(Radiographics. 2001;21:1257-1273.)
© RSNA, 2001
Multi–Detector Row CT of Thoracic Disease with Emphasis on 3D Volume Rendering and CT Angiography1
Leo P. Lawler, MD and
Elliot K. Fishman, MD
1 From the Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University, 601 N Caroline St, Rm 3254, Baltimore, MD 21287. Presented as an education exhibit at the 2000 RSNA scientific assembly. Received January 24, 2001; revision requested February 20 and received March 9; accepted March 16. Address correspondence to E.K.F. (e-mail: efishman@jhmi.edu).
 |
Abstract
|
|---|
Multi–detector row computed tomography (CT) with three-dimensional (3D) volume rendering provides a unique perspective on thoracic anatomy and disease. Multi–detector row CT allows shorter acquisition times, greater coverage, and superior image resolution. Three-dimensional volume rendering now permits real-time, interactive modification of relative pixel attenuation in an infinite number of planes and projections. In vascular imaging, this technique provides image quality that equals or surpasses that of conventional angiography. Its use has expanded to aid in diagnosis and surgical planning, often obviating conventional or digital angiography and reducing costs. It is reliable in depicting clot and the pulmonary vasculature and may also be used to evaluate thoracic venous anomalies (eg, pulmonary arteriovenous malformations) and to plan therapy. Airway imaging with multi–detector row CT with 3D volume rendering is particularly useful in the planning and follow-up of stent placement. In diffuse lung disease, this technique can increase nodule detection and help differentiate between small nodules and vessels. It is also helpful in imaging the musculoskeletal system and the thoracic cage. Multi–detector row CT with 3D volume rendering has enhanced the conventional roles of thoracic CT and challenged the supremacy of other imaging modalities. It will likely play a leading role in future radiologic research and practice.
Index Terms: Computed tomography (CT), angiography, 56.12116, 56.12117 • Computed tomography (CT), clinical effectiveness • Computed tomography (CT), image processing, 474.12117, 56.12117, 60.12117 • Computed tomography (CT), volume rendering, 474.12117, 56.12117, 60.12117 • Thorax, anatomy, **.922 • Thorax, CT, 474.12116, 474.12117, 56.12116, 56.12117, 60.12116, 60.12117
|
Introduction
|
|---|
Among the various thoracic anatomic structures and the diseases that affect them, there now remain few entities that can elude the combined potential of multi–detector row computed tomography (CT) and three-dimensional (3D) volume rendering. Technologic progress in this area has created the capacity for fast acquisition of unprecedented, high-quality data sets and the potential to view them in 3D in a practical, real-time, and powerful way (1,2). In many cases, although axial images may be adequate for diagnosis, volume-rendered 3D images can provide more information about the nature of the disease and better communication with clinicians.
In this article, we discuss and illustrate the application of multi–detector row CT with 3D volume rendering in the normal anatomy and related diseases of the thorax, which we have arbitrarily divided into the vasculature, lung, and chest wall. We will show how this technique has enhanced the conventional roles of thoracic CT, challenged the supremacy of other imaging modalities, and changed the radiologic approach to disease processes.
|
Imaging Technique
|
|---|
The protocols described in this article reflect our personal experience, which was gained largely with a Plus 4 Volume Zoom scanner by Siemens (Iselin, NJ) with an adaptive detector array design and highly flexible pitch. Our images are volume rendered on a prototype 3D Virtuoso workstation (Siemens), which is commercially available. When contrast material was required, we used (nonionic) Omnipaque 350 (Nycomed Amersham, Princeton, NJ).
The adaptive detector array design peculiar to Siemens scanners permits the user to choose from multiple detector sizes and to vary the pitch (travel per gantry rotation/section collimation) from 1 to 8. The choice of collimators will determine the range of section thicknesses that are available after the scan is performed. The main advantage of this design is the capacity to obtain thinner sections without compromising z-axis coverage. In many cases, multi–detector row CT with 3D volume rendering represents an advance over the classic two-dimensional (2D) study, and in all cases the nearly isotropic matching of in-plane resolution and section thickness means that alternatives to transaxial imaging are not only possible but, in many situations, more logical.
In pulmonary imaging, we routinely use 1- or 2.5-mm-thick detectors depending on the clinical indication. With 1-mm detectors we acquire 1.25-mm-thick sections, and with 2.5-mm detectors we obtain 3-mm-thick sections. We use a rotation time of 0.5 seconds and, most commonly, a pitch of 6. In the thorax, breath-hold imaging coverage and temporal resolution have been enhanced by the peculiar capacity of multi–detector row CT to achieve high pitch without widening the section thickness profile. With the capacity to cover a distance of 24 cm in 8 seconds, the need for hyperventilation techniques is eliminated, and the volume of intravenous contrast material and need for pediatric sedation are decreased (3).
