New mammography research reveals that the combination of digital breast tomosynthesis (DBT) with synthetic mammography (SM) offers a higher true-positive recall rate, a lower false-positive recall rate and a higher recall positive predictive value (PPV) than digital mammography alone.
For the exploratory analysis from the randomized TOmosynthesis plus SYnthesized MAmmography (TOSYMA) trial, recently published in Radiology, researchers reviewed data from 49,762 women who had digital mammography (DM) screening and 49,715 women who had DBT and SM. There was a 24-month follow-up period for women with false-positive findings, according to the study.
While the study authors noted comparable overall recall rates between the DM and DBT plus SM cohorts, they found a higher true-positive recall rate for DBT plus SM (8.4 per 1,000 women vs. 6.2 per 1,000 women) as well as a lower false-positive recall rate (40.2 per 1,000 women vs. 43.7 per 1,000 women).
The researchers pointed out that differences with the false-positive recall rate were particularly pronounced with the first round of mammography screening (94 per 1,000 women for DBT plus SM vs. 109.6 per 1,000 women for DM).
The DBT plus SM cohort had a higher false-positive biopsy rate in comparison to those who had DM only (7.8 per 1,000 women vs. 1.8 per 1,000 women). However, the study authors also noted a higher biopsy PPV with DBT plus SM (51.7 percent vs. 50.6 percent).
“ … The recall and false-positive recall rates were particularly lower with DBT plus SM in first-round screening participants. The positive predictive value of recall was consistently higher with DBT plus SM. Mildly raised false-positive biopsies with DBT plus SM were compensated with the detection of more breast cancers,” wrote lead study author Stefanie Weigel, M.D., who is affiliated with the Clinic for Radiology and Reference Center for Mammography Munster at the University of Munster and University Hospital Munster in Munster, Germany, and colleagues.
Three Key Takeaways
- Improved detection with DBT plus SM. The combination of digital breast tomosynthesis (DBT) and synthetic mammography (SM) yielded a higher true-positive recall rate and a consistently higher recall positive predictive value (PPV) compared to digital mammography (DM) alone.
- Lower false positives, especially in first-round screening. DBT plus SM significantly reduced false-positive recalls compared to DM, with the most pronounced reduction seen in women undergoing their first round of screening.
- Slight increase in false-positive biopsies but better PPV. While DBT plus SM was associated with a higher false-positive biopsy rate, this was offset by a higher biopsy PPV and greater overall breast cancer detection.
In an accompanying editorial, Benoit Mesurolle, M.D., and Mona El Khoury, M.D., maintained that the lower recall and false-positive recall rates from the first mammography screening round with DBT plus SM were unusual to see given the similar overall recall rates between the two cohorts.
“This finding is novel compared with the widely accepted understanding that prevalent screening examinations tend to have higher recall rates than incident screening examinations, likely due to the challenge imposed by the unavailability of prior examinations, which are known to facilitate comparison. European and Canadian guidelines set the acceptable rates at under 7% and under 10% for prevalent screening examinations, respectively, compared with under 5% and 7% for incident screening examinations,” noted Dr. Mesurolle, a radiologist at Republic Health Center in Clermont-Ferrand, France, and Dr. El Khoury, an associate professor of radiology at the University of Montreal in Canada.
(Editor’s note: For related content, see “Large Mammography Study Shows Mixed Results with AI in Breast Cancer Screening,” “Reducing Mammography Workload by Nearly 40 Percent? What a New Hybrid AI Study Reveals” and “Digital Breast Tomosynthesis Study Assesses Impact of Architectural Distortion on Malignancy Rates.”)
In regard to study limitations, the authors acknowledged the lack of assessment for inter-reader agreement and conceded that comparing DBT screening with prior DBT exams and not DM examinations may have affected recall indications. The researchers also cautioned that the subgroup analyses were exploratory and lacked sufficient power for definitive interpretation.
