysema (vanishing lung syndrome): imaging findings in nine patients. AJR Am J Roentgenol 1994;162:279–282. 55. Sharma N, Justaniah AM, Kanne JP, et al. Vanishing lung syndrome (giant bullous emphysema): CT findings in 7 patients and a literature review. J Thorac Imaging 2009;24:227–230. doi:10.1097/ RTI.0b013e31819b9f2a 56. Matsuoka S, Yamashiro T, Washko GR, et al. Quantitative CT assessment of chronic obstructive pulmonary disease. Radiographics 2010;30:55–66. 57. Coxson HO, Rogers RM. Quantitative computed tomography of chronic obstructive pulmonary disease. Acad Radiol 2005;12:1457–1463. 58. Lynch DA, Newell JD. Quantitative imaging of COPD. J Thorac Imaging 2009;24:189–194. doi:10.1097/RTI.0b013e3181b31cf0 59. Gevenois PA, De Vuyst P, de Maertelaer V, et al. Comparison of computed density and microscopic morphometry in pulmonary emphysema. Am J Respir Crit Care Med 1996;154:187–192. 60. Gevenois PA, de Maertelaer V, De Vuyst P, et al. Comparison of computed density and macroscopic morphometry in pulmonary emphysema. Am J Respir Crit Care Med 1995;152:653–657. 61. Mets OM, Murphy K, Zanen P, et al. The relationship between lung function impairment and quantitative computed tomography in chronic obstructive pulmonary disease. Eur Radiol 2012;22: 120–128. 62. Akira M, Toyokawa K, Inoue Y, et al. Quantitative CT in chronic obstructive pulmonary disease: inspiratory and expiratory assessment. AJR Am J Roentgenol 2009;192:267–272. 63. Feuerstein I, Archer A, Pluda JM, et al. Thin-walled cavities, cysts, and pneumothorax in Pneumocystis carinii pneumonia: further observations with histopathologic correlation. Radiology 1990;174: 697–702. 64. Panicek DM. Cystic pulmonary lesions in patients with AIDS [Editorial]. Radiology 1989;173:12–14.

65. Gurney JW, Bates FT. Pulmonary cystic disease: comparison of Pneumocystis carinii pneumatoceles and bullous emphysema due to intravenous drug abuse. Radiology 1989;173:27–31. 66. Im JG, Itoh H, Shim YS, et al. Pulmonary tuberculosis: CT findings— early active disease and sequential change with antituberculous therapy. Radiology 1993;186:653–660. 67. Nishimura K, Itoh H, Kitaichi M, et al. Pulmonary sarcoidosis: correlation of CT and histopathologic findings. Radiology 1993;189: 105–109. 68. Grenier P, Cordeau MP, Beigelman C. High-resolution computed tomography of the airways. J Thorac Imaging 1993;8:213–229. 69. Lynch DA, Newell JD, Tschomper BA, et al. Uncomplicated asthma in adults: comparison of CT appearance of the lungs in asthmatic and healthy subjects. Radiology 1993;188:829–833. 70. Matsuoka S, Uchiyama K, Shima H, et al. Bronchoarterial ratio and bronchial wall thickness on high-resolution CT in asymptomatic subjects: correlation with age and smoking. AJR Am J Roentgenol 2003;180:513–518. 71. Ooi GC, Khong PL, Chan-Yeung M, et al. High-resolution CT quantification of bronchiectasis: clinical and functional correlation. Radiology 2002;225:663–672. 72. Kim JS, Müller NL, Park CS, et al. Cylindrical bronchiectasis: diagnostic findings on thin-section CT. AJR Am J Roentgenol 1997;168: 751–754. 73. Kang EY, Miller RR, Müller NL. Bronchiectasis: comparison of preoperative thin-section CT and pathologic findings in resected specimens. Radiology 1995;195:649–654. 74. Naidich DP, McCauley DI, Khouri NF, et al. Computed tomography of bronchiectasis. J Comput Assist Tomogr 1982;6:437–444. 75. Westcott JL, Cole SR. Traction bronchiectasis in end-stage pulmonary fibrosis. Radiology 1986;161:665–669. 76. Franquet T, Giménez A, Monill JM, et al. Primary Sjögren’s syndrome and associated lung disease: CT findings in 50 patients. AJR Am J Roentgenol 1997;169:655–658.