We use various reconstruction or volume-rendering algorithms to highlight the structure or disease of interest. The 3D images are interactively displayed and edited in real time. Volume rendering applies shades of gray to the pixels of varying attenuation that are found in a data set, and trapezoidal transfer functions allow user modification of the relative contribution of various pixel values for the area of interest. Thus, all the information initially acquired is used for the final reconstruction. Unlike a threshold technique such as maximum intensity projection (MIP) or shaded surface display, volume rendering will take into account pixels that are only partially filled with contrast material (4,5). Although there are no data directly comparing it with other 3D techniques in thoracic applications, volume rendering has proved advantageous in nonthoracic applications (6,7). MIP images will not allow depth perception or understanding of interstructural relationships unless movable images are used (Fig 1). Likewise, all tissues will be represented based on their Hounsfield values so that, unlike with a threshold technique, simultaneous depiction of the vasculature and the airways on individual images is possible. Stereo imaging display and endoluminal perspectives may on occasion further enhance the appreciation of 3D relationships.
View larger version (163K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. Volume-rendered imaging versus MIP imaging. (a) Left lateral oblique volume-rendered image of the mediastinum demonstrates small coronary artery bypass vessels (thin solid arrows). The relationship of the descending thoracic aorta (thick solid arrow) to the left pulmonary artery branches (open arrow) can also be appreciated. (b) On a left lateral oblique MIP image of the mediastinum, the descending thoracic aorta is seen (arrow), but its relationship to the pulmonary artery cannot be appreciated. The small bypass vessels are lost in the attenuation of the heart chambers.
|
|
View larger version (146K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. Volume-rendered imaging versus MIP imaging. (a) Left lateral oblique volume-rendered image of the mediastinum demonstrates small coronary artery bypass vessels (thin solid arrows). The relationship of the descending thoracic aorta (thick solid arrow) to the left pulmonary artery branches (open arrow) can also be appreciated. (b) On a left lateral oblique MIP image of the mediastinum, the descending thoracic aorta is seen (arrow), but its relationship to the pulmonary artery cannot be appreciated. The small bypass vessels are lost in the attenuation of the heart chambers.
|
|
The images are immediately sent to the workstation as the patient is scanned. Depending on the institution, the images may be produced by the radiologist or by a trained technologist. There is a steep learning curve, and most training courses take 3 days. Time-consuming evaluation of regions of interest is not required with the use of clip editing planes, and there is true, real-time change of perspective with use of the mouse, allowing the 3D images to be produced in a timely and efficacious manner. For example, we will routinely produce four or five images of the thoracic aorta in under 10 minutes along with a conventional axial imaging study, so that 3D information is available the same day if required. Multi–detector row CT with volume rendering is now sufficiently flexible that there are many more tailored, high-quality techniques for chest imaging than are necessary in any one patient, so protocols have been designed to standardize our procedure in each case (Table).
|
Clinical Applications
|
|---|
Thoracic Vasculature
Multi–detector row CT with 3D volume rendering probably enjoys the widest clinical acceptance in vascular imaging (8). Whether normal anatomic structures and their variants (Figs 2 –5) (9) or the diseases that affect them are at issue, image quality equals or surpasses that of conventional angiography, which had been regarded as the standard of reference. Because volume rendering faithfully incorporates all the acquired data, one may see vessels that would otherwise be hidden with threshold-based shaded surface display or MIP techniques, both of which sacrifice some data. MIP imaging studies also lack 3D information unless motion display is used.
View larger version (174K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a. Aortic nipple. Superior axial (a) and sagittal (b) volume-rendered images show the left superior intercostal vein (arrow) extending between the left brachiocephalic vein and the posterior mediastinum.
|
|
View larger version (165K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b. Aortic nipple. Superior axial (a) and sagittal (b) volume-rendered images show the left superior intercostal vein (arrow) extending between the left brachiocephalic vein and the posterior mediastinum.
