INTRODUCTION: Treatment decision-making for intermediate-risk prostate cancer (CaP) is mainly based on grade and tumor involvement on systematic biopsy. We aimed to assess the added value of multi-parametric magnetic resonance imaging (mpMRI) and targeted biopsy (TB) features for predicting final pathology and for improving the well-established favourable/unfavourable systematic biopsy-based sub-classification. MATERIALS AND METHODS: From a prospective database of 377 intermediate risk CaP cases, we evaluated the performance of the standard intermediate risk classification (IRC), and the predictive factors for unfavourable disease on final pathology aiming to build a new model. Overall unfavourable disease (OUD) was defined by any pT3-4 and/or pN1 and/or grade group (GG) ≥ 3. RESULTS: The standard IRC was found to be predictive for unfavourable disease in this population. However, in multivariable analysis regression, ECE on mpMRI and GG ≥3 on TB remained the 2 independent predictive factors for OUD disease (HR = 2.7, P = 0.032, and HR = 2.41, P = 0.01, respectively). By using the new IRC in which unfavorable risk was defined by ECE on mpMRI and/or GG ≥3 on TB, the proportion of unfavorable cases decreased from 62.3% to 34.1% while better predicting unfavorable disease in RP speciments. The new model displayed a better accuracy than the standard IRC for predicting OUD (AUC: 0.66 vs. 0.55). CONCLUSIONS: The integration of imaging and TB features drastically improves the intermediate risk sub-classification performance and better discriminates the unfavourable risk group that could benefit from more aggressive therapy such as neo-adjuvant and/or adjuvant treatment, and the favourable group that could avoid over-treatment. External validation in other datasets is needed.
INTRODUCTION: Treatment decision-making for intermediate-risk prostate cancer (CaP) is mainly based on grade and tumor involvement on systematic biopsy. We aimed to assess the added value of multi-parametric magnetic resonance imaging (mpMRI) and targeted biopsy (TB) features for predicting final pathology and for improving the well-established favourable/unfavourable systematic biopsy-based sub-classification. MATERIALS AND METHODS: From a prospective database of 377 intermediate risk CaP cases, we evaluated the performance of the standard intermediate risk classification (IRC), and the predictive factors for unfavourable disease on final pathology aiming to build a new model. Overall unfavourable disease (OUD) was defined by any pT3-4 and/or pN1 and/or grade group (GG) ≥ 3. RESULTS: The standard IRC was found to be predictive for unfavourable disease in this population. However, in multivariable analysis regression, ECE on mpMRI and GG ≥3 on TB remained the 2 independent predictive factors for OUD disease (HR = 2.7, P = 0.032, and HR = 2.41, P = 0.01, respectively). By using the new IRC in which unfavorable risk was defined by ECE on mpMRI and/or GG ≥3 on TB, the proportion of unfavorable cases decreased from 62.3% to 34.1% while better predicting unfavorable disease in RP speciments. The new model displayed a better accuracy than the standard IRC for predicting OUD (AUC: 0.66 vs. 0.55). CONCLUSIONS: The integration of imaging and TB features drastically improves the intermediate risk sub-classification performance and better discriminates the unfavourable risk group that could benefit from more aggressive therapy such as neo-adjuvant and/or adjuvant treatment, and the favourable group that could avoid over-treatment. External validation in other datasets is needed.