Symptomatic local recurrence of prostate carcinoma after radiation therapy

BACKGROUND: Symptomatic local recurrence of prostate carcinoma (SLRPC) after radiation therapy (RT) is associated with morbidity and debilitating symptoms that have a substantial impact on the patient's quality of life. Most reports on the results of RT for localized prostate carcinoma (PC) do not address this endpoint. The objective of this study was to determine the incidence of SLRPC and to identify the risk factors for this endpoint. METHODS: The medical charts of 1006 patients who received RT for localized PC at the University of Texas M. D. Anderson Cancer Center between 1987 and 1997 were reviewed. Local symptoms were defined as hematuria, voiding symptoms, urinary obstruction, and pelvic pain. Progressive symptoms accompanied by either confirmatory histology or cystoscopic findings were attributed to PC. Univariate and multivariate analyses using Cox proportional hazards models were applied to identify risk predictors. RESULTS: Among 964 patients for whom follow-up data were available, 277 patients had prostate-specific antigen (PSA) progression, and 45 patients died of PC during a median follow-up of 9.4 years. In total, 33 patients (3.4%) developed SLRPC. In patients who experienced biochemical progression, the actuarial 5-year incidence of SLRPC was 8.3%. Among the patients who had developed SLRPC, 23 patients (69.7%) died of PC at a median of 25.3 months from the onset of local symptoms. Adverse histologic tumor subtypes (ductal, small cell, and sarcomatoid) were associated significantly with SLRPC (hazard ratio, 8.4; 95% confidence interval, 2.99-23.63). Clinical T classification at diagnosis, Gleason score, and initial PSA level showed a trend toward an increased hazard ratio. CONCLUSIONS: SLRPC after radiotherapy therapy was an uncommon but clinically significant event. Aggressive histologic subtypes were predictive of this endpoint. Clinical T classification, Gleason score, and initial prostate-specific antigen levels also may have predictive value.