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. Aberrant right subclavian artery. Superior axial (a) and coronal (b) volume-rendered images demonstrate an aberrant right subclavian artery arising as the last artery of the aortic arch and extending behind the trachea and esophagus (arrow).
|
|
View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. Aberrant right subclavian artery. Superior axial (a) and coronal (b) volume-rendered images demonstrate an aberrant right subclavian artery arising as the last artery of the aortic arch and extending behind the trachea and esophagus (arrow).
|
|
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. Normal anatomy mimicking disease in a patient undergoing evaluation for right paratracheal widening. (a) Coronal volume-rendered image shows a right paratracheal prominence caused by the azygos vein (arrow). (b) Inferior axial volume-rendered image helps confirm that the widening is due to the azygos vein (arrow). (c) Right sagittal volume-rendered image shows the azygos vein as it arches forward to join the superior vena cava (arrow). (d) Coronal volume-rendered image demonstrates the azygos vein (arrow).
|
|
View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. Normal anatomy mimicking disease in a patient undergoing evaluation for right paratracheal widening. (a) Coronal volume-rendered image shows a right paratracheal prominence caused by the azygos vein (arrow). (b) Inferior axial volume-rendered image helps confirm that the widening is due to the azygos vein (arrow). (c) Right sagittal volume-rendered image shows the azygos vein as it arches forward to join the superior vena cava (arrow). (d) Coronal volume-rendered image demonstrates the azygos vein (arrow).
|
|
View larger version (142K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5c. Normal anatomy mimicking disease in a patient undergoing evaluation for right paratracheal widening. (a) Coronal volume-rendered image shows a right paratracheal prominence caused by the azygos vein (arrow). (b) Inferior axial volume-rendered image helps confirm that the widening is due to the azygos vein (arrow). (c) Right sagittal volume-rendered image shows the azygos vein as it arches forward to join the superior vena cava (arrow). (d) Coronal volume-rendered image demonstrates the azygos vein (arrow).
|
|
View larger version (150K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5d. Normal anatomy mimicking disease in a patient undergoing evaluation for right paratracheal widening. (a) Coronal volume-rendered image shows a right paratracheal prominence caused by the azygos vein (arrow). (b) Inferior axial volume-rendered image helps confirm that the widening is due to the azygos vein (arrow). (c) Right sagittal volume-rendered image shows the azygos vein as it arches forward to join the superior vena cava (arrow). (d) Coronal volume-rendered image demonstrates the azygos vein (arrow).
|
|
Aorta.—Soon after spiral CT was introduced, its role in noninvasive imaging of the aorta became clear (10). Subsequent studies and reviews confirmed this role and suggested improvements that would likely accompany the development of multisection imaging (11–13). The use of 3D volume rendering has expanded to meet needs in diagnosis and surgical planning (14). In many cases, this obviates conventional angiography, and cost reduction has been demonstrated (15).
In cases of aortic aneurysm or dissection, we routinely achieve 1.25-mm section thickness with four 1-mm collimators and reconstruct the images at 1-mm intervals. When required, imaging from the arch to the bifurcation is possible with administration of a single bolus of contrast material. The temporal resolution now eliminates artifact from aortic pulsation, and misregistration from respiration and cardiac pulsation is minimized. The caliber and course of a normal and diseased aorta and branch vessels are well depicted, as is their internal architecture (Figs 6–8). The infinite internal and external imaging perspectives that are possible with administration of a single bolus of contrast material constitute a significant advantage over conventional angiography (16). By modifying attenuation, one can optimize visualization of either the vasculature or of the viscera they supply.
View larger version (154K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6. Aortic dissection. Right sagittal oblique volume-rendered image shows a type A dissection arising at the origin of the left subclavian artery (short solid arrow) and extending into the descending thoracic aorta. Both true (open arrow) and false (long solid arrow) lumina are patent on either side of the intimal flap.
|
|
View larger version (141K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7a. Aortic transection. (a) Axial volume-rendered image reveals deformity of the midthoracic descending aorta (long arrow) with an associated mediastinal hematoma (short arrow). (b) Sagittal reconstructed image shows acute angulation of the disrupted aortic wall (arrow).
|
|
View larger version (150K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7b. Aortic transection. (a) Axial volume-rendered image reveals deformity of the midthoracic descending aorta (long arrow) with an associated mediastinal hematoma (short arrow). (b) Sagittal reconstructed image shows acute angulation of the disrupted aortic wall (arrow).
|
|
View larger version (129K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8a. Aortic pseudoaneurysm. (a) Right sagittal volume-rendered image demonstrates a calcified, contained pseudoaneurysm arising from the concavity of the aortic arch (arrow). (b) Right midline sagittal volume-rendered image shows the calcified pseudoaneurysm in the inferior aortic arch (long arrow) with central flow (short arrow).
