Prognostic Implications of Partial Sampling of Radical Prostatectomy Specimens: Comparison of 3 Methods

We analyzed the prognostic implications of positive margins and extraprostatic extension missed by different methods of partially sampling prostatectomy specimens.The study group consisted of 1,499 patients treated with radical prostatectomy. All specimens were processed uniformly and submitted entirely. For each patient with a positive margin or extraprostatic extension we determined whether these pathological characteristics would have been diagnosed had the specimen been examined by 3 partial sampling techniques. The Harrell concordance index was used to quantify the predictive performance of the Cox models based on the potential findings of the different sampling methods.Partial sampling methods 1 and 2, which included the examination of alternate slides, missed 13% to 21% of positive margins and 27% to 46% of extraprostatic extensions. The effect on biochemical recurrence-free survival of these undetected pathological features was similar to that of positive margins and extraprostatic extension that would have been diagnosed by corresponding techniques. Method 3, which sampled the entire posterior region, the mid anterior prostate and the rest of the ipsilateral anterior gland (if sizeable tumor was seen), detected 95% of positive margins and 94% of extraprostatic extensions. The extraprostatic extension missed by this method was not associated with a significant increase in the risk of biochemical recurrence. The Harrell concordance index of the multivariate models was 0.806, 0.797, 0.795 and 0.804 based on the results of complete sampling, and methods 1, 2 and 3, respectively.Examining alternate sections of prostatectomy specimen results in missing clinically important positive margins and extraprostatic extension. It decreases our ability to predict biochemical recurrence-free survival.