| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorial |
1 Department of Diagnostic Radiology, Yale University School of Medicine, 333 Cedar St, 2-332SP, New Haven, CT 06520.
Index terms: Computed tomography (CT), image display and recording • Computed tomography (CT), image processing, 761.12111, 761.12114, 761.12115 • Editorials • Liver neoplasms, CT, 761.12111, 761.12114, 761.12115, 761.33 • Liver neoplasms, secondary, 761.33
 |
Introduction |
|---|
 TOP Introduction References |
|---|
The standard soft-tissue display window maps an attenuation range encompassing intraabdominal fat and contrast material–enhanced tissues to discrete gray levels and is useful for surveying the entire abdomen. Findings may be recognized in the mesenteric fat, as well as the contrast-enhanced hepatic parenchyma. However, hepatic lesions with attenuation characteristics that differ minimally from those of the surrounding hepatic parenchyma may be difficult to recognize with this window setting, owing to minimal gray-level contrast. The liver window maps a subset of this attenuation range to the 256 shades of gray. The window level is set equal to the attenuation of the hepatic parenchyma, and a narrow window width is chosen to encompass the attenuation range expected for lower attenuation and higher attenuation lesions. Thus, hepatic lesions may be displayed with greater gray-level contrast and be more conspicuous with a liver window than with a soft-tissue window.
In this issue of Radiology, Mayo-Smith et al (1) discuss the added utility of liver windows for detecting hepatic disease in a routine clinical setting. The authors do not dispute the improved gray-level contrast for hepatic lesion detection with liver windows. Rather, they evaluate the clinical benefit of this improved gray-level contrast in their practice. Within 8 months, the authors evaluated 1,175 consecutive abdominal CT scans and found additional lesions with liver windows in 36 patients (3.1%). A change in diagnosis occurred with use of liver windows in 20 patients (1.7%).
Liver windows are likely to be of particular value in detecting small lesions with attenuation characteristics that are not substantially different from those of background hepatic parenchyma. When the lesion diameter is less than the effective section thickness, volume averaging of the lesion with adjacent normal liver elevates the attenuation of the lesion and decreases its contrast with background tissue, thereby decreasing its conspicuousness. These presumptions hold true in the study by Mayo-Smith et al, as the average size of additional liver lesions detected with liver windows was only 8 mm with use of 5–10-mm-thick sections.
Another feature that will decrease the conspicuousness of hepatic lesions is the vascularity of the lesions relative to the phase of contrast enhancement (2). In the study by Mayo-Smith et al, all scans were obtained in the portal venous phase of enhancement when hypervascular tumors may become inconspicuous. Although liver windows may be of particular benefit under these circumstances, the majority of additional lesions detected with liver windows in this study were hypovascular.
With any imaging test, the benefits of modifications that improve diagnostic performance must be weighed against the costs of these modifications. To put the results of Mayo-Smith et al in perspective, Urban et al (3) previously noted a 7% (21 of 318) increase in liver lesion detection when spiral CT images were reconstructed with 50% overlap. The cost associated with doubling the number of reconstructed images for a 7% gain in liver lesion detection is probably warranted in patients for whom preoperative localization of hepatic tumors is desired prior to resection. However, the material and labor costs associated with doubling the number of reconstructed images is substantial. These include the film and archive media, the technologist's time to reconstruct and print the images, and the radiologist's time to review the images. In addition, there results a decrease in throughput owing to the increased examination time required to reconstruct and print images with a 50% overlap.
Given these costs, most practices that use film probably do not follow this procedure on a routine basis. Yet, liver windows are used routinely in many practices in spite of a 3% gain in liver lesion detection (1). Although the costs of this practice are not as great as those associated with reconstruction of images with 50% overlap, the costs may be substantial, particularly if the images are printed manually or the display window is adjusted individually.
Mayo-Smith et al recognize a low but finite benefit of using liver windows and conclude that routine use may be reserved for patients with an oncologic history or a high clinical suspicion of hepatic disease. As no patients with a history of trauma, rule out abscess, rule out hemorrhage, or rule out aneurysm or diverticulitis had additional benefit with liver windows, the authors suggest that liver windows may be omitted in such patients without a high clinical suspicion of hepatic disease. Although this approach is reasonably well supported, I have observed personally that radiologists tend to be partial to their style of practice in this regard. Those who routinely use liver windows favor them strongly, whereas those who do not rely on liver windows seem equally partial to the use of standard soft-tissue windows for detecting both intra- and extrahepatic disease.
This partisan approach may relate to beliefs grown from years of film-based image viewing. Traditionally, practices that use film have printed images at all window settings that may be required at image review; many routinely use soft-tissue, lung, bone, and liver windows for all abdominal CT examinations. The material and labor costs associated with this practice are not trivial, and the belief that liver windows are necessary in all cases may be, in part, a self-fulfilling prophecy. Although the material and technologist's labor costs are virtually eliminated with soft-copy image review on an imaging workstation, some labor costs remain associated with the time required for a physician to review images at multiple window settings. However, the ability to interactively manipulate display window settings on soft-copy review stations may improve image review beyond that afforded with standard window settings. The radiologist may manipulate the display window setting to a more "optimal" setting depending on the patient's body habitus and the degree of contrast enhancement (4).
Of course, manipulation of the display window setting implies that the full 12 bits of attenuation data have been preserved at the imaging workstation. Some lossy compression strategies perform 12- to 8-bit conversion and are commonly used when transmitting CT images with teleradiology systems. Once attenuation data have been compressed to 8 bits, radiologists may alter only the brightness and contrast of the image display and not the display window width and level. As long as 12-bit CT data are preserved on an imaging workstation, the radiologist can tailor the display window to the characteristics of each abdominal CT examination, recognizing the clinical history and the magnitude of contrast enhancement.
|
Footnotes |
|---|
See also the article by Mayo-Smith et al (pp 601–604) in this issue.
Received November 20, 1998;
accepted November 23, 1999.
|
References |
|---|
|
TOPIntroductionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