7

HRCT Findings: Decreased Lung Attenuation I M P O R T A N T

T O P I C S

MOSAIC PERFUSION  187

MIXED DISEASE AND THE HEADCHEESE SIGN  191

THE MOSAIC ATTENUATION PATTERN: DIFFERENTIATION OF MOSAIC PERFUSION FROM GROUND-GLASS OPACITY  190

AIR TRAPPING ON EXPIRATORY HRCT  191

Abbreviations Used in This Chapter COPD CPTE DIP DLCO FEF FEV1 FVC GGO HU LIP PFT PH PI RB-ILD

chronic obstructive pulmonary disease chronic pulmonary thromboembolism desquamative interstitial pneumonia diffusing capacity forced expiratory flow forced expiratory volume in 1 second forced vital capacity ground-glass opacity Hounsfield units lymphoid interstitial pneumonia pulmonary function test pulmonary hypertension pixel index respiratory bronchiolitis-interstitial lung disease

In this chapter, we will discuss causes, patterns, and diseases associated with decreased lung attenuation, not due to lung destruction or cystic airspaces, which were reviewed in Chapter 6. These include mosaic perfusion, the mosaic attenuation pattern, the headcheese sign, and air trapping on expiratory scans.

MOSAIC PERFUSION Lung density and attenuation are partially determined by the volume of blood present within pulmonary vessels. Thus, regional differences in lung perfusion in patients with airways disease or pulmonary vascular disease can result in inhomogeneous lung opacity on high-­resolution computed tomography (HRCT) (1–5). Because this is often patchy or “mosaic” in distribution, with different regions of lung being of differing attenuation, it has

been termed mosaic perfusion (6) or mosaic oligemia (7), although the former term is most appropriate (8). This can result from airways or vascular disease. Areas of relatively decreased lung opacity seen on HRCT can be of varying sizes and sometimes appear to correspond to lobules, segments, lobes, or an entire lung (Figs. 7-1 to 7-5). In almost all cases, mosaic perfusion is seen in association with diseases causing regional decreases in lung perfusion. However, differences in attenuation between normal and abnormal lung regions recognizable on HRCT are accentuated by compensatory increased perfusion of normal or relatively normal lung areas. Mosaic perfusion is most frequent in patients with airways diseases that result in focal air trapping or poor ventilation of lung parenchyma (Figs. 7-1 to 7-4) (1–3); in these patients, areas of poorly ventilated lung are poorly perfused because of reflex vasoconstriction or because of a permanent reduction in the pulmonary capillary bed. In our experience, this finding has been most common in patients with bronchiolitis obliterans (constrictive bronchiolitis) (Figs. 7-2 to 7-4) or other diseases associated with small airways obstruction such as cystic fibrosis, infections, bronchiectasis, and cellular bronchiolitis, but it can also be seen as a result of large bronchial obstruction (9–11). Mosaic perfusion also occurs in association with pulmonary vascular obstruction such as that caused by chronic pulmonary embolism (Fig. 7-5) (7,12,13). Regardless of its cause, when mosaic perfusion is present, pulmonary vessels in the areas of decreased opacity usually appear smaller than vessels in relatively dense areas of lung (3,13) (Figs. 7-1 to 7-5). This discrepancy reflects differences in regional blood flow and can be quite helpful in distinguishing mosaic perfusion from patchy ground-glass opacity (GGO), which can have a similar appearance. In patients with ground-glass opacity, vessels usually appear equal in size throughout the 187

A

FIGURE 7-3  HRCT in a 9-year-old boy with postinfectious bronchiolitis obliterans. Patchy areas of mosaic perfusion are visible, with decreased vascular size within the lucent regions. Some lucent regions are lobular; this is typical of an airway abnormality.

B FIGURE 7-1  A and B: Mosaic perfusion in two patients with cystic fibrosis. In each patient, vessels appear larger in relatively dense lung regions, a finding of great value in making the diagnosis of mosaic perfusion. The relatively dense lung regions are normally perfused or overperfused because of shunting of blood away from the abnormal areas. Also note that abnormal airways (i.e., bronchiectasis, bronchial wall thickening, tree-in-bud) are visible in relatively lucent lung regions. These areas are poorly ventilated and poorly perfused.

A

FIGU RE 7-2  HRCT in a patient with bronchiolitis obliterans related to rheumatoid arthritis. Bronchiectasis is visible, along with patchy lung attenuation, a finding that reflects mosaic perfusion. Note that the pulmonary vessels in the lucentappearing peripheral left lung (black arrows) are smaller than vessels in the denser medial left lung (white arrows).

188

B FIGURE 7-4  A and B: HRCT in a bone marrow transplant recipient with bronchiolitis obliterans. Patchy areas of mosaic perfusion are visible, associated with findings of bronchiectasis. In patients with bronchiolitis obliterans, bronchiectasis is commonly visible.

C hapter 7 HRCT Findings: Decreased Lung Attenuation

189

in some patients with mosaic perfusion as the cause of inhomogeneous lung attenuation. In a blinded study by Arakawa et al. (15) of patients with inhomogeneous lung opacity of various causes, only 68% of patients with airways or vascular disease were believed to show small vessels in areas of low attenuation. In patients with inhomogeneous lung attenuation, if a confident diagnosis of mosaic perfusion or patchy ground-glass opacity cannot be made, the abnormal appearance may be referred to using the less specific term mosaic attenuation pattern (4), described in what follows. Decreased lung attenuation, when diffuse, can reflect panlobular emphysema or diffuse airways disease with air trapping (1,2). This distinction can be very difficult to make.

