Abstract Purpose To determine whether post-implant MRI-based dosimetry of the Dominant Intra-prostatic Lesion (DIL) could best predict the occurrence of PSA bounce after prostate brachytherapy. Methods and materials We selected 66 patients with a low risk prostate cancer treated with125 I prostate brachytherapy as monotherapy. Post-implant dosimetry based on day 30 CT-scan and multiparametric MRI co-registration was generated: planned D 90, D 95, V 100, V 150 values were calculated for each DIL. Bounce was defined as a PSA elevation ⩾0.2 ng/mL from the previous baseline value followed by a decrease to or below the prior nadir with no additional treatment. Results After a median follow-up of 35.5 months (range 13.2–72.5), a PSA bounce occurred in 24 (36.4%) patients. The mean planned D 90 of the DIL was significantly lower in bouncers: 196 ± 61 Gy vs. 234 ± 62 Gy, p = 0.018. The mean planned V 150 of the DIL was 56 ± 32% for bouncers while it was 75 ± 30% for non-bouncers, p = 0.026. Conclusion A lower planned D 90 or V 150 in the DIL were predictive of PSA bounce after prostate brachytherapy. PSA bounce could be caused by delayed cell death related to sublethal damage accumulation in the tumor.