Proposal for a predictive model of erectile function after permanent (125)I prostate brachytherapy for localized prostate cancer.

To determine the risk factors for EDin men treated by prostate brachytherapy (PB) for localized prostate cancer and to propose a model to predict post-implant erectile function. Out of a series of 270 sexually active men treated by PB, 241 (89%) (mean age=66 years (range, 43-80)) accepted to participate in a mail-based study on erectile function. The risk factors for erectile dysfunction were determined by regression analysis and a predictive model was proposed. The performance of the model was determined in this population and subsequently verified in a population of 50 men treated by PB in another treatment center. The risk factors for ED after PB were age, the pre-implant IIEF score and prostate volume. In the studied population, the final model to predict a post-treatment IIEF-5 score, using these factors, had a sensitivity of 69% and a specificity of 68% associated to an area under the ROC curve (AUC) of 0.75. The same performance was obtained in another treatment center. Age, pre-implant IIEF-5 score and prostate volume may be used to predict post-implant erectile function in patients treated by PB.