Talk: Gleason grading of biopsies using an attention-based multi-resolution model ensembled with LGBM and XGBoost
Date:
In this talk we present our work in PANDA challenge. We developed an automated prostate Gleason grading algorithm based on an attention-based multi-resolution model ensembled with LGBM and XGBoost. Our model was trained on patch-based tissue samples extracted from whole slide images (WSIs). A two-stage attention-based multiple instance learning (MIL) model using weakly supervised region of interest (ROI) detection was developed for ISUP-grade prediction. The model was trained on multiple resolutions, with the lower resolution to identify suspicious regions that were further examined at higher resolution. To make the model more robust, we ensembled the MIL model with LGBM and XGBoost models, whose feature extractors were trained to predict the primary and secondary Gleason scores.