July 2001, VOLUME 177
NUMBER 1

Recommend & Share

July 2001, Volume 177, Number 1

Technical Innovation

Paddle-Wheel CT Display of Pulmonary Arteries and Other Lung Structures
A New Imaging Approach

+ Affiliation:
1 All authors: Department of Radiology, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Ave., Boston, MA 02215.

Citation: American Journal of Roentgenology. 2001;177: 195-198. 10.2214/ajr.177.1.1770195

Next section

Multidetector helical CT technology permits high-speed and high-resolution imaging data of the entire thorax to be acquired during a single breath-hold [1,2,3]. The lungs can be scanned in 12 sec with four detectors, each with collimation of 2.5 mm, or in 24 sec if 1.25-mm collimation is used [1,2,3]. Even faster systems are being developed [4]. State-of-the-art equipment already provides an almost cuboidal voxel size of about 1 × 1 × 1.25 mm. Thus, reconstructed nonaxial images have a resolution comparable with that of the best axial images [1,2,3].

The diagnostic value of high-speed CT can be enhanced by using the original axial helical CT data set to construct a new set of planar reconstructions arranged in a “paddle-wheel” pattern, in which all planes pass through a central horizontal axis between the two lungs and hilum. In this article, we introduce the paddle-wheel CT imaging method.

Paddle-Wheel Method
Previous sectionNext section

This reconstruction method uses a paddle-wheel arrangement, in which planar slabs pivot around a central horizontal axis between the lung hilum (Fig. 1A). Each slab is 15- to 20-mm thick, depending on the size of the patient. Both lungs can be covered completely using 20 reconstruction image slabs, with 9° rotations between successive slabs for a total of 180° (Fig. 1B). Because all planes pass through the central axis, the large central hilar structure of interest appears on every image. In addition, images from each plane also display the two specific sets of anatomic branches that happen to fan outward in opposite directions from that central structure at the same angle of rotation as the particular image plane. For anatomic correlation, we number the planes in a clockwise fashion on the lateral view, starting along the major fissures at an angle of approximately 45° from the horizontal.

figure
View larger version (79K)
Fig. 1A. Schematic drawings of paddle-wheel principles. Drawing shows how paddle-wheel slabs “pivot” on central hilar structure. All slabs will display this central structure from different angles whereas each individual slab will display those branches that fan outward in opposite directions at same angle as specific slab.

figure
View larger version (81K)
Fig. 1B. Schematic drawings of paddle-wheel principles. Drawing of lateral view of lungs with 20 superimposed numbered slabs providing complete coverage of both lungs. Numbering starts on major fissure at approximately 45° angle and runs clockwise. Preferred positions of central axes are at bifurcation of pulmonary artery trunk, A, for pulmonary arteries; at rear of left atrium, V, for pulmonary veins; and at tracheal carina, B, for central airways.

This display method provides close accord between the CT images and lung anatomy. Lobes, segments, fissures, branching airways, arteries, and veins appear in a natural sequence of 20 numbered images that follow their logical order and anatomic orientation, thus simplifying anatomic correlation (Fig. 1C). The upper half of each image obtained using this method displays a section of lobes above the major fissure in both lungs. The lower half shows a section of lower lobe segment of each lung.

figure
View larger version (60K)
Fig. 1C. Schematic drawings of paddle-wheel principles. Drawing of lateral view of lungs with superimposed numbered slabs corresponding to lung segments. A represents central axis of pulmonary arteries. Upper half of each image displays section of segment above major fissure in each lung. Lower half shows section in lower lobe segment of each lung. Typically, each segment has five slabs that converge toward hilum of lung.

Typically, each segment is covered by five slabs that converge toward the hilum of the lung. Specifically, the upper halves of slabs 1-5 display segments of the middle lobe and lingula, those of slabs 6-10 display the anterior segment, slabs 11-15 show the apical segment, and slabs 16-20 show the posterior segment of each upper lobe. The lower halves of slabs 1-5 show the superior segments of the lower lobes, whereas slabs 6-10 reveal the posterior segments, slabs 11-15 display the lateral and medial segments on the right but mainly the lateral on the left side (because of the heart), and slabs 16-20 show the anterior segments of both lower lobes. Obviously, an appropriate correction of the relationship between the image number and lung anatomy is necessary if the patient has undergone lung resection or experienced significant lung collapse or effusion.

