Ehsan Samei PhD, Presenter: Nothing to Disclose

Ananth Poolla MS, Abstract Co-Author: Nothing to Disclose

To evaluate the comparative performance of a high-fidelity liquid crystal display (LCD) and a cathode ray tube (CRT) for mammography applications, and to assess the impact of the LCD viewing angle on the detection accuracy.

The comparison used a monochrome CRT (Barco MGD 521) and a color LCD (IBM T-221). Ninety 1k x 1k ROIs were selected from a database of digital mammograms acquired with a digital mammography system (Fischer): 30 without any abnormality present, 30 with subtle masses, and 30 with microcalcification clusters. Lesions were at the approximate center of the ROIs. Using the known angular response of the LCD, two additional versions of the images were also created by manipulating the grayscale appearance of each image such that the image, when viewed from a direction normal to the display surface, would appear as if it is viewed from a 30˚ and 50˚ off-axis viewing angle. With post-processing presentation identical to those of the commercial mammography system, the randomized images were independently scored by 3 experienced mammographers for the presence of lesions using a 0-100 grading scale. Both displays were calibrated to the DICOM GSDF standard and readings were undertaken in a controlled ambient lighting condition. The scores were evaluated to compare the display modes using an unpaired t-test analysis assuming un-equal variance.

For masses and microcalcifications, the detection rate showed a 3% and a 2% drop from CRT to LCD, respectively. However, the differences were not statistically significant (p>0.5). The viewing angle data showed no statistically-discernable trends suggesting that within the 0-50˚ viewing angle range, the variation in the contrast response of the LCD has little or no impact on the detection of mammographic lesions in a controlled observer experiment.

While CRTs and LCDs have different display characteristics in terms of angular response and noise, those characteristics seem to have little influence on the detection of mammography lesions. The results suggest comparable performance in clinical implementations of the two devices.

E.S.,A.P.: This work was supported by a grant from the NIH (1 R44 CA099068) and additional support by Fischer Imaging Corporation.

Samei, E, Poolla, A, Comparative Evaluation of LCD and CRT Displays for Digital Mammography.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4418955.html