Tuesday, December 1 | 11:20 a.m.-11:30 a.m. | RC315-14 | Arie Crown Theater
In this presentation, researchers will describe how adding digital breast tomosynthesis (DBT) to 2D mammography appears to steadily improve the positive predictive value of biopsies.A team led by Dr. Madhavi Raghu of Yale University compared the positive predictive value of biopsies performed (PPV3) following diagnostic workup before and after DBT was incorporated into Yale's imaging department. The study included four time frames: one year of 2D mammography (June 2010 to June 2011) and three consecutive years of 2D-DBT mammography (January-December 2012, 2013, and 2014).
A total of 3,567 2D diagnostic mammograms were conducted between 2010 and 2011. In 2012, 3,385 2D-DBT mammograms were performed; in 2013, 4,542 were performed; and in 2014, 4,507 were performed. Raghu's group found that DBT had a clear effect on PPV3 over the study time frame.
Diagnostic workup with tomosynthesis demonstrates a significant and steady increase in the positive predictive value of biopsies performed, the researchers concluded.
![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=100&q=70&w=100)
![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&dpr=2&fit=crop&h=167&q=70&w=250)
