Supplementary Data S2 from Quantitative and Morphology-Based Deep Convolutional Neural Network Approaches for Osteosarcoma Survival Prediction in the Neoadjuvant and Metastatic Settings

Blinded annotations by two pathologists of HPCs. Each HPC was annotated 4 times by each pathologist based on 100 tiles drawn randomly each time from 3 different sets. The 100 tiles were randomly selected from the NYU training (twice), from the NYU test set, and from the Charles’ University external cohort.