AN ALGORITHM FOR PREDICTING NONORGAN CONFINED PROSTATE-CANCER USING THE RESULTS OBTAINED FROM SEXTANT CORE BIOPSIES WITH PROSTATE-SPECIFIC ANTIGEN LEVEL

Purpose: We determined the enhanced ability to predict nonorgan confin ed prostate cancer using several histopathological and quantitative nu clear imaging parameters combined with serum prostate specific antigen (PSA). Materials and Methods: Several independent pathological and qu antitative image analysis variables obtained from sextant biopsy speci mens, as well as preoperative PSA were used. The study population incl uded 210 patients with pathologically staged disease (192 with PSA). A ll variables were examined by univariate and multivariate logistic reg ression analyses to assess ability to predict disease organ confinemen t status. Results: Univariate logistic regression analysis demonstrate d that, in decreasing order, quantitative nuclear grade, preoperative PSA, total percent tumor involvement, number of positive sextant cores , preoperative Gleason score and involvement of more than 5% of a base and/or apex biopsy were significant (p less than or equal to 0.006) f or prediction of disease organ confinement status. Backward stepwise l ogistic regression was applied to these univariately significant varia bles, including deoxyribonucleic acid ploidy, to calculate a multivari ate model for prediction of disease organ confinement status. This alg orithm had a sensitivity of 85.7%, specificity 71.3%, positive predict ive value 72.9%, negative predictive value 84.7% and area under the re ceiver operating characteristic curve 85.9%. Conclusions: Information from pathological study of sextant prostate biopsies, preoperative PSA blood test and a new image analysis variable termed quantitative nucl ear grade can be combined to create a multivariate algorithm that can predict more accurately nonorgan confined prostate cancer compared to previously reported methods.