| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 All authors: Lynn Sage Comprehensive Breast Center, Northwestern Memorial Hospital, Chicago, IL 60611; and Department of Radiology, Northwestern University Feinberg School of Medicine, 251 E Huron St., Galter Pavilion, 13th Fl., Chicago, IL 60611.
Received August 10, 2005;
accepted after revision August 31, 2005.
Address correspondence to E. A. Berns
(eberns{at}radiology.northwestern.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Technologist study acquisition time from examination initiation to release of the screenee was measured for both screen-film and digital mammography (100 cases each) in routine clinical practice. The total interpretation time for screening mammography was also measured for 183 hard-copy screen-film cases and 181 soft-copy digital cases interpreted by a total of seven breast imaging radiologists, four experienced breast imagers, and three breast imaging fellows.
RESULTS. Screening mammography acquisition time averaged 21.6 minutes for screen-film and 14.1 minutes for digital, a highly significant 35% shorter time for digital than screen-film (p < 10-17). The average number of images per case acquired with digital mammography was higher than that for screen-film mammography (4.23 for screen-film, 4.50 for digital; p = 0.047). The total interpretation time averaged 1.4 minutes for screen-film mammography and 2.3 minutes for digital mammography, a highly significant 57% longer interpretation time for digital (p < 10-11). In addition, technical problems delaying interpretation were encountered in none of the 183 screen-film cases but occurred in nine (5%) of the 181 digital cases.
CONCLUSION. Compared with screen-film mammography, the use of digital mammography for screening examinations significantly shortened acquisition time but significantly increased interpretation time. In addition, more technical problems were encountered that delayed the interpretation of digital cases.
Keywords: breast cancer • digital imaging • digital mammography • mammography • screen-film mammography • screening
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although digital mammography offers the potential advantage over screen-film mammography of separating image acquisition, display, and storage [3, 4], the clinical advantage of digital over film has yet to be shown [5-9]. In addition, the effect of digital mammography on image acquisition and interpretation times in the clinical setting has not been studied to date.
A single study compared interpretation times of digital with screen-film mammography using a locally developed soft-copy review workstation and a retrospectively collected set of 63 cases enriched with biopsy-proven cancers and benign findings [10]. All cases had prior films that were digitized for display on the review workstation along with the current digital study. None of the reviewers had prior experience performing soft-copy evaluations of digital mammography. The authors of that study concluded that soft-copy digital display was unlikely to significantly change the accuracy or speed of interpretation. No attempt was made to study image acquisition times.
The purpose of this study was to compare the image acquisition and interpretation times of screening studies acquired with digital mammography and screen-film mammography. The cases used consisted of a nonoverlapping set of clinically acquired digital and screen-film mammograms. The accuracy of the interpretations of the two techniques was not measured.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Digital mammograms were obtained on a Senographe 2000D (GE Healthcare) digital mammography system. Screen-film examinations were performed on one of five screen-film units located in the same facility: three 800 T units (GE Healthcare), one DMR unit (GE Healthcare), and one Mammomat 3000 unit (Siemens Medical Solutions). All screenfilm images were processed on one of two dedicated mammography daylight processors (Kodak Miniloader 2000P, Kodak) located within 50 ft (15 m) of each mammography room. Three technologists experienced in acquiring both screen-film and digital mammography examinations participated in the timing study of both techniques.
For the study of image interpretation times, a research assistant timed the total time from the location of the images on the alternator or soft-copy review workstation to the completion of the mammography report. All screening examinations were interpreted in Batch mode. Screen-film images were preloaded on a mammography alternator, and digital images were preloaded on a review workstation that used Solaris software ([version 3.1] GE Healthcare). The same dictation and reporting system was used for both screen-film and digital mammography. Interpretation time was recorded by a research assistant, who played no other role in the interpretation task, for 183 screen-film cases and 181 nonoverlapping digital cases. All cases were routine screening examinations performed clinically at our site. Seven different Mammography Quality Standards Act (MQSA)-qualified radiologists participated as the interpreting physicians in the timing study, four having extensive experience with both screen-film and digital mammography and three breast imaging fellows with more limited experience with both techniques. Each radiologist interpreted between 13 and 40 screen-film examinations and between 13 and 30 digital examinations. Figure 1 shows each reviewer's experience interpreting screen-film and digital mammograms over the previous year. Other than the fellows, all radiologists had several years of experience interpreting both techniques.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Twelve (12%) of the 100 screen-film cases required more than four views, whereas 20 (20%) of the 100 digital cases required more than four views. The average number of screen-film images per case (4.23) was significantly less than the average number of digital images per case (4.50, p = 0.047). This difference in the number of images per case was primarily because the single digital mammography detector was smaller (19 x 23 cm) and more often required tiling to accommodate the entire breast in each projection than the screen-film mammography detector, which had both 18 x 24 cm and 24 x 30 cm cassettes.
Each examination in the interpretation timing study had between four and eight images for both screen-film and digital mammography. Twenty-three (13%) of the 183 screen-film cases had more than four films, and 20 (11%) of the 181 digital cases had more than four images. The mean number of films for the 183 screen-film examinations was 4.31; the mean number of images for the 181 digital examinations was 4.19. The difference in the number of films interpreted per examination was not statistically significant (p = 0.16). The fraction of cases with prior images did not differ significantly: 81% of the patients undergoing screen-film examination had previously obtained films, 95% of the patients undergoing digital examinations had prior digital images (p = 0.50), 73% of the patients undergoing digital imaging had prior screen-film images; and 22% of the patients undergoing digital imaging had both prior screen-film and prior digital images. The fraction of cases with repeated images and the fraction of cases requiring either clinical or technical recall did not differ significantly between screen-film and digital cases, indicating that the longer interpretation times required for digital were not due to these causes.
