Tomosynthesis on Par with Mammography in Detecting Cancer

Tomosynthesis does not appear to have an effect on the number of radiologist-recommended breast biopsies, despite the number of screening evaluations completed by the radiologist, according to a presentation at the 2016 annual meeting of the American College of Radiology.

Researchers from the Cleveland Clinic Akron General in Ohio performed a retrospective chart review to determine the positive predictive value (PPV) for conventional mammography (CM) versus tomosynthesis (TS) in patients who underwent a stereotactic breast biopsy, in order to determine if reading TS had changed the radiologist's threshold for recommending biopsy.

The study included chart reviews of 29 patients who underwent TS (mean age 58) and 348 patients who underwent CM (mean age 52). Eleven radiologists read the CM and five of the 11 read the TS. The parameters included the date of the most recent screening evaluation prior to biopsy, radiologist evaluating the exam, recommendation towards biopsy, and BI-RADS score. Following this, the biopsy results and cancer detection rates per 100 patients (CDR) were dichotomized into no disease/benign or malignant. The PPVs of TS versus CM and individual radiologists were compared.

The results showed that 79 patients in the CM cohort and six in the TS cohort were diagnosed with cancer. The CDR for CM was 22.7% and 20.7% for TS. The PPVs were similar and the confidence intervals (CI) overlapped between TS 22.2% and year one CM 22.2%, year two CM 24.3%, and overall CM 23.3%. The PPVs were also similar between screening evaluations completed in year 1 (CM) and year 2 (TS + CM), 22.2% and 24%, respectively. The PPVs were similar and the CI's overlapped among the three radiologists who evaluated the largest number of screening exams.

The researchers concluded that TS may not be more effective than CM in detecting breast cancer, given the similar CDRs and PPVs. They do, however, say that since TS is still a relatively new screening modality, the outcomes warrant the need for further studies with multicenter analysis and larger cohorts.