Are Partin tables suitable for Chinese patients with prostate cancer? / 中华医学杂志(英文版)

ABSTRACT

<p><b>BACKGROUND</b>Recently, the number of patients with prostate cancer who needed to be treated with radical prostatectomy increased rapidly in China. There is still a difference between clinical staging and the post-operative final pathologic staging; hence, an excellent tool for accurately predicting the pathologic stages of prostate cancer is needed urgently in clinical practice. The Partin tables are the most popular and widely used tool for predicting the pathologic stages of prostate cancer because of its high accuracy and ease of implementation. The aim of this study was to externally validate the accuracy of the three versions of the Partin tables in predicting the post-operative pathologic stages in Chinese patients with prostate cancer.</p><p><b>METHODS</b>We retrospectively analyzed the data from 203 patients with prostate cancer who underwent radical prostatectomies between June 2000 and May 2012. The accuracies of the three versions of the Partin tables in predicting the post-operative pathologic stages in Chinese patients with prostate cancer were evaluated using the area under the receiver operating characteristic curve (AUC).</p><p><b>RESULTS</b>Using the 1997, 2001, and 2007 Partin tables for predicting the current cases, the AUC of organ confinement (OC) was 0.877, 0.788, and 0.726; the AUC of extracapsular extension (ECE) was 0.525, 0.615, and 0.608; the AUC of seminal vesicle invasion (SVI) was 0.875, 0.649, and 0.820; and the AUC of pelvic lymph node invasion (LNI) was 0.808, 0.758, and 0.735 respectively.</p><p><b>CONCLUSIONS</b>The accuracies of the three versions of Partin tables in predicting OC, SVI, and LNI were good, especially the 2001 Partin table for SVI. In contrast, the accuracy of the three versions of the Partin tables in predicting ECE was fair. The 1997 Partin table was much better than the 2007 table in predicting OC, and the 2001 table in predicting SVI. The 2007 Partin table did not show any advantages.</p>