Total interpretation time averaged 1.4 minutes for screen-film mammography and 2.3 minutes for digital mammography, a 0.9-minute difference (57%), which was highly statistically significant (p = 3.4 x 10-12). Table 1 summarizes these results.
|
Each individual radiologist had a shorter interpretation time for screen-film than digital mammography, with differences between the two techniques ranging from 38% to 86% (Fig. 2). All individual reviewer time differences were statistically significant (p < 0.013), except for one experienced reviewer who was timed on only 13 screen-film and 13 digital cases (p = 0.13). The four experienced reviewers had an average interpretation time of 1.2 minutes per case for screen-film mammography and 2.0 minutes per case for digital mammography, a statistically significant 0.8-minute difference (64%) per case (p = 1.8 x 10-10). The three fellows had an average interpretation time of 1.8 minutes per case for screen-film mammography and 2.6 minutes per case for digital mammography, a statistically significant 0.8-minute difference (51%) per case (p = 3.2 x 10-5).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The time savings of digital mammography for screening examination acquisition does not necessarily apply to diagnostic mammography because often the duration of the diagnostic examination depends on factors beyond image acquisition, such as the need for the radiologist to complete interpretation before release of the patient. At our institution, the technologist is occupied with a single patient until the interpretation is complete, so speeding image acquisition does not necessarily translate into shorter diagnostic examination time for the patient or technologist.
Unlike a previous study of interpretation time [8], our study found a significantly longer time for interpretation of soft-copy digital images than hard-copy screen-film images. Several factors may account for our different result. First, our study involved actual clinical cases interpreted on a commercial review workstation, rather than a study set interpreted on an in-house-designed review workstation. Second, all of our reviewers had substantial experience interpreting digital images before the study rather than being new to the technique.
Third, even though Senographe 2000D digital images can be displayed in full resolution when one 4.4-megapixel image is displayed on each 5-megapixel monitor, clinical practice at our institution is to view each image in electronically magnified form (using Quadrant Zoom or 2x Zoom and Pan modes) to facilitate identification of microcalcifications on digital images. This is the digital equivalent of using a 1.5-2x handheld magnifier for screen-film images. Performing this electronically magnified search for microcalcifications on digital images requires more image manipulation and is more time consuming than holding a magnifier to review screen-film images. It is not clear whether it was routine to view each digital image in a magnified mode in the previous interpretation timing study [8].
Fourth, although the researchers of the previous study [8] had prior images for both digital and film interpretation, they presented film images (printed from digital) with prior films on the alternator and digitized prior film images for soft-copy display beside the current digital images. In short, films were displayed with prior films and soft-copy images were presented with prior soft-copy images. In our study, 81% of the screen-film cases with prior images all had prior screen-film images, and 95% of the digital cases with prior images all had prior films displayed on an alternator. Comparison with prior images was likely easier with film, where both film sets were mounted on an alternator, than with digital, where prior screen-film images on an alternator were compared with current digital images on a soft-copy review workstation, as our clinical practice requires.
Finally, 22% of our digital cases had prior images both on film and on digital soft copy. This required an additional comparison of current digital images with both prior films on an alternator and prior digital images on the soft-copy review workstation. For this subset of cases, the mean time for interpretation was 2.8 minutes versus 2.4 minutes for the subset of digital cases with priors only on screen-film. Thus, having prior images that had been obtained using both techniques added time compared with having prior images that had been obtained on film only, although this difference was not statistically significant.
Although these results apply to digital images acquired on the Senographe 2000D acquisition system and viewed on the review workstation, other manufacturers may have different results depending on the timing of their image acquisition and the detailed operation of their review workstation. It is likely, however, that the general results of this study will transfer to other fixed-detector digital systems used with soft-copy interpretation: Acquisition with fixed digital systems (not computed radiography systems) will be faster and soft-copy interpretation of digital images will be slower, at least until all relevant prior images are available in digital format. The first part of this statement is based on the fact that all fixed-detector digital acquisition systems eliminate processing, and therefore speed acquisition compared to film. The second part of this statement is based on the commonality of tasks required for soft-copy review of digital mammograms, including the need to examine each view in multiple display formats and the need to compare soft-copy digital with prior film studies.
These results suggest that review workstation manufacturers should improve functionality to make soft-copy interpretation of digital mammograms more efficient. In the meantime, because careful soft-copy interpretation of digital mammograms in clinical practice will incur longer interpretation times than screen-film screening studies, radiologists should adjust their time allocation for interpretation of digital studies.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. M. Nishikawa, S. Acharyya, C. Gatsonis, E. D. Pisano, E. B. Cole, H. S. Marques, C. J. D'Orsi, D. M. Farria, K. M. Kanal, M. C. Mahoney, et al. Comparison of Soft-copy and Hard-copy Reading for Full-Field Digital Mammography Radiology, April 1, 2009; 251(1): 41 - 49. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. M. Haygood, J. Wang, E. N. Atkinson, D. Lane, T. W. Stephens, P. Patel, and G. J. Whitman Timed Efficiency of Interpretation of Digital and Film-Screen Screening Mammograms Am. J. Roentgenol., January 1, 2009; 192(1): 216 - 220. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
