Preliminary report: Comparison of performance in combined screening with digital mammography and ultrasound versus conventional screening with digital mammography alone among Korean women aged 40–59 years old

Objective To compare the performance of combined screening with digital mammography and ultrasound, using modified criteria for breast ultrasound screening, versus conventional screening with digital mammography alone in Korea. Methods From May 2016 to March 2018, 5,969 asymptomatic women aged 40–59 years old were enrolled in a prospective cohort study from 12 university-affiliated hospitals participating in the MUST-BE (Mammography and Ultrasonography STudy for Breast cancer screening Effectiveness) trial. Full-field digital mammography and radiologist-performed hand-held breast ultrasound were obtained in all cases. For this study, new criteria for breast ultrasound screening were established via modification of the BI-RADS (Breast Imaging Reporting and Data System) criteria. In the combined screening, the same radiologist interpreted both mammography and ultrasound; in the conventional screening, a second radiologist in another hospital interpreted the mammography on a monthly basis without any ultrasound findings or follow-up results. We identified positive cases, including those probably benign or inconclusive and those suspicious, and compared the recall rate, positive predictive value (PPV), and cancer detection rate (CDR) per 1,000 women screened by one of the two screening methods. The reference standard was obtained from histological results and a 12-month breast ultrasound follow-up period. Results There were 421 probably benign and 287 suspicious cases in the combined screening, and 618 inconclusive and 48 suspicious cases in the conventional screening. We performed biopsy in 231 and 89 women and detected 45 and 31 cancer patients in the combined and conventional screenings, respectively. The recall rate was 11.9% and 11.2%, respectively (p=0.03). The preliminary PPV1 was 6.4% and 4.7%, respectively (p=0.17) and the CDR was 7.5 and 5.2 per 1,000 women screened, respectively (p=0.11). Conclusions Using the MUST-BE criteria, the combined screening with digital mammography and ultrasound might result in similar PPV1 and CDR compared to conventional screening with digital mammography alone but a higher recall rate.