|
|
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8b. Aortic pseudoaneurysm. (a) Right sagittal volume-rendered image demonstrates a calcified, contained pseudoaneurysm arising from the concavity of the aortic arch (arrow). (b) Right midline sagittal volume-rendered image shows the calcified pseudoaneurysm in the inferior aortic arch (long arrow) with central flow (short arrow).
|
|
In the setting of thoracic aneurysms, volume-rendered imaging can enhance understanding of vessel dilatation, mural thrombus, and branch vessel patency compared with axial imaging alone. Volume-rendered imaging has become nearly indispensable for evaluation of endovascular stent placement in abdominal aneurysms and will likely also become so for interventional management of thoracic aneurysms. It may also prove beneficial in postoperative evaluation of the aorta because it clearly depicts stent placement (17).
The extent of an intimal flap and its relationship to arch vessels as well as the patency of both the true and false lumina are clearly defined with imaging planes customized to the aortic course and angioscopic views (Fig 6) (16,17–20). CT has proved more sensitive and specific than transesophageal echocardiography or magnetic resonance imaging (21), and, although 3D image quality may be improved with multi–detector row CT, there are as yet no data showing that it has any greater sensitivity in detecting intimal tears.
The ability to produce high-quality images of the lung parenchyma and chest wall from the same data set is invaluable in the emergency setting in patients with suspected thoracic aortic trauma (Fig 7). Three-dimensional imaging has also been suggested as being more sensitive than multiplanar reconstruction for the detection of pseudoaneurysms (Fig 8) (18), although axial imaging will often suffice.
Coronary Arteries.—Quantification of coronary calcium was introduced by Agatston et al (22) and was shown to be related to the presence of coronary artery disease. The finding that lipid-lowering drugs could alter the calcified plaque burden (23) and possibly change the natural history of atherosclerosis intensified efforts to detect and quantify coronary calcification. High correlation was noted between findings at conventional CT and those at electron-beam CT (24), leading to the use of spiral imaging. Cardiac-gated multi–detector row CT is proving a worthy competitor to electron-beam CT in this setting (Fig 9).
View larger version (133K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9a. Arterial calcification. (a) Superior 2D MIP image shows calcified right coronary (R), circumflex (C), and left coronary (L) arteries. (b) Superior volume-rendered image demonstrates the calcified left anterior descending (short arrow) and circumflex (long arrow) arteries.
|
|
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9b. Arterial calcification. (a) Superior 2D MIP image shows calcified right coronary (R), circumflex (C), and left coronary (L) arteries. (b) Superior volume-rendered image demonstrates the calcified left anterior descending (short arrow) and circumflex (long arrow) arteries.
|
|
The high temporal resolution of 250 msec with 500-msec rotation times, multisection acquisition, and electrocardiographic gating decrease the effects of cardiac and respiratory motion and may reduce interstudy variability. There is still a need for large, prospective cohort studies to evaluate whether this technique should be implemented as a cost-effective screening tool in addition to conventional risk factor assessment (25). Multi–detector row CT with 3D volume rendering will likely play a leading role in research involving coronary CT angiography and imaging of noncalcified, lipid-laden plaques (26).
Pulmonary Embolism.—Pulmonary embolism accounts for 10% of all hospital deaths and is a contributing factor in another 10% (27), yet 50% of pulmonary emboli are not diagnosed antemortem (28). Part of the problem is that our diagnostic tests are simply not good enough. Approximately 70%–75% of ventilation-perfusion scans are considered indeterminate, and oftentimes, "problem-solving" pulmonary angiography is recommended but not performed (29). In addition, problem-solving conventional angiography is increasingly being questioned as a true standard of reference (30).
Electron-beam CT and spiral CT have proved reliable in depicting clot (31,32) and the pulmonary vasculature (33), even in patients in whom thromboembolism is not clinically suspected (34,35). In 1997, Mayo et al (36) demonstrated that CT rivaled scintigraphy and Remy-Jardin et al (37) showed that it rivaled conventional angiography in detecting pulmonary thromboembolism. Although there is ongoing debate concerning the methods used in such studies, CT pulmonary angiography has been established as a quick, reliable, safe, and cost-effective technique (38) and was incorporated into the latest Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study. In cases in which pulmonary thromboembolism is ruled out, CT is helpful in identifying alternative pathologic conditions that may have similar manifestations.
