AUPA: weakly supervised approach for streamlining breast cancer diagnostic workflow by WSI histological type classification for efficient IHC triage

In routine breast cancer (BC) diagnostics, pathologists often review each case twice—first to determine the need for immunohistochemical (IHC) stains, and a second time to emit the final diagnosis—creating significant workload and delays. We present AUPA, an Artificial Intelligence-based system designed to streamline this process by identifying the most invasive histological type in Whole Slide Images of H&E-stained breast biopsies, enabling automatic IHC stain requests. AUPA leverages a weakly supervised method, trained directly on final diagnostic labels without the need for manual annotations. It achieves over 91% sensitivity and specificity across histological types in internal validation and shows strong generalizability in two external pilot studies. In a real-world setting, the system could determine if IHC stains should be ordered (sensitivity and specificity > 96%) and which stains (sensitivity and specificity > 81%). Compared to expert pathologists and state-of-the-art models, AUPA performs competitively, especially distinguishing in situ from benign cases. Designed for real-world deployment, AUPA is fully integrated into the Digital Imaging and Communications in Medicine (DICOM) standard, ensuring interoperability with hospital IT systems. In validation data, it could have saved up to 43 hours of pathologist time. AUPA represents a scalable solution for improving diagnostic efficiency in BC workflows.