Nomogram predicting the probability of early recurrence after radical prostatectomy for prostate cancer.

PURPOSE

We developed a nomogram predicting the probability of early biochemical recurrence after radical prostatectomy because early recurrence predisposes to distant metastasis and prostate cancer related mortality. Identifying patients at risk for early recurrence may improve prognosis as early institution of adjuvant therapy may reduce the risk of progression.

MATERIALS AND METHODS

From January 1992 to December 2005, 2,911 patients underwent radical prostatectomy for localized prostate cancer. Cox regression models addressing biochemical recurrence after radical prostatectomy were used to identify significant predictors. Age, prostate specific antigen, pathological Gleason sum, surgical margin, extracapsular extension, seminal vesicle invasion and lymph node invasion were considered. A nomogram predicting the probability of biochemical recurrence-free survival within 2 years after radical prostatectomy was developed. Data from an independent center were used for external validation (2,875).

RESULTS

In both cohorts combined during the first 2 years 11.0% (639) of all patients experienced relapse which accounted for 58.5% of all observed biochemical recurrence. In the development cohort except for age all covariates represented significant predictors of biochemical recurrence after radical prostatectomy. Pathological Gleason sum 7 or greater, seminal vesicle invasion and lymph node invasion were the most powerful predictors of biochemical recurrence. The accuracy (c-index) of the nomogram predicting biochemical recurrence-free survival within 2 years after radical prostatectomy was 0.82 in the external validation cohort.

CONCLUSIONS

Two-thirds of all instances of relapse occur during the first 2 years after radical prostatectomy. Those patients can be highly accurately identified with our nomogram. They might benefit the most from adjuvant treatment and could be the ideal candidates for adjuvant treatment trials.