There is better visualization of smaller pulmonary artery branches with a smaller collimation of 2 mm (39). Multi–detector row CT with 1-mm collimation and a pitch of 6 allows coverage from the aortic arch to the diaphragm in 10 seconds. Single–detector row spiral CT will take 20 seconds for such z-axis coverage with 2-mm collimation. Thus, the limitation of single–detector row CT of not depicting clot in the smaller segmental and subsegmental branches should be improved upon. When a very quick study is required (eg, in a poor breath holder), we use 2.5-mm collimation to rule out clot in larger vessels. If the clinical question concerns segmental or subsegmental clot, 1-mm collimation is used. Temporal resolution at multi–detector row CT limits artifacts, especially in lower-lobe pulmonary artery branches near the heart. The duration of the breath hold needed for single–detector row spiral CT is reduced by almost one-half for multi–detector row CT, and the high pitch means that simultaneous CT venography of the lower extremities is feasible (40). Although good axial imaging will suffice in the majority of cases, the ability to render 3D images helps in viewing those vessels that course obliquely through the imaging plane and in following vessels to their origin to differentiate between arteries and veins (Fig 10) (36). Also, we find that defining the extent of clot in a vessel is useful for understanding the clot burden and for planning surgery in chronic thromboembolic pulmonary hypertension (Fig 11).
View larger version (91K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10a. Pulmonary arteries. (a) Axial 2D image shows a questionable area of low attenuation in the tortuous branching regions at the origin of the right middle lobe arteries (left arrow). The lingular branches are poorly defined (right arrow). (b) On a superior axial volume-rendered image, the contrast material-filled right middle lobe vessels are seen in their entirety (long white arrow), and the lingular vessels (short white arrow) are clearly distinguishable from the pulmonary veins (black arrow). S = superior vena cava.
|
|
View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10b. Pulmonary arteries. (a) Axial 2D image shows a questionable area of low attenuation in the tortuous branching regions at the origin of the right middle lobe arteries (left arrow). The lingular branches are poorly defined (right arrow). (b) On a superior axial volume-rendered image, the contrast material-filled right middle lobe vessels are seen in their entirety (long white arrow), and the lingular vessels (short white arrow) are clearly distinguishable from the pulmonary veins (black arrow). S = superior vena cava.
|
|
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Pulmonary embolism. (a) Axial 2D image shows acute low-attenuation clot as a filling defect in the contrast material-filled right and left pulmonary arteries (arrows). (b) Anterior volume-rendered image depicts acute clot in the right pulmonary artery extending into upper and lower branches (arrow).
|
|
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Pulmonary embolism. (a) Axial 2D image shows acute low-attenuation clot as a filling defect in the contrast material-filled right and left pulmonary arteries (arrows). (b) Anterior volume-rendered image depicts acute clot in the right pulmonary artery extending into upper and lower branches (arrow).
|
|
Venous Anatomy.—Multi–detector row CT with 3D volume rendering has been used for venous angiography of the liver and kidney, especially when planning tumor resection or transplant surgery. Similarly, the venous structures of the thorax may be visualized, and many of the advantages of this technique for arterial imaging discussed earlier apply to venous imaging as well. The inherently tortuous, variable-caliber nature of veins makes them well suited for volume imaging.
Superior vena cava obstruction, often related to tumor, is frequently seen at multi–detector row CT with 3D volume rendering. This technique may be used to establish the extent of tumor involvement and document the extent of collateral vessel formation (Figs 12, 13). Similarly, other causes of venous obstruction such as thoracic outlet syndrome and subclavian thrombosis can be evaluated (41). Multi–detector row CT with 3D volume rendering is valuable in studies of venous anomalies such as duplications and pulmonary arteriovenous malformations and in planning therapy (Fig 14). CT with shaded surface display has been shown to be accurate in defining the angioarchitecture of pulmonary arteriovenous malformations, and volume rendering has improved on these results, allowing better depiction of the smaller-caliber structures (42,43). The ability to interactively change projections to evaluate specific vessels allows classification of malformations as simple or complex and is helpful in making management decisions (43).
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13. Superior vena cava obstruction in a patient with clot involving the azygos and left brachiocephalic veins as well as the superior vena cava. Coronal 3D volume-rendered image shows extensive pericardial collateral vessels (arrows).
|
|
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 14. Pulmonary arteriovenous malformation in a patient with Osler-Weber-Rendu disease. Inferior axial volume-rendered image reveals a vascular nidus in the right lung (long black arrow) with large feeding and draining vessels (short arrows).
|
|
Airways and Lung Parenchyma
The ability to perform multiplanar imaging of the lung with nearly identical image quality regardless of the plane chosen is useful in imaging of the airways and lung parenchyma. Likewise, the ability to select either less noisy, thicker sections for survey studies or thinner high-resolution images for evaluating diffuse disease is of particular interest. Chest coverage of 24 cm in 8 seconds at a pitch of 6 is of obvious benefit in poor breath holders.
