Predicting clinical end points: treatment nomograms in prostate cancer

Abstract

Due to the generally indolent nature of prostate cancer, patients must decide among a wide range of treatments, which will significantly affect both quality of life and survival. Thus, there is a need for instruments to aid patients and their physicians in decision analysis. Nomograms are instruments that predict outcomes for the individual patient. Using algorithms that incorporate multiple variables, nomograms calculate the predicted probability that a patient will reach a clinical end point of interest. Nomograms tend to outperform both expert clinicians and predictive instruments based on risk grouping. We outline principles for nomogram construction, including considerations for choice of clinical end points and appropriate predictive variables, and methods for model validation. Currently, nomograms are available to predict progression-free probability after several primary treatments for localized prostate cancer. There is need for additional models that predict other clinical end points, especially survival adjusted for quality of life.