A

Mosaic Perfusion Due to Airways Disease In patients with mosaic perfusion resulting from airways disease, abnormally dilated or thick-walled airways (i.e., bronchiectasis) may be visible in the relatively lucent lung regions, thus suggesting the correct diagnosis (3,5,11,15). In one study (16), abnormal airways were seen in 70% of patients with airways disease and mosaic lung attenuation (Figs. 7-1, 7-2, and 7-4). Mosaic perfusion can be seen in a variety of airways diseases, including bronchiectasis, cystic fibrosis, hypersensitivity pneumonitis, and constrictive bronchiolitis. In patients with mosaic perfusion secondary to airways disease, lobular areas of low attenuation are common (Fig. 7-3); these are much less frequent in patients with vascular obstruction. Air trapping on expiratory scans, described later in this chapter, is often helpful in confirming the diagnosis.

B

Mosaic Perfusion Due to Vascular Disease

C FIGU RE 7-5  Mosaic perfusion with patchy lung attenuation in two patients with pulmonary embolism. A: Multidetector-row HRCT in a patient with extensive acute pulmonary embolism; patchy mosaic perfusion is visible. Vessels appear smaller in the lucent lung regions. B and C: In a patient with chronic pulmonary embolism, vessels appear smaller in the large low-attenuation regions in the peripheral lung. Large lucent regions are typical of CPTE.

lung. For example, in a series of 48 patients with mosaic perfusion primarily due to airways disease, Im et al. (14) observed smaller vessels in areas of low attenuation in 93.8% of cases. It must be pointed out, however, that decreased vessel size may be subtle and difficult to observe

#152201  Cust: LWW  Au: Webb  Pg. No. 189

Heterogeneous lung attenuation is common in patients with acute or chronic pulmonary thromboembolism (CPTE), and decreased vessel size in relatively lucent regions is often visible (Fig. 7-5B,C) (17,18). In a study of pulmonary parenchymal abnormalities in 75 patients with CPTE, 58 patients (77.3%) showed mosaic perfusion with normal or dilated arteries in areas of hyperattenuation (13); areas of relatively increased attenuation averaged −727 ­Hounsfield units (HU), whereas areas of decreased attenuation averaged −868 HU. In another study of patients with pulmonary hypertension (PH) due to CPTE, PH of other causes, and a variety of other pulmonary diseases, HRCT was believed to show mosaic perfusion in all patients with CPTE (19). Considerably more variation in vessel size in different lung regions was also visible in the patients with CPTE. Overall, HRCT had sensitivities of 94% to 100% and specificities of 96% to 98% in diagnosing CPTE (19). Mosaic perfusion is less frequent in patients with acute pulmonary embolism (Fig. 7-5A) (20,21). For example, Arakawa et al. (22) reported mosaic perfusion in 47% of patients with pulmonary embolism, most of whom had

DESIGN SERVICES OF

190

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

acute pulmonary embolism; in most of these patients, the mosaic perfusion was likely related to bronchoconstriction and air trapping. Mosaic perfusion may also be seen in patients with large-vessel vasculitis resulting in pulmonary artery stenosis. These include Takayasu arteritis and giant cell arteritis (23). The frequency with which mosaic perfusion is seen on CT in patients with various causes of PH has been studied by Sherrick et al. (24). Of 23 patients with PH caused by vascular disease, 17 patients (74%) had mosaic perfusion; 12 of these had chronic pulmonary embolism. Of 21 patients with PH associated with lung disease, 1 patient (5%) had mosaic perfusion. Among 17 patients with PH caused by cardiac disease, 2 patients (12%) had mosaic perfusion (24). In patients with vascular disease as a cause of mosaic perfusion, areas of low attenuation are usually larger than lobules, representing segments, lobes, or larger nonanatomic lung regions (e.g., the peripheral lung) (Fig. 7-5). In patients with mosaic perfusion occurring in association with CPTE, enlargement of the main pulmonary arteries may be visible because of PH (see Chapter 22).

