To compare cancer detection, recalls, and positive predictive value of biopsies (PPV3) prompted by contrast-enhanced mammography (CEM) alone, tomosynthesis (DBT) alone, or CEM+DBT in women with personal history of breast cancer (PHBC).

From 10/23/19 through 3/31/22, 1103 eligible women with PHBC consented to a prospective IRB-approved, HIPAA-compliant clinical trial of CEM and DBT. 81 (7.3%) women did not complete any CEM: 70 (6.3%) due to failed iv access and 11 (1.0%) reduced kidney function (eGFR <45 mL/min). 1022 women completed first round screening, of whom 14 (1.4%) had contrast reaction; 383 women completed a second annual CEM (4, 1.0% reaction) to date; and 44, a third (1, 2.3% hives). Two radiologists interpreted each study: R1 reviewed DBT first and R2, CEM. We assessed incremental cancer detection rate (ICDR), cancer type/nodal status, recall rates and PPV3 of biopsies performed, and interval cancer rates. Note these results represent interim analyses of the ongoing study.

The 1022 women ranged in age from 32-82 years (mean 62.9; median 64.1). 32 women were diagnosed with cancer: 27 in year 1 (12 only on CEM, ICDR 12/1000), 3 in year 2 (3 only on CEM, 7.8/1000), and 2 in year 3 (1 only on CEM, 23/1000). There were two interval cancers (2.0/1000): a previously biopsied (enhancing) metastatic axillary node dismissed as benign by both radiologists, and recurrent ILC at scar (enhancing on CEM) not recognized. 25/32 (78%) cancers were seen by R1 on CEM, 13 also on DBT. Another 4 (13%) cancers were identified by R2 on CEM; 1 cancer (3%, DCIS) was recalled only on DBT (overall DBT sensitivity 14/32, 44%). In year 1, adding CEM increased recall from 90/1022 (8.8%) to 141 (13.8%, p<0.0001) for R1 and from 19/383 (5.0%) to 34/383 (8.9%) in year 2 (p<0.0001). At lesion level, PPV3 (biopsies performed) was: 18/60 (30%) prompted only by CEM; 14/37 (38%) by both DBT and CEM; and 5/18 (28%) only by DBT (overall 32/97, 33% by CEM, and 19/55, 35%, by DBT, p=0.86). Of 16 women with cancers seen only on CEM, 14 (88%) were invasive, median size 0.7 cm (range 0.1-2.1); 7 with nodal staging were N0.

In women with PHBC, adding CEM significantly improved detection of early, mostly invasive breast cancer compared to DBT alone. False positiveprompted by CEM are substantial, at 5.0% of women in year 1, but decrease with incidence screens.

Annual screening with CEM appears to have excellent performance characteristics in women with PHBC with little additional yield from DBT. Interval cancer rate to date at 2.0/1000 is lower than literature rates of 3.2 to 3.6 per 1000 with mammography alone.