The use of clinical data to predict response to therapy

Clinical data are useful for predicting response to therapy and prognosis of patients with malignant disease. Multivariate statistical methods have proven to be useful for objectively evaluating those variables which are most useful for predicting outcome. Once variables have been identified, models can be created which predict for individuals the expectations for response and survival and such models with known clinical variables are useful for evaluating the contribution of newly developed information such as in vitro testing or a new laboratory test in terms of its contribution to the ability to predict response. The models are then useful for evaluation of treatment effects by comparing observed effects of new treatments to expectations for each patient.