A novel approach for cancer outcome prediction using personalized classifier

Cancer patients can be classifier into different subtypes but an obvious question can ask, whether these predefined subtypes can be helpful to detect the risk of disease in advance which is a well known problem in cancer biology. In this article we address these issues of using subtypes for disease outcome prediction and propose a personalized classification approach by relaxing the predefined subtype idea. Finally, we find that our proposed method is helpful to develop prediction model which can be useful in clinical practice to predict disease outcome. Thus the idea proposed here has better accuracy and application than conventional approaches for cancer outcome prediction.