Memorial Sloan Kettering Cancer Center

Computational pathology model predicting ten lymphoma subtypes from H&E whole-slide images and ordering the matching immunohistochemistry panel.

Self-supervised brain MRI foundation model built on DINOv3, pretrained on roughly 6.6 million unlabeled axial slices for neuroimaging tasks.

Distilled few-step protein backbone generator that adapts Score Identity Distillation to Proteina for over 20x faster de novo structure sampling.

Multimodal slide-level pathology foundation model trained by clinical-dialogue supervision on 2.3M whole-slide images and 14M Q&A pairs.

3D segmentation foundation model for female genito-pelvic anatomy, reading T2-weighted MRI and radiotherapy planning CT with one shared encoder.

Breast cancer recurrence risk read from an H&E slide and six routine clinical variables, using a frozen pan-cancer pathology encoder.