Comparative performance of modern digital mammography systems in a large breast screening program. Issue 12 (27th November 2013)

• Record Type: Journal Article
• Authors: Yaffe, Martin J.; Bloomquist, Aili K.; Hunter, David M.; Mawdsley, Gordon E.; Chiarelli, Anna M.; Muradali, Derek; Mainprize, James G.
• Is Part Of: Medical physics. Volume 40:Issue 12(2013)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2013-11-27
• Subjects: Digital mammography -- Computed radiography -- Digital radiography -- Cancer; biological tissues -- cancer -- diagnostic radiography -- dosimetry -- image resolution -- mammography -- phantoms -- transfer functions; digital mammography -- computed radiography -- image quality -- breast imaging; Biological material, e.g. blood, urine; Haemocytometers; Image detection systems -- Medical X‐ray imaging -- Image sensors -- X‐ray detectors -- Film mammography -- Medical image noise -- Digital mammography -- Cancer; Medical physics -- Periodicals; Medical physics; Geneeskunde; Natuurkunde; Toepassingen; Biophysics; Periodicals; Periodicals; Electronic journals; 610.153
• DOI: 10.1118/1.4829516
• Languages: English
• ISSNs: 0094-2405

Abstract

Abstract : Purpose: : To compare physical measures pertaining to image quality among digital mammography systems utilized in a large breast screening program. To examine qualitatively differences in these measures and differences in clinical cancer detection rates between CR and DR among sites within that program. Methods: : As part of the routine quality assurance program for screening, field measurements are made of several variables considered to correlate with the diagnostic quality of medical images including: modulation transfer function, noise equivalent quanta, d′ (an index of lesion detectability) and air kerma to allow estimation of mean glandular dose. In addition, images of the mammography accreditation phantom are evaluated. Results: : It was found that overall there were marked differences between the performance measures of DR and CR mammography systems. In particular, the modulation transfer functions obtained with the DR systems were found to be higher, even for larger detector element sizes. Similarly, the noise equivalent quanta, d′, and the phantom scores were higher, while the failure rates associated with low signal‐to‐noise ratio and high dose were lower with DR. These results were consistent with previous findings in the authors' program that the breast cancer detection rates at sites employing CR technology were, on average, 30.6% lower than those that used DR mammography. Conclusions: : While the clinical study was not large enough to allow a statistically powered system‐by‐system assessment of cancer detection accuracy, the physical measures expressing spatial resolution, and signal‐to‐noise ratio are consistent with the published finding that sites employing CR systems had lower cancer detection rates than those using DR systems for screening mammography.