Elizabeth Roy MD, Presenter: Nothing to Disclose

David Motiuk BS, Abstract Co-Author: Nothing to Disclose

Paul William Burrowes MD, Abstract Co-Author: Nothing to Disclose

Bobbie Docktor, Abstract Co-Author: Nothing to Disclose

The Calgary Health Region converted from screen-film mammography (SFM) to digital mammography (DM) in 2005. This retrospective study was designed to determine how this conversion has affected positive predictive values (PPVs) for cancerous and precancerous breast lesions.

In the Calgary region, biopsies for mammographic calcifications are only done at Foothills Medical Centre (FMC) by a small group of mammographers employing homogeneous biopsy techniques. From FMC’s breast cancer database, we reviewed core biopsy data performed for mammographic calcifications in the years 2002-2004 (SFM years) and 2008-2010 (DM years). Mammographic masses were excluded. We determined PPVs for each of SFM years and DM years for detection of cancerous lesions (PPV3 for calcifications). We calculated the PPVs of SFM and DM for detection of high-risk lesions, including ADH, ALH, LCIS and papilloma collectively (precancerous lesions). The detection rates of benign lesions (excluding precancerous lesions) after biopsy were also calculated. Statistical analysis was performed using two-tail z-tests.

A total of 3778 biopsies in 3544 patients were reviewed. The difference in overall detection rate of cancer after biopsy for mammographic calcification between SFM (PPV3=24.7%) and DM (PPV3=23.8%) was not statistically significant (P=.53). On further analysis, however, the PPV for precancerous lesions increased (P<.0001) in DM (11.6%) versus SFM (7.8%). No significant difference (P=.065) was found in detection rate of benign lesions.

When comparing DM to SFM, we found no significant change in PPV3 with respect to calcifications. However, with DM, there was a statistically significant increase in detection of lesions considered at risk for future malignancy. Given that the natural history of these premalignant lesions is incompletely understood, the significance of this finding is in question. This potential trend could be further strengthened by determining PPV1 for cancerous and precancerous lesions with respect to calcifications. Future direction may include a study of regional pathology practices over the eight-year study period to exclude interval change as a confounding factor.

With DM, more of the patients recommended for biopsy based on mammographic calcifications may be getting diagnosed with premalignant lesions, with no significant change in detection of cancer.

Roy, E, Motiuk, D, Burrowes, P, Docktor, B, Digital vs Screen Film Mammography: Impact on Positive Predictive Values Following Transition.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12021565.html