多參數磁振造影增加預測達文西攝護腺全切除術後病理結果的準確度
方仁愷、謝博帆、黃志平、吳錫金、葉進仲、楊啟瑞、陳冠亨、張議徽、黃煒軒、李聖偉、賴俊佑、張菡1、林維卿2、張兆祥*
中國醫藥大學附設醫院泌尿部，1病理部，2放射部
Multiparametric Magnetic Resonance Imaging increases predictive accuracy of oncological outcomes after robot-assisted radical prostatectomy
Jen-Kai Fang, Po-Fan Hsieh, Chi-Ping Huang, Hsi-Chin Wu, Chin-Chung Yeh, Chi-Rei Yang, Guan-Heng Chen, Yi-Huei Chang, Wei-Hsuan Huang, Sheng-Wei Li, Chun-Yu Lai, Han Chang1, Wei-Ching Lin2, Chao-Hsiang Chang*
Department of Urology, 1Pathology and 2Radiology, China Medical University Hospital, Taichung, Taiwan
Purpose: Robot-assisted radical prostatecotmy (RARP) is one of the standard treatment for localized prostate cancer (PCa). Cinical models including Partin table and Cancer of the Prostate Risk Assessment(CAPRA) score have been developed to predict pathological outcomes and recurrent rate after radical prostatectomy. In this study we aimed to analyze the added value of preoperative multiparametric magnetic resonance imaging (mpMRI) in predicting oncological outcomes.
Materials and Methods: We retrospectively reviewed patients with PCa who underwent RARP and bilateral pelvic lymph node dissection (BPLND) from September 2015 to December 2017. Exclusion criteria was patients with metastatic PCa and those without preoperavei mpMRI. Prostate mpMRI was read by an experienced uroradiologist according to the criteria of Prostate Imaging Reporting and Data System (PI-RADS) version 2. Pathology was reported by an expert uropathologist. Multivariate logistic regression was developed based on clinical variables, PartinTable, CAPRA score and MRI to predict oncological outcomes, including the occurence of extracapsular extention (ECE), seminal vesicle invasion (SVI), lymph node metastasis and biochemical failure. We draw the ROC curves of the clinical and imaging predictive models and compared the AUC curve of these models.
Results: From September 2015 to December 2017, 155 patientes underwent RARP and BPLND. Fourteen patients were excluded due to no prepoerative mpMRI. At last,141 patients were analyzed. For predicting ECE, SVI and lymph node metastasis, CAPRA score had higher odds ratio compared to PARTIN table (1.161 vs. 1.006, 1.816 vs. 1.055, 1.241 vs. 0.919, respectively). mpMRI was also an independent predictor of ECE and SVI. Furthermore, according to ROC curve, the combination of CAPRA score and mpMRI had more AUC than CAPRA alone for predicting ECE (0.793 vs 0.747), SVI(0.927 vs 0.84) and lymph node metastasis(0.819 vs 0.802)
Conclusions: mpMRI before RARP could be used to predict ECE and SVI. In addition, the combination of mpMRI and CAPRA score could help us better predict the oncological outcomes.