THE MOSAIC ATTENUATION PATTERN: DIFFERENTIATION OF MOSAIC PERFUSION FROM GROUND-GLASS OPACITY The presence of inhomogeneous lung attenuation on HRCT is a common finding; in one study, inhomogeneous lung opacity was the predominant HRCT abnormality in 19% of scans reviewed (15). This appearance

can be a diagnostic dilemma, resulting from several possible causes. These include (a) ground-glass opacity, (b) mosaic perfusion resulting from airways obstruction and reflex vasoconstriction, (c) mosaic perfusion resulting from vascular obstruction, or (d) a combination of these (i.e.,  mixed disease). If the cause of the inhomogeneous lung opacity is not readily apparent, it may be described using the terms mosaic pattern or mosaic attenuation pattern (4,25). However, most cases of inhomogeneous opacity can be correctly classified as one of these four possibilities based on HRCT findings (15,16). On inspiratory scans, it is often possible to distinguish between ground-glass opacity, mosaic perfusion caused by airways disease, and mosaic perfusion caused by vascular disease (Fig. 7-6). In two studies (15,16), an accurate distinction was possible in more than 80% of cases based on HRCT findings. The use of expiratory scanning, described in the next section, is of further value in making this distinction. The most important HRCT finding in determining the presence of mosaic perfusion is reduced vessel size in lucent lung regions. If reduced vessel size is visible in lucent regions, a confident diagnosis of mosaic perfusion can usually be made. Also, in patients with mosaic perfusion, some lung regions may appear too lucent to be normal, but this is somewhat subjective and based on experience with the window settings used for scan viewing. In patients with mosaic perfusion resulting from airways disease, abnormally dilated or thick-walled airways (i.e., bronchiectasis) may be visible in the relatively lucent lung regions (Figs. 7-1 to 7-3), suggesting the proper

Inhomogeneous lung opacity (mosaic attenuation pattern)

Decreased vessel size Some lung regions too lucent No reticulation or CL nodules

Vessels of uniform size Some lung regions too dense Associated reticulation or CL nodules

Mosaic perfusion

Ground-glass opacity

Dilated PA Large areas of lucency

Abnormal airways Lobular lucencies

Vascular disease

Airways disease

Chronic PE

Small airways disease Large airways disease

Ground-glass opacity Differential

FIGU RE 7-6  Algorithmic approach to inhomogeneous lung opacity. CL, centrilobular; PE, pulmonary embolism; PA, pulmonary artery.

#152201  Cust: LWW  Au: Webb  Pg. No. 190

DESIGN SERVICES OF

C hapter 7 HRCT Findings: Decreased Lung Attenuation

diagnosis; this is visible in approximately 70% of cases and can be very helpful in the diagnosis (10,26–29). Furthermore, lobular areas of lucency are common in patients with airways disease (Fig. 7-3). In a study by Im et al. (14) of 48 consecutive patients with lobular areas of low attenuation seen on HRCT, 46 (95%) had symptoms related to respiratory disease, such as productive cough (n = 25) and hemoptysis (n = 18). Only two patients with this appearance, one with CPTE and one with Takayasu arteritis combined with bronchiectasis, had pulmonary vascular disease. In patients with vascular obstruction (e.g., CPTE) as a cause of mosaic perfusion, dilatation of central pulmonary arteries may be present as a result of PH, lobular areas of lucency are typically absent, and larger areas of low attenuation are usually visible (Fig. 7-5B,C). Ground-glass opacity may be accurately diagnosed as the cause of inhomogeneous lung opacity if it is associated with other findings of infiltrative disease such as consolidation, reticular opacities (i.e., the crazy-paving pattern), or nodules (see Chapter 5). Also, a pattern in which the areas of higher attenuation are centrilobular almost always represents ground-glass opacity with a centrilobular distribution. This pattern is not seen with mosaic perfusion resulting from airways disease; it is very uncommonly the result of vascular disease with mosaic perfusion. Ground-glass opacity may also result in very illdefined and poorly marginated areas of increased opacity, lacking the sharply marginated and geographic appearance sometimes seen in patients with mosaic perfusion. Ground-glass opacity can often be diagnosed simply because the lung looks too dense, although this is quite subjective and depends on using consistent window settings and being familiar with the appearance of normal lung parenchyma.

MIXED DISEASE AND THE HEADCHEESE SIGN In occasional patients with the mosaic attenuation pattern, inspiratory scans show a patchy pattern of variable lung attenuation, representing the combination of ground-glass opacity (or consolidation) and reduced lung attenuation as a result of mosaic perfusion. This combination of mixed attenuation, including the presence of mosaic perfusion, often gives the lung a geographic appearance and has been termed the headcheese sign because of its resemblance to the variegated appearance of a sausage made from the chopped parts of the head of a hog or other animal (Figs. 7-7 to 7-9) (30,31). If you can be sure that both ground-glass opacity or consolidation and mosaic perfusion are visible (rather than one or the other), the headcheese sign is present. Air trapping is commonly visible on expiratory scans when the headcheese sign is seen (Fig. 7-9). The headcheese sign is often indicative of mixed infiltrative and obstructive disease, usually associated with bronchiolitis (30,31). In patients with this appearance, the presence of ground-glass opacity or consolidation

