Breast imaging software developer Volpara Solutions is highlighting a study from the Netherlands showing a strong linear relationship between decreased screening performance and volumetric breast density.
The study involved more than 110,000 mammograms from the Dutch breast screening program. Researchers found that higher volumetric breast density, as automatically measured by VolparaDensity software, affects mammography performance measures, such as sensitivity and the rates of recall, false positives, and interval cancers.
The study was published in Breast Cancer Research and Treatment (23 December 2016) and was carried out in the Adapting Breast Cancer Screening Strategy Using Personalized Risk Estimation (ASSURE) project, funded by the European Union. The results suggest that because fully automated measurements of breast density have much higher reproducibility than visual assessment, this automatic method could help with implementing density-based supplemental screening.
Dr. Carla van Gils and a team of researchers examined screening performance in the Dutch screening program using a consecutive series of 111,898 digital mammograms from 53,239 women (ages 50-75). Volumetric breast density was automatically assessed using VolparaDensity software.
The study included 667 screen-detected and 234 interval cancers. Of all the tumors, 84.3% were invasive cancers. Mammographic sensitivity was 85.7%, 77.6%, 69.5%, and 61.0% across density categories 1 to 4, respectively. Higher breast density showed a strong linear relationship with all screening performance measures, except for positive predictive value. The rates for recalls, false positives, and total and interval breast cancers were greater in higher breast density categories than in lower density categories.
The group found that a woman was nearly seven times more likely to have an interval cancer if her breasts were extremely dense versus very fatty, and approximately twice as likely to have a false-positive result if her breasts were extremely dense versus very fatty.










![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=112&q=70&w=112)






