171 Background: To evaluate the accuracy of preoperative magnetic resonance imaging (MRI) in patients who underwent pelvic lymph node dissection (PLND). Methods: The data of 1528 patients who underwent radical prostatectomy and PLND from 2003 to 2017 in our institution were retrospectively reviewed. We evaluated the various clinicopathologic variables including preoperative MRI and pathologic lymph node metastasis (LNmet). The prediction model for pathologic lymph node (LNmet) was assessed using logistic regression analyses and the areas under receiver operating characteristic curves (AUCs) were evaluated. Results: The mean age of our cohort was 66.4 ± 6.7 years. Positive preoperative MRI finding was observed in 9.4% (145/1528) of patients. 5.3% (81/1528) of patients accompanied confirmed final pathologic LNmet. Sensitivity and specificity of preoperative MRI were 30.8% and 91.7%, respectively. Multivariate analysis showed that preoperative MRI findings, clinical stage, and biopsy Gleason score were independent predictors for pathologic LNmet. Preoperative PSA, which was significantly related to pathologic LNmet in univariate analysis, failed to achieve independent predictor status in multivariate analysis. A better prediction model with greater accuracy was achieved by applying multivariate ROC analysis that included MRI findings, clinical stage, and biopsy Gleason score (AUC: 0.799 vs 0.613, p<0.01). The corresponding newer prediction model showed better sensitivity(77.78%) and specificity(70.74%) within threshold value range when it was compared with the predictor model using conventional Partin triad. Conclusions: Preoperative MRI findings for pathologic LNmet showed limited prediction power, yet the predicting power was significantly increased when additional factors such as clinical stage, and biopsy Gleason score were included in the analysis. [Table: see text]