#152201  Cust: LWW  Au: Webb  Pg. No. 191

191

is caused by lung infiltration, whereas the presence of mosaic perfusion with decreased vessel size is usually caused by small airway obstruction. The most common causes of this pattern (Table 7-1) are hypersensitivity pneumonitis (Figs. 7-8A and 7-9), sarcoidosis, atypical (viral or mycoplasma) infections with associated bronchiolitis (Fig. 7-7), desquamative interstitial pneumonia (DIP) (Fig. 7-8B) or respiratory ­bronchiolitis-interstitial lung disease (RB-ILD) associated with ground-glass opacity and bronchiolar obstruction, and sometimes lymphoid interstitial pneumonia (LIP) with follicular bronchiolitis; occasionally, pulmonary edema may result in this pattern. Each of these diseases results in an infiltrative abnormality and may be associated with airway obstruction. This combination may also be seen in patients with two abnormalities, an infiltrative process coexisting with an airways disease, such as asthma.

AIR TRAPPING ON EXPIRATory HRCT Obtaining HRCT scans at selected levels after expiration may be useful in (a) the diagnosis of air trapping in patients with obstructive lung disease, including chronic obstructive pulmonary disease (COPD) ­(15,28,29,32–34), (b) the diagnosis of airways disease unassociated with distinct morphologic abnormalities on inspiratory images (35), (c) distinguishing mosaic perfusion from ground-glass opacity (15,16), and (d) allowing the diagnosis of mixed infiltrative and obstructive diseases (Fig. 7-9) (15,36,37).

Diagnosis of Air Trapping Expiratory HRCT scans have proved useful in the evaluation of patients with a variety of lung diseases characterized by obstruction of airflow (10,11). Air trapping visible using dynamic expiratory or postexpiratory HRCT techniques has been recognized in patients with emphysema (27,38,39), COPD (32–34), asthma (40–44), cystic fibrosis (45), bronchiolitis obliterans and bronchiolitis obliterans syndrome (2,5,27,35,46–55), the cystic lung diseases associated with Langerhans histiocytosis and tuberous sclerosis (56), bronchiectasis (27,57), airways disease related to AIDS (58), and small airways disease associated with thalassemia (59). Expiratory HRCT has also proved valuable in demonstrating the presence of bronchiolitis in patients with primarily infiltrative diseases such as hypersensitivity pneumonitis (37,60–62), sarcoidosis (36,63), and pneumonia. Air trapping may also be seen in patients with acute pulmonary embolism or CPTE; Arakawa et al. (22) reported this finding in 60% of patients with pulmonary embolism, presumably due to bronchoconstriction. It must be understood that limited air trapping can be seen in normal subjects, particularly in the superior segments of the lower lobes, in scattered secondary lobules, and in dependent lung regions. Abnormal air trapping differs from this in extent. Expiratory CT techniques and normal findings are described in ­Chapters 1 and 2.

DESIGN SERVICES OF

192

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

A

C

B

D

FIGU RE 7-7  A–D: Headcheese and the headcheese sign. A and B: A slice of headcheese shown in color and black and white has a variegated appearance, consisting of chunks of different meats from the head of a hog. In the black-and-white image, some areas appear dark, some appear light, and some are gray. C and D: Mycoplasma pneumonia with the headcheese sign. Inspiratory images show inhomogeneous lung attenuation consisting of ground-glass opacity and multiple lobular areas of lucency due to mosaic perfusion, secondary to bronchiolitis. Note small or invisible vessels in lucent regions. Air trapping was present on expiratory scans.

Lung Attenuation Abnormalities in the Diagnosis of Air Trapping In normal subjects, lung increases significantly in attenuation during expiration (Figs. 2-30 to 2-32; Fig. 7-10). In the presence of airway obstruction and air trapping, the lung remains lucent on expiration and shows little change in a cross-sectional area. Areas of air trapping are seen as relatively low in attenuation on expiratory scans. On expiratory HRCT, the diagnosis of air trapping is easiest to make when the abnormality is patchy in distribution, and normal lung regions can be contrasted with abnormal, lucent lung regions (Figs. 7-11 to 7-15) (3,11). Areas of air trapping can be patchy and nonanatomical; can correspond to individual secondary pulmonary ­lobules, segments, and lobes; or may involve an entire lung

#152201  Cust: LWW  Au: Webb  Pg. No. 192

(26,64). Air trapping in a lobe or lung is usually associated with large airway or generalized small airway abnormalities, whereas lobular or segmental air trapping is associated with diseases that produce small airway abnormalities (26). Pulmonary vessels within the low-attenuation areas of air trapping often appear small relative to vessels in the more opaque normal lung regions (26). In patients with airways disease or emphysema who have a diffuse abnormality, expiratory heterogeneity in lung attenuation may not be visible, but air trapping can be detected by measuring the degree of lung attenuation change occurring with expiration (10,26–29,54). Areas of air trapping show significantly less attenuation increase than seen in normal lung on expiratory scans (50). The  normal mean attenuation difference