The transhilar axis of rotation can be precisely positioned for the best display of specific anatomic structures. For the pulmonary arteries and lungs, the preferred axis of rotation is at the bifurcation of the pulmonary artery trunk (Figs. 1B and 2A,2B,2C,2D). For the pulmonary veins, the preferred axis of rotation is along the posterior margin of the left atrium where the veins converge (Fig. 1B). For the trachea and central airways, the most useful axis is on the carina (Fig. 1B). The different axes are all available from the original helical data set and can be easily located on the CT workstation using a left lateral scout image and appropriate axial and midline sagittal reference images (Fig. 1B).

figure
View larger version (147K)
Fig. 2A. Paddle-wheel CT angiographic series of pulmonary arteries in 40-year-old woman with shortness of breath. Possibility of pulmonary embolus had been excluded. Four typical images from full sequence of 20 images show appearance of pulmonary vessels on individual slabs. In practice, upper halves of each image are first examined in sequence, from front to back, to view all segments above major fissure. Then, lower halves are examined in sequence for all lower lobe segments, from back to front. Full sequence is normally viewed on monitor in cine mode.

figure
View larger version (161K)
Fig. 2B. Paddle-wheel CT angiographic series of pulmonary arteries in 40-year-old woman with shortness of breath. Possibility of pulmonary embolus had been excluded. Four typical images from full sequence of 20 images show appearance of pulmonary vessels on individual slabs. In practice, upper halves of each image are first examined in sequence, from front to back, to view all segments above major fissure. Then, lower halves are examined in sequence for all lower lobe segments, from back to front. Full sequence is normally viewed on monitor in cine mode.

figure
View larger version (172K)
Fig. 2C. Paddle-wheel CT angiographic series of pulmonary arteries in 40-year-old woman with shortness of breath. Possibility of pulmonary embolus had been excluded. Four typical images from full sequence of 20 images show appearance of pulmonary vessels on individual slabs. In practice, upper halves of each image are first examined in sequence, from front to back, to view all segments above major fissure. Then, lower halves are examined in sequence for all lower lobe segments, from back to front. Full sequence is normally viewed on monitor in cine mode.

figure
View larger version (184K)
Fig. 2D. Paddle-wheel CT angiographic series of pulmonary arteries in 40-year-old woman with shortness of breath. Possibility of pulmonary embolus had been excluded. Four typical images from full sequence of 20 images show appearance of pulmonary vessels on individual slabs. In practice, upper halves of each image are first examined in sequence, from front to back, to view all segments above major fissure. Then, lower halves are examined in sequence for all lower lobe segments, from back to front. Full sequence is normally viewed on monitor in cine mode.

IV contrast medium is required for imaging intraluminal abnormalities of the pulmonary arteries and veins. We routinely use a single bolus of 75 mL of contrast medium (320 mg/mL), injected at a rate of 3.5 mL/sec. A test injection with a small bolus is normally used to determine the optimal imaging delay for maximum pulmonary artery enhancement; the optimal delay is usually between 10-20 sec, depending on the patient's cardiac function. Maximum-intensity-projection reconstructions, which depict the highest pixel values [5], are excellent for showing the smaller intrapulmonary vessels (Fig. 2A,2B,2C,2D). Typically, the larger central vessels are displayed well on the standard axial images with average window settings, but they can also be displayed by obtaining additional thin paddle-wheel slabs with average window settings.

For visualizing the major airways, the inherent contrast with the surrounding mediastinal tissues is sufficient to produce clear views of the extrapulmonary tracheobronchial tree using average lung window settings or minimum-intensity-projection settings, which depict the lowest pixel values. Small intrapulmonary bronchi are also best visualized by the use of minimum-intensity-projection settings, particularly if silhouetted by peribronchial disease.

When viewing the images, branching structures can generally be followed without interruption from the hilum to the pleura on any single image or perhaps on two adjacent images. Because the 15- to 20-mm slabs increasingly overlap as they approach the axis of rotation, the larger central structures are displayed on multiple slabs, each viewed from a slightly different projection angle. The 20 images that cover the entire lung volume can be displayed in their natural anatomic order on two standard-sized radiographs or even on a single radiograph using smaller images (Fig. 2A,2B,2C,2D). However, the series of images is ideally viewed back and forth on a monitor in cine fashion.

Advantages and Potential Disadvantages
Previous sectionNext section

Technically, the paddle-wheel method can be easily implemented on all helical CT equipment. However, the improved resolution and speed of multidetector scanners now produce images of distinctly superior quality [4]. The image reconstruction protocol is simple and is easily learned by technologists familiar with standard CT reconstruction methods. In our experience, the protocol is rapidly learned and takes only a few minutes to perform. Fewer representative axial images need to be documented on radiographs, and the use of parallel sagittal, coronal, or oblique reconstructions can be minimized or eliminated (Fig. 3A,3B). As a result, there is a potential for substantial reduction of image reconstruction time and film costs. Less contrast medium is required because the scan time associated with fast multidetector CT systems is shorter.

figure
View larger version (116K)
Fig. 3A. Comparison of anterior and posterior lung regions in 53-year-old man with shortness of breath. Possibility of pulmonary embolus was excluded. Paddle-wheel reconstruction (A) and conventional coronal CT reconstruction (B). With paddle-wheel reconstructions, pulmonary vessels are displayed in continuity from hilum to lung periphery. Conventional reconstructions display vessels in fragments.

figure
View larger version (116K)
Fig. 3B. Comparison of anterior and posterior lung regions in 53-year-old man with shortness of breath. Possibility of pulmonary embolus was excluded. Paddle-wheel reconstruction (A) and conventional coronal CT reconstruction (B). With paddle-wheel reconstructions, pulmonary vessels are displayed in continuity from hilum to lung periphery. Conventional reconstructions display vessels in fragments.

