Stratifying patient outcomes with RCC using AI based on whole genome sequencing and pathological information

1. Obtain whole genome sequencing for 12 patients with different histologies of renal cell carcinoma
2. Use AI to merge DNA data together Dicom Pixelized histologic data to predict clinical outcomes
3. Apply the histologic algorithm thus obtained to a confirmation cohort of 120 patients from within the biobank 

Summary

Renal cell carcinoma is a heterogeneous disease with the soul predictive elements being pathologic data (grade and stage) for localized disease, and clinical data (Performance status and blood work) for advanced disease. Such information, although being the best available, is very limited and crude; it has weak predictive value. Our aim is to build on genetic data obtained from whole genome sequencing and combine it with pathological information using artificial intelligence in a limited but representative set of patients in order to optimize prediction of clinical outcome. The AI machine learning thus developed will then be tested for validity in a larger cohort of patients using their histological information to predict their outcomes. Gaining predictive power will allow tailoring patient treatment to maximize efficacy and avoid undue adverse effect. It will also help provide a more accurate prognosis; ultimately resulting in optimized care.