Christiane Mary Hakim MD, Presenter: Nothing to Disclose

David Gur PhD, Abstract Co-Author: Nothing to Disclose

Jules Henry Sumkin DO, Abstract Co-Author: Scientific Advisory Board, Hologic, Inc

Marie Adele Ganott MD, Abstract Co-Author: Nothing to Disclose

Denise Chough MD, Abstract Co-Author: Nothing to Disclose

Victor J. Catullo MD, Abstract Co-Author: Nothing to Disclose

Amy Elizabeth Kelly MD, Abstract Co-Author: Nothing to Disclose

Dilip Digambar Shinde MD, Abstract Co-Author: Nothing to Disclose

Luisa Paula Wallace MD, Abstract Co-Author: Nothing to Disclose

To assess the impact of the availability of PRIOR examinations and digital breast tomosynthesis (DBT), individually and jointly, on decisions to recall a woman during mammographic interpretations.

Under an IRB approved HIPAA compliant protocol eight radiologists independently interpreted 153 FFDM, PRIOR and DBT examinations in a paired design, fully crossed, mode balanced study. During one mode three sequential ratings were provided using FFDM only, followed by FFDM+ DBT, followed by FFDM+DBT+PRIORS. During the second mode FFDM only, then FFDM+PRIORS, then FFDM+PRIORS+DBT were viewed and rated. The enriched dataset of 153 verified (two year follow up) examinations included 43 benign cases clinically recalled, 60 negative cases, and 50 verified cancer cases. Recall recommendations under each mode and interaction between the effect of PRIORS and DBT on decisions were assessed at a 0.05 significance level using generalized linear model (proc glimmix, SAS. V.9.3) accounting for case and reader variability.

Regardless of the order of presentation, there were no statistically significant differences between the multiplicative effects of the availability of DBT and PRIORS on assessment of cancer-free cases (p=0.71, interaction term in logit-link model). DBT, whether presented before or after PRIORS, reduced recall rate recommendations of negative/benign cases. Average recall rates in all non-cancer cases were significantly reduced (25%; from 0.51 to 0.38, p=0.03) with the addition of DBT first and with addition of PRIORS first (35%, from 0.51 to 0.33 p<0.01). For a combination of FFDM with PRIORS, addition of DBT reduced recall rate by 30% (from 0.33 to 0.23, p=0.02) and addition of PRIORS to FFDM+DBT reduced the recall rate by 34% (from 0.38 to 0.25, p<0.01). Availability of DBT did not have significant effects on sensitivity whether added to FFDM only (p=0.38) or to FFDM+PRIORS (p=0.13). Notably, availability of PRIORS reduced sensitivity when added to FFDM only (p<0.01) or to FFDM+DBT (p=0.09).

PRIORS and DBT are independent, primary contributing factors in reducing recall recommendations during mammographic interpretations.

Comparison with prior imaging and review of DBT are primary, independent contributing factors in decision making. Each lead independently to a significant decrease in unnecessary recalls.

Hakim, C, Gur, D, Sumkin, J, Ganott, M, Chough, D, Catullo, V, Kelly, A, Shinde, D, Wallace, L, Impact of Availability of PRIOR Examinations and DBT Individually and Jointly on the Interpretation of Non-recalled, Clinically Recalled and Cancer-verified Mammograms.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14002675.html