MB-DSMIL-CL-PL: Scalable Weakly Supervised Ovarian Cancer Subtype Classification and Localisation Using Contrastive and Prototype Learning with Frozen Patch Features
New AI method improves ovarian cancer subtype classification by 70.4% without sacrificing training speed.
Researchers led by Marcus Jenkins and Michal Mackiewicz developed MB-DSMIL-CL-PL, a new AI model for ovarian cancer analysis. It uses contrastive and prototype learning with frozen patch features to achieve a 70.4% F1 score improvement for instance-level classification and 16.9% better AUC for tumor localization. The system maintains scalability by avoiding computationally expensive end-to-end training, allowing pathologists to classify cancer subtypes and locate tumors more accurately from histopathology slides.
Why It Matters
This addresses critical pathology lab bottlenecks, enabling faster, more precise cancer diagnosis without requiring massive computational resources.