Despite emerging evidence and insights, ductal carcinoma in situ (DCIS) continues to perplex even the most experienced of breast radiologists.
It can be a toss-up as to whether to treat DCIS upon diagnosis in women, with anywhere between 20% and 50% of cases progressing to invasive breast cancer. Research is also exploring surveillance imaging options for women who may opt out of surgery or in low-risk cases.
Lars Grimm, MD, joins today's show to discuss these challenges in DCIS care. He is an associate professor of radiology at the Duke University School of Medicine and is a member of the Duke Cancer Institute. He also serves as vice president of the Society of Breast Imaging.
Lars Grimm, MD, discusses challenges in overtreating or undertreating women with DCIS, including what current health guidelines say about managing this tricky early-form breast cancer.
Among Grimm's research projects include the Comparing Two Treatment Approaches for Women with Low-Risk Ductal Carcinoma In Situ (COMET) study, for which he is the lead radiologist. This study aims to further investigate DCIS care strategies for women.
Grimm shares updates from the study, emerging risk stratification tools, and considerations for women seeking surgery or alternative care approaches in this area.
Watch the full episode below.
![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)
