Researchers from Microsoft Research and Paige, now part of Tempus, developed PRISM2, a pathology foundation model trained on both tissue images and language from real pathology reports, as described in a study published in Nature Medicine. The model was designed to help AI learn from available text data and support future pathology research, moving beyond single-task systems that require rebuilding for each new application. In testing, PRISM2 matched or exceeded the performance of specialized cancer-detection systems on benchmark tasks including prostate cancer, breast cancer, and breast lymph node metastasis detection, without creating a separate model for each task. The full PRISM2 model weights are publicly available for research use on Hugging Face. The approach is based on the idea that pathology is language-driven as well as visual, pairing images with report-derived information to create millions of question-and-answer examples. The resulting model can work with images only or images and text, allowing users to interact through prompts rather than task-specific software. This differs from earlier pathology foundation models that focused primarily on visual representations. Researchers say this could make it easier to build and adapt future pathology tools without starting from scratch for each new application.