Joanna Stern undergoes AI-assisted breast cancer screening
Updated
Updated · The Wall Street Journal · May 4
Joanna Stern undergoes AI-assisted breast cancer screening
1 articles · Updated · The Wall Street Journal · May 4
At Mount Sinai in Manhattan, radiologist Laurie Margolies reviewed 95 mammogram and ultrasound images with ScreenPoint and Koios tools, finding low cancer risk but recommending a six-month ultrasound follow-up.
Stern said dense breasts and a family history, including a mother who survived breast cancer three times, drove the test; one stable right-breast mass was flagged suspicious by AI but later proved benign.
The report portrays AI as augmenting rather than replacing radiologists, citing claims and studies that it can improve detection in dense breasts while still missing some cancers and generating false positives.
AI finds more cancers, but can we trust a tool that is more often wrong for Black and older women?
If AI helps radiologists, why does it increase their workload and legal risks when they disagree with its findings?
From Trial to Clinic: Real-World Validation and Challenges of AI in Breast Cancer Screening
Overview
In early 2026, Joanna Stern underwent AI-assisted breast cancer screening, sparking widespread public discussion about AI in healthcare. The AI technology, trained on vast mammogram data, improved cancer detection rates, reduced false positives, and eased radiologist workloads by up to 63.6%. Clinical trials confirmed AI's accuracy matched traditional double reading, detecting more cancers and fewer interval cancers. Despite these benefits, barriers like clinician trust, high costs, and data bias challenge adoption. Looking ahead, AI-powered risk-stratified screening promises personalized care and better resource use, reshaping radiologists' roles and addressing workforce shortages. However, ensuring equity, privacy, and clear communication remains essential for responsible implementation.