Airways.—Since the demise of bronchography, whole-lung airway imaging has largely been limited to axial imaging at selected levels of interest. Attempts at 3D imaging have been hampered by poor data quality and suboptimal rendering techniques. Stair-step and other artifacts understandably led to some skepticism regarding the role of 3D imaging in evaluation of the thorax. However, high-quality, high-fidelity images can now be produced efficiently and have been used for problem solving in airway disease (44–46).
Unlike with single–detector row CT, the narrow collimation and overlapping reconstruction used in multi–detector row CT permit routine imaging beyond the central airways to the subsegmental bronchi of the peripheral airways (Fig 15). Airway disease often affects multiple disparate branches of the tracheobronchial tree. It is important to be able to perform whole-lung volume-rendered imaging with ease, as opposed to being confined to a small, preselected area of interest as in the past. Unlike "all-or-none" thresholding binary classification, volume rendering allows the percentage of different tissue types to be reflected in the final image while maintaining 3D spatial relationships. The trapezoidal transfer function creates a histogram of displayed Hounsfield values, permitting control over window width and level, attenuation, and brightness so that various combinations of tissues may be displayed. Whole-lung volume-rendered imaging is far less susceptible to volume averaging artifacts, which are a particular problem in airway imaging due to the air–soft tissue interface.
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15a. Normal airways. Axial (a), coronal (b), and sagittal (c) volume-rendered images demonstrate normal central and peripheral airways. Note the sheet overlying the chest wall.
|
|
View larger version (190K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15b. Normal airways. Axial (a), coronal (b), and sagittal (c) volume-rendered images demonstrate normal central and peripheral airways. Note the sheet overlying the chest wall.
|
|
View larger version (136K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15c. Normal airways. Axial (a), coronal (b), and sagittal (c) volume-rendered images demonstrate normal central and peripheral airways. Note the sheet overlying the chest wall.
|
|
The ability to visualize beyond severe narrowing and the ease with which small pediatric airways can be imaged with volume rendering makes the technique a complement to bronchoscopy (Figs 16, 17). A preprocedure bronchoscopic map has obvious potential in defining the number, location, and extent of lesions within an individual bronchus as well as the overall lobar and segmental distribution of areas of stenosis and bronchiectasis. Such luminal abnormalities of obliquely coursing airways are better seen on volume-rendered views. Functional imaging is possible, allowing determination of the airway caliber at various stages of the respiratory cycle. As with vascular structures, endoluminal views may also be generated. Anomalous branching of the airway is better appreciated on volume-rendered views, and simultaneous CT angiography may help clearly define the interrelationship of vascular structures in cases of rings and slings (Figs 16, 17). We have found airway imaging to be particularly useful in the planning and follow-up of stent placement (Fig 18) (47,48).
View larger version (80K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16a. Normal anatomic variant in a patient who was referred for evaluation of an impression noted on the right middle portion of the trachea. (a) Coronal volume-rendered image of the airway reveals an impression (arrow). (b) Inferior axial volume-rendered image reveals that the tracheal impression is caused by the right aortic arch (arrow).
|
|
View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 16b. Normal anatomic variant in a patient who was referred for evaluation of an impression noted on the right middle portion of the trachea. (a) Coronal volume-rendered image of the airway reveals an impression (arrow). (b) Inferior axial volume-rendered image reveals that the tracheal impression is caused by the right aortic arch (arrow).
|
|
View larger version (106K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18a. Tracheobronchial strictures due to acid ingestion. (a) Volume-rendered image demonstrates tracheobronchial strictures (short arrow) as well as a stent in the left main bronchus (long arrow). (b) Volume-rendered image shows a granulation shelf (arrow) over the stent opening in the left main bronchus. Interim placement of a new tracheal stent was performed.
|
|
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 18b. Tracheobronchial strictures due to acid ingestion. (a) Volume-rendered image demonstrates tracheobronchial strictures (short arrow) as well as a stent in the left main bronchus (long arrow). (b) Volume-rendered image shows a granulation shelf (arrow) over the stent opening in the left main bronchus. Interim placement of a new tracheal stent was performed.
|
|
Diffuse Lung Disease.—For diffuse lung disease, we have found that whole-lung imaging and interpretation facilitate understanding and communication of disease extent (Fig 19). MIP imaging has shown some merit in the evaluation of patients with emphysema (49,50), both in detecting disease and in describing its extent. The nearly isotropic data sets provide almost identical imaging quality in any plane; this is valuable in determining the distribution of diffuse lung diseases in which many of the axial landmarks of the airway and fissures are distorted. One can use larger 2.5-mm detectors for rapid lung evaluation at the cost of some spatial resolution or incorporate 1-mm detectors for high-resolution images (eg, in interstitial lung disease).