DESIGN SERVICES OF

C hapter 7 HRCT Findings: Decreased Lung Attenuation

193

A A

B FIGU RE 7-8  The headcheese sign. A: In a patient with hypersensitivity pneumonitis, lobular areas of lucency (red arrows) reflect mosaic perfusion associated with cellular bronchiolitis and bronchiolar obstruction. Patchy areas of ground-glass opacity and ground-glass opacity centrilobular nodules are visible. B: Ground-glass opacity and lobular mosaic perfusion in a patient with DIP.

B FIGURE 7-9  The headcheese sign and air trapping in hypersensitivity pneumonitis. A: Lobular areas of lucency reflect mosaic perfusion associated with cellular bronchiolitis and bronchiolar obstruction. Diffuse ground-glass opacity is present, with some centrilobular nodules visible. B: On a postexpiratory scan, air trapping is seen within the lucent lobules. Lung showing ground-glass opacity increases in attenuation. TABLE 7-1  Differential Diagnosis of Mixed Disease (the Headcheese Sign) Diagnosis

Comments

Hypersensitivity pneumonitis

HP the most common cause; lobular MP and air trapping with patchy or centrilobular GGO; MP and air trapping also seen with fibrotic disease

Atypical pneumonia

Patchy GGO or consolidation with MP and air trapping; viral and mycoplasma infections most typical

Sarcoidosis

MP and air trapping commonly seen; GGO uncommon

DIP and RB-ILD

Air trapping in a few percent of patients, associated with GGO and cysts

LIP

Patchy ground-glass may be seen with LIP; follicular bronchiolitis may result in MP or air trapping

Pulmonary edema

An uncommon manifestation; air trapping may reflect a coexistent airway abnormality or “cardiac asthma”

#152201  Cust: LWW  Au: Webb  Pg. No. 193

DESIGN SERVICES OF

194

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

A

B

FIGU RE 7-10  Normal postexpiratory HRCT. Inspiratory image (A) shows homogeneous lung attenuation. B: After expiration, there has been a significant reduction in lung volume associated with an increase in lung attenuation. Lung attenuation remains homogeneous. Note flattening of the posterior tracheal membrane.

A

B

FIGURE 7-11  Inspiratory (A) and expiratory (B) HRCT in a patient with postinfectious bronchiolitis obliterans. A: On inspiration, the lungs appear heterogeneous in

attenuation due to mosaic perfusion. B: On expiration, marked inhomogeneity in lung attenuation is noted, with multifocal air trapping. Many regions of air trapping appear to be lobular.

A

B

FIGU RE 7-12  Mosaic perfusion and air trapping in a patient with chronic airway infection and bronchiolitis. A: Inspiratory HRCT shows large regions of mosaic

perfusion in the upper lobes. B: Image from a low-dose dynamic expiratory HRCT shows air trapping in the regions that appeared lucent on the inspiratory scan. Note bowing of the posterior tracheal membrane due to expiration.

#152201  Cust: LWW  Au: Webb  Pg. No. 194

DESIGN SERVICES OF

C hapter 7 HRCT Findings: Decreased Lung Attenuation

A

195

B

FIGU RE 7-13  Expiratory air trapping in a patient with bronchiolitis obliterans. A: Inspiratory scan is normal. B: Postexpiratory scan shows patchy lung attenuation with the relatively lucent regions representing regions of air trapping. Normally ventilated areas increase significantly in attenuation on expiration.

A

B

FIGU RE 7-14  Postexpiratory air trapping in a patient with asthma. A: An inspiratory scan is normal. B: Routine postexpiratory scan obtained during suspended respiration after a forced exhalation scan shows patchy air trapping.

B

A

FIGU RE 7-15  Postexpiratory air trapping in a patient with bronchiolitis obliterans related to smoke inhalation. A: An inspiratory scan is normal. B: A low-dose dynamic expiratory scan shows patchy air trapping. Note anterior bowing of the posterior tracheal membrane, a good indication of forceful exhalation.

between full inspiration and expiration usually ranges from 80 to 300 HU. On dynamic scans, a lung attenuation change of less than 70 or 80 HU between full inspiration and full exhalation may be regarded as abnormal (Fig.  7-16). On simple postexpiratory scans, a lung attenuation change of less than 70 HU sometimes may be seen in normals. Lung attenuation change is most simply measured using small (1–2 cm) regions of interest on both inspiratory and expiratory scans (28). Measuring the change in overall lung attenuation from inspiration to expiration may be used in patients with diffuse air trapping (28) but is clearly less sensitive in

