In prostate cancer management, key clinical variables—such as Gleason score, tumor laterality, margin status, and pathologic staging—are often embedded in unstructured pathology and imaging reports. Manual extraction for research or audit is time-consuming and prone to variability. An AI-powered system is presented to automate data extraction from prostate biopsy and robotic prosta-tectomy reports, enabling high-quality and scalable outcome analysis. The system integrates OCR for digitizing scanned reports, large language models (LLMs) for clinical interpretation, and rule-based logic for validation. Document specific extractors (e.g., for MRI, micro-ultrasound, biopsy, prostatectomy) capture relevant variables, which are then standardized and linked by patient identifiers. All structured data are stored in a relational database with traceability to the original reports
Urology, Prostate Cancer, Artificial Intelligence