Converting a promising in-house AI research tool into a medical device requires careful planning, says Dr. Filippo Pesapane, PhD, a radiologist at the European Institute of Oncology (IEO) in Milan. He spoke on this topic at ECR 2024, and in a video interview, he elaborated on the key points to consider during this process.
Pesapane, who is a member of AuntMinnieEurope.com's editorial advisory board, also provided an update on his other research studies, including those looking at the practical applications of AI in breast imaging. Additionally, he reflected on his relocation from King's College Hospital and Royal Marsden Hospital in London to Milan, and looked ahead to the 51st Italian Society of Medical Radiology (SIRM) National Congress, which begins in Milan on 20 June 2024. Outside of radiology, he recently became a new parent, and he spoke about how this has changed his life.









![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.auntminnieeurope.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)









