Predicting Time to First Treatment in Chronic Lymphocytic Leukemia Using Machine Learning Survival and Classification Methods

摘要

Dealing with censored data is an important consideration for disease prognosis modeling. This is particularly true when diseases have highly heterogeneous presentations and prognosis. Algorithms used to develop prognostic models must be robust to censored data. We explore methods to deal with censoring in a highly heterogeneous disease - chronic lymphocytic leukemia. Although survival analysis is the standard method for estimating survival times, binary classifiers can potentially yield better predictive accuracy, depending on the outcome specified.