Biostatistical modeling using traditional variables and genetic biomarkers for predicting the risk of prostate carcinoma recurrence after radical prostatectomy

BACKGROUND Approximately 50-60% of patients treated with radical prostatectomy for clinically localized prostate carcinoma are found to have microscopic disease that is not organ-confined, and a significant portion of these patients will relapse. Multiple studies have attempted to identify these high risk patients by evaluating many potential prognostic variables. These studies, however, have not included the more recent molecular biomarkers found to be independent predictors of disease recurrence. METHODS One hundred thirty-two patients who underwent radical prostatectomy at one center between 1986 and 1993 were subjected to a multivariable Cox regression analysis to determine the preoperative and postoperative variables that remain significant predictors for the likelihood of serologic recurrence. The preoperative variables included in the model were age, race, and prostate specific antigen(PSA); the postoperative variables were Gleason sum, nuclear grade, pathologic stage (capsular status), p53 tumor suppressor gene expression, bcl-2 protooncogene expression, and proliferative biomarker Ki-67 expression. Biomarkers were also evaluated separately. RESULTS A model was developed using only variables that remained significant predictors for the likelihood of recurrence. The following equation calculated the relative risk of recurrence: Rw = exp [(0.70 × Race) + (0.79 × PSA{4.1-10}) + (1.34 × PSA{>10}) + ( 0.91 × Organ confinement) + (0.65 × p53{1,2+}) + (1.45 × p53{3,4+}) + (0.70 × bcl-2)]. This equation categorized men into 3 distinct risk groups (low: Rr 15.0). CONCLUSIONS This equation allows patients at high risk for PSA recurrence to be identified shortly after radical surgery. These patients at high risk for serologic recurrence and eventual progression may be considered for currently accepted adjuvant therapy or enrollment in clinical trials for the newer investigational therapies for locally advanced prostate carcinoma. Cancer 1997; 79:952-62. © 1997 American Cancer Society.