Systematic assessment of the ability of the number and percentage of positive biopsy cores to predict pathologic stage and biochemical recurrence after radical prostatectomy.
OBJECTIVES: We hypothesized that the number and/or percentage of positive cores, proxies of tumor volume, could improve the ability to predict pathologic stages and/or biochemical recurrence (BCR). To test this hypothesis, we examined radical retropubic prostatectomy (RRP) data from three centers on two continents. MATERIAL AND METHODS: Clinical data from men undergoing RRP at three different institutions were used to predict pathologic stages and BCR. Univariable and multivariable logistic analyses and Cox regression analyses were used. Predictive accuracy (PA) was assessed with the area under the receiver operating characteristics curve estimates, which were subjected to 200 bootstraps to reduce overfit bias. The statistical significance of PA gains was assessed with the Mantel-Haenszel test. RESULTS: The number and the percentage of positive cores were independent predictors of virtually all pathologic stage outcomes and of BCR. In PA analyses, the percentage of positive cores improved the PA of pathologic stage predictions and of BCR predictions between 0.06% and 1.49%. Conversely, the number of positive cores improved the PA of pathologic stage predictions and of BCR predictions between 0.36% and 1.14%. CONCLUSIONS: The information derived from biopsy cores is important and can improve the ability to predict pathologic stage and BCR. It appears that the percentage of cores is most helpful in stage predictions. Conversely, the number of cores appears to improve mostly BCR predictions. Consideration of both variables might not be helpful because of the similarity of information they encode.
OBJECTIVES: We hypothesized that the number and/or percentage of positive cores, proxies of tumor volume, could improve the ability to predict pathologic stages and/or biochemical recurrence (BCR). To test this hypothesis, we examined radical retropubic prostatectomy (RRP) data from three centers on two continents. MATERIAL AND METHODS: Clinical data from men undergoing RRP at three different institutions were used to predict pathologic stages and BCR. Univariable and multivariable logistic analyses and Cox regression analyses were used. Predictive accuracy (PA) was assessed with the area under the receiver operating characteristics curve estimates, which were subjected to 200 bootstraps to reduce overfit bias. The statistical significance of PA gains was assessed with the Mantel-Haenszel test. RESULTS: The number and the percentage of positive cores were independent predictors of virtually all pathologic stage outcomes and of BCR. In PA analyses, the percentage of positive cores improved the PA of pathologic stage predictions and of BCR predictions between 0.06% and 1.49%. Conversely, the number of positive cores improved the PA of pathologic stage predictions and of BCR predictions between 0.36% and 1.14%. CONCLUSIONS: The information derived from biopsy cores is important and can improve the ability to predict pathologic stage and BCR. It appears that the percentage of cores is most helpful in stage predictions. Conversely, the number of cores appears to improve mostly BCR predictions. Consideration of both variables might not be helpful because of the similarity of information they encode.