Research Round Up
AI vs. ultrasound
A single-institution, retrospective study of 5707 women with dense breasts and a negative mammogram at Seoul National University Hospital found that adding artificial intelligence (AI) is less effective than adding screening ultrasound (US). Supplemental physician-performed whole breast handheld US found more cancers than 2D digital mammography (MG) plus AI with an added incremental cancer detection rate of 2.2 (95% CI 1.1, 3.8) per 1000 screens. All 12 additional cancers seen by US alone were invasive and node negative. While sensitivity was higher for MG plus US than for MG plus AI (97% vs. 60%, p=.002), specificity was lower (77.6% vs. 95.3%, p<.001). Further research is needed to better understand the role of AI in women with dense breasts.
Contrast-enhanced mammography
Patel and colleagues at Mayo Clinic, Scottsdale, AZ, reported results of the first round of screening contrast-enhanced mammography (CEM) after negative digital breast tomosynthesis (DBT) in 460 women at elevated risk, of whom 408 (89%) had dense breasts. CEM prompted biopsy in 37 (8.0%) women, with 11/37 (30%) found to have cancer, for a supplemental cancer detection rate of 24/1000. Specificity of CEM was 393/449 (87.5%). Of 14 total malignancies found, 10 were invasive, median size 9 mm, all node negative; there was one interval cancer found on screening MRI.