View larger version (170K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 19a. Abnormal lung parenchyma and airways in a patient with cystic fibrosis. MIP (a) and coronal volume-rendered (b) images demonstrate whole-lung distribution of bronchiectasis (thin arrow) and large bulla formation (thick arrow).
|
|
View larger version (178K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 19b. Abnormal lung parenchyma and airways in a patient with cystic fibrosis. MIP (a) and coronal volume-rendered (b) images demonstrate whole-lung distribution of bronchiectasis (thin arrow) and large bulla formation (thick arrow).
|
|
This article would not be complete without a comment on the current interest in lung cancer screening (51,52). It is not yet clear if the ability to screen for cancer will change clinical outcomes or if current CT technology is overly sensitive and lacking in specificity for many nodules. However, if a practical, high-quality test is sought, multi–detector row CT should be considered, and low-dose population studies may be feasible. Three-dimensional imaging has been shown to increase nodule detection and can help differentiate between small nodules and vessels (53,54).
Chest Wall
Multi–detector row CT with 3D volume rendering has proved worthwhile in musculoskeletal imaging and can also be applied to imaging of the thoracic cage (Fig 20) (55,56). This technique has been shown to have value in imaging congenital and postsurgical anomalies, and similar principles may be applied in imaging other chest wall conditions (57,58). Whether it be in the bone thorax or in the soft tissues it supports, we have applied these principles in imaging congenital changes (Figs 21, 22), tumor (Fig 23), trauma (Figs 24, 25), and infectious diseases (Figs 26, 27) with trapezoidal transfer functions to highlight structures of interest. Unlike MIP imaging, multi–detector row CT with 3D volume rendering preserves the 3D interrelationship of data. The complex curving, overlapping anatomy of the shoulder girdle and thoracic cage is well suited to 3D imaging. In the trauma setting, complicated chest wall fracture can be evaluated and CT aortography performed in a timely manner in a single study with this technique. CT is well established as a useful modality for lung cancer staging, and the question of chest wall involvement can be addressed with 3D imaging (58).
View larger version (142K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 20a. Normal thoracic cage. Coronal (a) and right lateral (b) volume-rendered images demonstrate the normal sternum, costochondral cartilages (arrow in a), and sternoclavicular joints.
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 20b. Normal thoracic cage. Coronal (a) and right lateral (b) volume-rendered images demonstrate the normal sternum, costochondral cartilages (arrow in a), and sternoclavicular joints.
|
|
View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 23a. Chest wall malignancy. Thoracic inlet (a) and coronal (b) volume-rendered images show a right posterior lung cancer extending to the chest wall and thoracic spine (arrow).
|
|
View larger version (51K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 23b. Chest wall malignancy. Thoracic inlet (a) and coronal (b) volume-rendered images show a right posterior lung cancer extending to the chest wall and thoracic spine (arrow).
|
|
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 25a. Clavicular trauma. Coronal (a) and thoracic inlet (b) volume-rendered images show a displaced, comminuted fracture of the left midclavicle with inferior displacement (arrow).
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 25b. Clavicular trauma. Coronal (a) and thoracic inlet (b) volume-rendered images show a displaced, comminuted fracture of the left midclavicle with inferior displacement (arrow).
|
|
View larger version (159K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 27a. Sternoclavicular joint infection in an immunocompromised oncology patient with a central catheter in the right internal jugular vein. (a) Axial 2D image shows an ipsilateral sternoclavicular infection causing bone destruction, sequestra, and soft-tissue swelling (arrow). (b) Thoracic inlet volume-rendered image depicts the right sternoclavicular infection (arrow).
