Objectives: Following surgery, men with recurrent prostate cancer have an isolated elevation in serum prostate-specific antigen (PSA) well in advance of measurable metastatic disease. Rational patient selection for new forms of adjuvant therapy, for example, gene therapy, is imperative.
Methods: In a retrospective study of two cohorts, we used proportional hazards regression analysis to develop and validate a multifactor model for identifying men who are at high risk of cancer recurrence. The model cohort consisted of 216 men with clinical Stage T2b and T2c treated by 1 urologist. The validation cohort consisted of 214 men with Stage T2b and T2c disease.
Results: A model for log relative risk, Rw, used serum PSA with a sigmoidal transformation (PSAST), radical prostatectomy Gleason score (GS), and pathologic stage (PS) as specimen confined or nonspecimen confined: Rw = (PSAST x 0.06) + (GS x 0.54) + (PS x 1.87). Recurrence risk categories were determined as low risk if Rw is less than 4.0, intermediate risk if it is 4.0 to less than 5.75, and high risk if Rw is more than 5.75. The observed Kaplan-Meier actuarial analysis of the three risk groups correlated well with the predictions determined for the model cohort. We then validated this model independently using a second cohort of 214 men with similar age, stage, and grade treated by 3 different urologists at two different institutions.
Conclusions: The recurrence rates for men in the high-risk group are similar to those for men with positive lymph nodes and justifies exploration of experimental adjuvant therapy within this group using this model of patient selection.