The purpose of this research is to determine whether or not a CT scan can accurately predict a patient's chance of survival after a diagnosis of kidney cancer. This study utilizes two deep learning networks, one for extracting features from images and another for predicting how long an individual will live. This study's survival model is non-linear and non-proportional. Our methodology produces survival curves for each individual that are unique to them.