#152201  Cust: LWW  Au: Webb  Pg. No. 195

patients with patchy disease. In a study by Berger et al. (65), the difference between mean lung attenuation on inspiration and expiration was significantly larger in nonsmokers (128 HU) than in ex-smokers (77 HU) or current ­smokers (67 HU). A second method is to compare equivalent areas in each lung on expiratory scans. In healthy subjects, the mean difference in attenuation change between symmetric regions of the right and left lungs during exhalation was measured as 36 ± 14 HU (66). From this, a rightleft difference in attenuation increase during exhalation exceeding 78 HU (more than three standard deviations

DESIGN SERVICES OF

196

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

A

FIGU RE 7-16  Dynamic expiratory HRCT in a patient with cystic

fibrosis obtained using an electron-beam scanner. A: Six dynamic images from a sequence of 10, through the right upper lobe region, shown sequentially in a clockwise fashion from the upper left to lower left. On inspiration (top middle), lung opacity appears homogeneous. On expiration (lower left corner), a part of the anterior segment shows a normal increase in opacity, whereas the remainder of the upper lobe remains lucent. B: Timeattenuation curve measured in a lucent region of the upper lobe shows little change in attenuation during expiration.

B

greater than the mean) can be considered abnormal. This method is especially useful when air trapping is unilateral. Occasionally, lung attenuation decreases during expiration in regions of air trapping; a decrease of attenuation by as much as −258 HU has been reported during dynamic expiration (27). Although there is no definite

#152201  Cust: LWW  Au: Webb  Pg. No. 196

explanation for this phenomenon, several suggestions have been made (27). The most likely is that during exhalation, lung units trapping air compress small pulmonary vessels, squeezing blood out of the lung and decreasing lung perfusion. Another possible explanation is so-called pendelluft, in which air may pass from a normally ventilated lung unit

DESIGN SERVICES OF

C hapter 7 HRCT Findings: Decreased Lung Attenuation

to a partially obstructed lung unit during rapid expiration, resulting in an increased gas volume (26). Although measurement of lung attenuation can be used to diagnose air trapping, except in patients with diffuse air trapping (e.g., COPD, emphysema, large bronchial obstruction), the extent of air trapping rather than overall lung attenuation better predicts pulmonary function test (PFT) findings of obstruction (26,27).

Air-Trapping Score in the Diagnosis of Air ­Trapping The extent of air trapping present on expiratory scans can be measured using a semiquantitative scoring system, which estimates the percentage of lung that appears abnormal on each scan (5,27–29,43,57,59,66–68). Such systems have the advantage of being simple, quick, and easy to perform at the time of image interpretation. Furthermore, in one study (67), a simple 5-point scoring system was found to be associated with better interobserver agreement than a more detailed scoring system. Because the distribution of areas of air trapping can be heterogeneous, Bankier et al. (69) attempted to determine the number of expiratory slices necessary to accurately assess air trapping in patients with suspected bronchiolitis obliterans. Overall, the extent of air trapping increased from the upper to lower lung regions, with significant differences between regions (p < 0.001). It was found that expiratory imaging at fewer than three levels could show a result not representative of the overall extent of air trapping (69), although results varied in individual patients. In the scoring system proposed by Webb et al. (66) and Stern et al. (27), estimates of air trapping were made at three levels scanned using expiratory technique (at the aortic arch, carina, and 5 cm below the carina). At each level and for each lung, a 5-point scale is used to estimate the extent of air trapping visible subjectively: 0 = no air trapping; 1 = 1% to 25% of cross-sectional area of lung affected; 2 = 26% to 50% of affected lung; 3 = 51% to 75% of affected lung; 4 = 76% to 100% of affected lung. The air-trapping score is the sum of these numbers for the three levels studied and ranges from 0 to 24. In several studies using this method, significant differences were found in the extent of air trapping in normal patients and those with airway obstruction, and significant correlations were found between the extent of air trapping and PFT measures of airway obstruction (15,27,28). Other methods of visually scoring the extent of air trapping on expiratory scans have been used and validated (29,57). Lucidarme et al. (29) and Lee et al. (68) used a grid superimposed on the expiratory HRCT image and counted the number of squares containing lucent lung and the number encompassing the entire lung. The air-trapping score represented the ratio of air-trapping squares to the total number of squares overlying the lung and approximated the cross-sectional percent of abnormal lung. Excellent interobserver agreement was achieved using this method (29).

