External validation of the 2007 and 2001 Partin tables in Irish prostate cancer patients.

INTRODUCTION: External validation of predictive prostate cancer nomograms is of critical importance within distinct geographical locations, prior to their institution into routine clinical practice. We performed external validation of the 2007 and 2001 Partin tables in a cohort of Irish prostate cancer patients. PATIENTS AND METHODS: Men enrolled in the Irish Prostate Cancer Research Consortium (n = 175) and who had undergone radical prostatectomy between 2004 and 2008 were used to externally validate the 2007 and 2001 Partin tables. A comparative analysis of the clinical and pathological parameters of the Irish and Partin patient cohorts was performed. The reported receiver operating characteristic (ROC) curve derived area under the curve (AUC) values were used to assess for variations in predictive accuracy. Statistical analyses were calculated with R software. RESULTS: AUC values assigned to the differentiation of extra-prostatic extension and seminal vesicle invasion using the 2007 tables are 22 and 3%, respectively. The 2007 Partin tables showed superior accuracy for all parameters, excluding seminal vesicle invasion. CONCLUSION: Cumulatively the Partin tables showed poor discriminate ability for prediction of post-radical prostatectomy pathological outcomes in Irish men, necessitating caution in their clinical utilisation.