Comparison of Clinical and Automated Breast Density Measurements: Implications for Risk Prediction and Supplemental Screening
This is an excellent new paper from Mayo Clinic.
- Density is a significant risk factor for development of breast cancer, and this is again demonstrated using three different measurement techniques (Volpara, Quantra, clinical BI-RADS).
- Women with extremely dense breasts had approximately twice the risk of developing cancer compared to those with scattered dense breasts.
- VolparaDensity demonstrated an agreement with the average radiologist in judging BI-RADS 4th Edition density category of 0.57 (Quantra 0.46). In a 2000 study published in AJR, Berg found the interobserver agreement of radiologists’ visual assessment of breast density to be 0.43 (weighted kappa).
- Quantra read 37% of the women as “dense”, clinical BI-RADS 43% and Volpara 51%. Volpara used BI-RADS 4th Edition version so with 5th Edition reads, all those % will rise. The only paper to date looking at Volpara and 5th Edition (van der Waal, PLOS ONE, Sept 2015) demonstrated a kappa of 0.8.
So, this is another strong validation study demonstrating density is a risk factor, and it’s pleasing to see the high kappa comparing Volpara to the average radiologist. It will be very interesting to see more comparisons between Volpara and visual assessment of 5th Edition BI-RADS density in due course.