Automated Clear Cell Carcinoma Grade Classification with Prognostic Significance
Katherine Tian
Cupertino, CA
Age 17
Clear cell renal carcinoma (ccRCC) is the most common type of kidney cancer. One of the most important clinical factors in predicting ccRCC patient prognosis is the nuclear grade, an indication of how fast the tumor is likely to grow and spread. Correctly classifying ccRCC grade is paramount in guiding clinical management and molecular-based therapies. In this project, we developed an automated 2-tier ccRCC grade classification system using whole slide images (WSI) from The Cancer Genome Atlas (TCGA) ccRCC dataset. Our grade classification system has outperformed human pathologists, and can provide valuable second opinion to doctors to help prevent misdiagnosis.