Predictive modeling in prostate cancer. Conclusions and reflections.

Abstract

Validated prognostic factors are an integral part of any cancer diagnosis. The available proliferation of new markers and randomized clinical trial results leads to a complex decision making looking for optimal treatment and outcome results. Advanced mathematic models have been developed to weigh each specific prognostic factor into a single outcome number in nomograms modeled for specific predictive accuracy in key phases of the treated history of the disease. Despite this progress, it is important to realize that the results are an outcome stratification for groups rather than for the individual patient. It is clear that in the tumor prognosis, host and environmental factors need to be evaluated before any clinical decision is made. Cancer 2009;115(13 suppl):3160-2. (c) 2009 American Cancer Society.