ABSTRACT
With the use of transrectal ultrasound (TRUS) and prostate-specific antigen (PSA) testing, an increasingly large number of nonpalpable tumors are being discovered in the prostate.1 Some of these tumors are indolent but many are clinically significant, and deciding the most appropriate treatment strategy has been a challenge for both physicians and patients. Whatever the treatment option, whether watchful waiting, radical prostatectomy, brachytherapy, or external-beam radiation therapy (EBRT), the patient needs to be able to make an informed decision based on his predicted probability of survival, as well as his perception of his future quality of life. Prostate cancer is a classic example of a disease in which quality of life is as important as survival time, if not more so.2 A tool that aids in the decision-making process is of great value for the patient who must weigh various treatment strategies against their possible complications and comorbidities.3
Several instruments do exist,4,5 the most common of which are risk-grouping tables and specialized prediction models called nomograms. Risk-grouping tables place patients into categories based on the overall prognosis of the group. Nomograms, on the other hand, focus on the risk for the individual patient, and thus maximize predictive accuracy. Although risk-grouping tables are usually more paper-friendly, their construction is often flawed, resulting in loss of homogeneity in the group.6,7 In this chapter, we will outline the models available for predicting biochemical recurrence after local definitive therapy with radical prostatectomy, brachytherapy, and EBRT. Because many of the models that we will discuss in detail are nomograms, we will provide a description of the