Diagnostically, the primary advantage of the paddle-wheel display is that individual slab reconstructions are anatomically more meaningful and easier to interpret. Branching structures are displayed in continuity from hilum to pleura. Smaller arteries and veins are thus easily distinguished by their connection to their respective larger central vessels. Segmental anatomy is preserved. The relationship between a small lung nodule and an adjacent artery, vein, or bronchus is usually clear. Because anatomic relationships are now obvious, time required for interpretation should be reduced. The possibility of abnormalities being hidden by overlapping central structures is reduced because closely spaced slabs display these structures from multiple angles of rotation. This feature is an advantage when compared with conventional angiography, in which large central vessels are frequently superimposed, potentially obscuring small lesions.

This new method has some potential disadvantages. It certainly works best with relatively costly state-of-the-art multidetector technology. However, the paddle-wheel principle may also offer some benefit with less sophisticated equipment. In some practice settings, cine viewing may not always be available. As with axial helical CT, skillful selection of appropriate Hounsfield levels and windows remains important. For a short time, comparisons of the paddle-wheel reconstruction images with previously obtained parallel CT images depicting fragmented anatomic structures may be difficult. Finally, with multidetector studies, there is a potential for a slight increase in radiation dose to the patient, depending on the scan parameters chosen. These disadvantages must be weighed against the substantial benefits in image quality in future studies.

Conclusion
Previous sectionNext section

We have now instituted the paddle-wheel method of CT reconstruction imaging of the lungs as a routine supplement to the standard axial multidetector CT pulmonary angiography images for pulmonary embolism. We are exploring its role in imaging other disorders, including airway disease. We anticipate that large-scale clinical comparison studies of the traditional and the new paddle-wheel CT imaging methods will be undertaken by us and by radiologists in other clinical centers to quantify the relative clinical effectiveness of each method.

Address correspondence to M. Simon.

We thank Alesia Napier and her talented team of CT radiology technologists who have adopted the method as part of the routine imaging of pulmonary embolism. We also thank Robert Carr for help with the diagrams and Claire Martinez, Nancy Williams, and Diane Pliner for clerical assistance.

References
Previous sectionNext section
1. Hu H. Multi-slice helical CT: scan and reconstruction. Med Phys 1999; 26:5-18 [Google Scholar]
2. Taguchi K, Aradate H. Algorithm for image reconstruction in multi-slice helical CT. Med Phys 1998; 25:550-561 [Google Scholar]
3. Rubin GD, Shiau MC, Schmidt AJ, et al. Computed tomographic angiography: historical perspective and new state-of-the-art using multi detector-row helical computed tomography. J Comput Assist Tomogr 1999; 23[suppl 1]:S83-S90 [Google Scholar]
4. Rubin GD, Shiau MC, Leung A, et al. Aorta and iliac arteries: single versus multiple detector-row helical CT angiography. Radiology 2000; 215:670-676 [Google Scholar]
5. Prokop M, Oh S, Schanz A, Schaefer-Prokop CM. Use of maximum-intensity projections in CT angiography: a basic review. RadioGraphics 1997; 17:433-451 [Google Scholar]

Recommended Articles

Paddle-Wheel CT Display of Pulmonary Arteries and Other Lung Structures

Free Access, , ,
American Journal of Roentgenology. 2001;177:136-138. 10.2214/ajr.177.1.1770136
Citation | Full Text | PDF (2320 KB) | PDF Plus (2371 KB) 
Free Access, , , , , , ,
American Journal of Roentgenology. 2001;177:173-175. 10.2214/ajr.177.1.1770173
Citation | Full Text | PDF (1902 KB) | PDF Plus (1940 KB) 
Free Access,
American Journal of Roentgenology. 2001;177:221-227. 10.2214/ajr.177.1.1770221
Citation | Full Text | PDF (7265 KB) | PDF Plus (7325 KB) 
Free Access, , , , ,
American Journal of Roentgenology. 2014;202:479-492. 10.2214/AJR.13.11463
Abstract | Full Text | PDF (1418 KB) | PDF Plus (1472 KB) 
Free Access, ,
American Journal of Roentgenology. 2001;177:95-98. 10.2214/ajr.177.1.1770095
Citation | Full Text | PDF (5605 KB) | PDF Plus (5683 KB) 
Free Access,
American Journal of Roentgenology. 2001;177:53-59. 10.2214/ajr.177.1.1770053
Citation | Full Text | PDF (6878 KB) | PDF Plus (1043 KB)