Impact of the radical prostatectomy surgical technique and surgeon experience on freedom from cancer recurrence

4569 Background: We have shown a direct relation between positive margin (PM), morbidity outcomes and surgeon volume, technique after radical prostatectomy (RP). Significant variation occurs even among high volume providers. Our aim was to analyze the surgeon effects on cancer control after RP. Methods: We evaluated 8196 consecutive cT1–3NxMx naive men who underwent RP by one of 76 surgeons within 4 institutions between 1987 and 2003. We calculated the 5-yr probability of recurrence (BCR, PSA elevation >0.4 ng/ml × 2 or initiation of secondary therapy for a PSA rise) for each surgeon assuming a log-logistic survival distribution. A meta-analysis controlling for case mix: PSA, Gleason score, stage, PM and surgical expertise (i.e. cumulative number of surgeries performed) to evaluate for differences in BCR rates between surgeons was performed. We applied the I-square statistic to determine what proportion of the variation represented genuine differences v. chance alone. Results: 33 surgeons performed > 40 RP with 17 surgeons having > 100 procedures during the study period. BCR events were recorded in 1361 patients. The overall 5-yr freedom from BCR with 2524 patients remaining at risk was 80% (79%, 81%) . Extracapsular extension, seminal vesicle invasion, nodal metastasis, PM, Gleason score and PSA were independent predictors of BCR. The surgical volume also correlated independently with BCR. Importanntly we found significant variability on freedom from BCR between high volume surgeons. The I-squared statistic from the meta-analysis was 0.63. That is, approximately 63% of the difference in BCR rates among surgeons can be explained by genuine differences in surgical skill and approach, and approximately 37% is compatible with chance alone. For a sensitivity analysis, we repeated the analysis excluding surgeons who performed less than 100 procedures. In this sub-analysis, the I-squared statistic remained very significant at 0.48. Conclusions: Our data shows that in men treated by RP, the BCR outcomes of men are not exclusively determined by the biology and stage of prostate cancers (explained in most models), but to the surgical skill. Clinical trials evaluating BCR outcomes must prove equivalency among providers so that results are not biased by them. No significant financial relationships to disclose.