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 27b. Sternoclavicular joint infection in an immunocompromised oncology patient with a central catheter in the right internal jugular vein. (a) Axial 2D image shows an ipsilateral sternoclavicular infection causing bone destruction, sequestra, and soft-tissue swelling (arrow). (b) Thoracic inlet volume-rendered image depicts the right sternoclavicular infection (arrow).
|
|
|
Conclusions
|
|---|
The latest advances in multi–detector row CT and 3D volume rendering have much to offer the thoracic imager. Volume rendering is suitable for exploiting the advantages of high-quality multi–detector row CT data and is a necessity due to the large data sets involved. Not only is this technique expanding the established conventional roles of thoracic CT, but it has shown great potential in a host of new applications.
|
Footnotes
|
|---|
** indicates multiple body systems. 
Abbreviations: MIP = maximum intensity projection,
3D = three-dimensional,
2D = two-dimensional
|
References
|
|---|
-
Hu H. Multi-slice helical CT: scan and reconstruction. Med Phys 1999; 26:5-18.[Medline]
-
Calhoun PS, Kuszyk B, Heath DG, Carley JC, Fishman EK. Three-dimensional volume rendering of spiral CT data: theory and method. RadioGraphics 1999; 19:745-764.[Abstract/Free Full Text]
-
Frush DP, Donnelly LF, Chotas HG. Contemporary pediatric thoracic imaging. AJR Am J Roentgenol 2000; 175:841-851.[Free Full Text]
-
Magnusson M, Lenz R, Danielsson PE. Evaluation of methods for shaded surface display of CT volumes. Comput Med Imaging Graph 1991; 15:247-256.[Medline]
-
Napel S, Marks MP, Rubin GD, et al. CT angiography with spiral CT and maximum intensity projection. Radiology 1992; 185:607-610.[Abstract/Free Full Text]
-
Johnson PT, Halpern EJ, Kuszyk B, et al. Renal artery stenosis: CT angiography—comparison of real-time volume rendering and maximum intensity projection algorithms. Radiology 1999; 211:337-343.[Abstract/Free Full Text]
-
Hong KC, Freeny PC. Pancreaticoduodenal arcades and dorsal pancreatic artery: comparison of CT angiography with three-dimensional volume rendering, maximum intensity projection, and shaded surface display. AJR Am J Roentgenol 1999; 172:925-931.[Abstract/Free Full Text]
-
Smith PA, Fishman EK. Clinical integration of three-dimensional helical CT angiography into academic radiology: results of a focused survey. AJR Am J Roentgenol 1999; 173:445-447.[Abstract/Free Full Text]
-
Hopkins KL, Patrick LE, Simoneaux SF, Bank ER, Parks WJ, Smith SS. Pediatric great vessel anomalies: initial experience with spiral CT angiography. Radiology 1996; 200:811-815.[Abstract/Free Full Text]
-
Costello P, Ecker CP, Tello R, Hartnell CG. Assessment of the thoracic aorta by spiral CT. AJR Am J Roentgenol 1992; 158:1127-1130.[Abstract/Free Full Text]
-
Quint LE, Francis IR, Williams DM, et al. Evaluation of thoracic aortic disease with the use of heli- cal CT and multiplanar reconstructions: comparison with surgical findings. Radiology 1990; 201:37-41.[Abstract/Free Full Text]
-
Ledbetter S, Stuk JL, Kaufman JA. Helical (spiral) CT in the evaluation of emergent thoracic syndromes: traumatic aortic rupture, aortic aneurysm, aortic dissection, intramural hematoma and penetrating atherosclerotic ulcer. Radiol Clin North Am 1999; 37:575-589.[Medline]
-
Rubin GD, Shiau MC, Leung AN, Kee ST, Logan LJ, Sofilos MC. Aorta and iliac arteries: single versus multiple detector-row helical CT angiography. Radiology 2000; 215:670-676.[Abstract/Free Full Text]
-
Wu CM, Urban BA, Fishman EK. Spiral CT of the thoracic aorta with 3-D volume rendering: a pictorial review of current applications. Cardiovasc Intervent Radiol 1999; 22:159-167.[Medline]
-
Rubin GD, Armerding MD, Dake MD, Napel S. Cost identification of abdominal aortic imaging by using time and motion analyses. Radiology 2000; 215:63-70.[Abstract/Free Full Text]
-
Kimura F, Shen Y, Date S, Azemoto S, Mochizuki T. Thoracic aortic aneurysm and aortic dissection: new endoscopic mode for three-dimensional CT display of the aorta. Radiology 1996; 198:573-575.[Abstract/Free Full Text]
-
Rubin GD, Dake MD, Semba CP. Current status of three-dimensional spiral CT scanning for imaging the vasculature. Radiol Clin North Am 1995; 33:51-70.[Medline]
Top
Abstract
Introduction