#152201  Cust: LWW  Au: Webb  Pg. No. 197

197

In patients studied using postexpiratory HRCT, correlations between the air-trapping score and various PFT findings of obstruction range from approximately r = −0.4 to r = −0.6 (15,28,59,68); correlations are generally best when normal and abnormal patients are grouped together and when patients with emphysema are included among those with airway obstruction (28). Thus, in a study by Chen et al. (28), considering only patients with obstructive disease, air-trapping score correlated significantly with forced expiratory volume in 1 second (FEV1) (r = −0.78), FEV1/forced vital capacity (FVC) (r = −0.64), FVC (r = −0.61), and forced expiratory flow (FEF) at 25% to 75% of vital capacity (r = −0.65); when both normal and abnormal patients were considered together, correlations were higher, with r values measuring −0.89, −0.74, −0.77, and −0.81, respectively. In a study by Lucidarme et al. (29) of 74 patients with suspected chronic airways disease, expiratory air trapping was seen in 18 of 35 (51%) patients with severe airway obstruction (FEV1/FVC < 80%), in 21 of 29 (72%) patients with predominantly small airway obstruction (abnormal flow-volume curve and FEV1/FVC < 80%), and in 4 of 10 (40%) patients with normal PFT results. Air-trapping scores were 27%, 12%, and 8% for these groups, respectively, with significant negative correlations with FEV1 (r = −0.45), FEV1/FVC (r = −0.31), and FEF at 25% of vital capacity (r = −0.57). Lee et al. (68) studied 47 asymptomatic subjects using PFTs and expiratory HRCT; in all, PFTs were considered to be normal. In this study, the air-trapping grade correlated with FEV1/ FVC (r = −0.44). In a study of 70 patients with chronic purulent sputum production (57), the air-trapping score defined at a lobular level significantly correlated with values of FEV1 and FEV1/FVC. In a study of 33 patients developing bronchiolitis obliterans syndrome after hematopoietic stem cell transplantation, postexpiratory air trapping was the principal finding seen on CT, and its severity correlated with PFTs. In this study, HRCT scans were visually ranked for degree of air trapping and also scored for findings of bronchial wall thickening, bronchiectasis, and centrilobular opacities (5). The degree of air trapping on expiratory HRCT correlated significantly with FEV1 (r = −0.52; p = 0.002), FEV1/ FVC (r = −0.57; p < 0.001), residual volume (r = −0.62; p < 0.001), and carbon monoxide diffusion capacity (p = 0.023). ­Bronchial wall thickening occurred in 73%, predominantly in lower lobes (p = 0.007), but was mild. Bronchiectasis occurred in 42.4% and centrilobular opacities in 39.4%.

Lung Area Changes in the Diagnosis of Air Trapping Decreased reduction in lung area with expiration correlate with the presence of air trapping and lung volume change, but area change measurements are less easily obtained during the clinical interpretation of expiratory HRCT.

DESIGN SERVICES OF

198

s e c t i o n II Approach to HRCT Diagnosis and Findings of Lung Disease

Robinson and Kreel (70) showed that a significant correlation exists between changes in cross-sectional lung area measured using CT and lung volume (r = 0.569). The percentage decrease in lung cross-sectional area that occurred during exhalation also correlates with the attenuation increase (66,70). In a study using dynamic ultrafast HRCT (66), a significant correlation between cross-sectional lung area and lung attenuation was found for each of three lung regions evaluated (upper lung: r = 0.51, p = 0.03; mid-lung: r = 0.58, p = 0.01; lower lung: r = 0.51, p = 0.05). Usually, areas of air trapping show little or no area and volume change during exhalation and can help identify areas of air trapping. In one study of nine cases of SwyerJames syndrome (2), expiratory CT scans in areas of abnormal lung showed no significant lung volume change, and mediastinal shift toward the normal lung was also seen. In a study by Lucidarme et al. (29) of 74 patients with suspected chronic airways disease and 10 normal subjects, an area reduction score was measured, representing the reduction in cross-sectional lung area from inspiration to expiration. Area reduction scores were 18%, 30%, and 35%, respectively, for groups of patients with severe airway obstruction (FEV1/FVC < 80%), predominantly small airways obstruction (abnormal flow-volume curve and FEV1/FVC = 80%), and normal PFT results. In the normal subjects, the area reduction score was 43%. Area reduction score correlated significantly with all PFT indexes (r = 0.35–0.66) except total lung capacity.

Air Trapping in Normals Air trapping can be seen in normal subjects, although its extent is limited. Air trapping in one or more secondary pulmonary lobules is not uncommon. Also, focal areas of relative lucency can be seen in normal subjects on expiratory scans in the superior segments of the lower lobes, in the lingula or middle lobe, and in dependent lung at the lung bases (26,66,68,71–73). Tanaka et al. (71) found a frequency of air trapping of 64% in asymptomatic patients with normal pulmonary function. In a study by Chen et al. (28), focal areas of air trapping, including the superior segments of the lower lobes, were visible in 61% of patients having normal PFTs. In a study by Lee et al. (68), air trapping was seen in 52% of 82 asymptomatic subjects with normal PFTs. The frequency of air trapping increased with age (p