AI-Based Extraction of Retrospective Pathology Data for Breast Cancer Research

In this MEDtalk, Agnar Martin Bjørnstad, PhD student in Medical Technology & E-health at Akershus University Hospital, presents an AI-based approach for extracting and structuring retrospective pathology data from clinical records. Combining OCR, rule-based methods and large language models, the project aims to support the development of multimodal models for breast cancer recurrence risk assessment.