The median age and BMI at baseline, median prostate-specific-antigen at diagnosis of prostate cancer, and median prostate volume at PI were 66 years (range; 50-78 years), 23.7 kg/m2 (range; 16.0-39.9 kg/m2), 6.3 ng/ml (range; 1.9-38.3 ng/ml), and 21.6ml (range; 8.6-57.6 ml), respectively. The median followup duration was 84 months (range; 1-168). The median Gleason score was 7 (range; 5-9). Based on modified D’Amico risk classification, 135 patients (43.1%), 157 patients (50.2%) and 21 patients (6.7%) were low risk, intermediate risk and high risk, respectively. Neoadjuvant ADT was administered to 259 patients (82.7%). The PI-alone consisted of 189 patients (60.4%) and the PI + EBRT consisted of 124 patients (39.6%). The dosimetry data of BED, UD5 were 191.8Gy (range; 121.0-261.5Gy) and 220.4Gy (range; 110.5-395.3Gy), respectively. The mean IPSS, OABSS, maximal voiding rate (Qmax), voided volume, and post-voided residual before PI were 6.6 (range 0-29), 3.1 (range; 0-10), 18.9 ml/s (range; 6.5-49.5ml/s), 293.1ml (range; 11.6-844ml), and 14.8ml (range; 0-74ml), respectively. IPSS increased at 3 months after PI with the mean score 17.3, and the score decreased at 36 months with the mean score 7.7. Non-linear regression analysis revealed that duration was significantly related change of IPSS (p<0.001) and 36 months after treatment with PI the score remained consistent level. In addition, pretreatment IPSS significantly modified the association between change of IPSS and month (p=0.023). OABSS also increased at 3 months after PI with the mean score 7.0, and the score decreased at 9 months after PI with the mean score 4.0. Non-linear regression analysis revealed that duration was significantly related change of OABSS (p<0.001). Interestingly, non-linear regression analysis revealed that OABSS increased again after 60 months unlike IPSS. In addition, pretreatment OABSS significantly modified the association between change of OABSS and month (p=0.017). Non-linear regression analysis of uroflowmetry revealed transient deterioration of Qmax, voided volume, and post-voided residual, and these factors recovered until 36 months after PI. It is noteworthy that Qmax and post-voided residual remain after improvement, voided volume became decreased again. Qmax, voided volume and post-voided residual did not modify the relationship between outcome and month significantly, but we could confirm weak tendency for effect modification (p-value for interaction was p=0.06, p=0.151 and p=0.057.).