Session Item

Unlike radical prostatectomy, radiotherapy lacks a definitive pathological assessment of performance, and consequently patients can be under- or overtreated. The purpose of this study was to evaluate the ability of radiomic features to improve the accuracy of non-invasive prediction of pathological features of prostate cancer with prostatectomy as confirmation.

A representative subset of 100 patients from a cohort of roughly 1500 who have undergone mpMRI of the prostate and prostatectomy in our Institution since 2015 was selected. The prostate of each patient was segmented from T2-weighted MR images by an expert radiologist, and used in the extraction of 1810 radiomic features (PyRadiomics 3.0.1, AIM Harvard), successively reduced to 50. Gradient-boosted decision tree models were separately trained using clinical features (age, prostate volume, iPSA, PI-RADS category, biopsy-based total Gleason score, ISUP grade, and risk class) alone and in combination with the radiomic features to predict surgical marginal status (R0 vs R1), the presence of pathological lymph nodes (pN0 vs pN1), pathological tumor stage (pT2 vs pT3), and ISUP grade group (≤3 vs ≥4) and validated with 32-times repeated 5-fold cross validation. The models were evaluated and compared in terms of their AUC values.

The addition of radiomics features led to increases of AUC ranging between 0.061 (pT) and 0.139 (ISUP grade group) as illustrated in Figure 1 and summarized in Table 1. All AUC gains were statistically significant at a level of at least 0.0001 (Mann-Whitney U-test).

Figure 1. ROC curves and AUC values for the prediction models of different outcomes.

Table 1. AUC values and 95% confidence intervals over repeated validation folds of the trained models*

*all differences between clinical and radiomic models significant at p<0.0001 (Mann-Whitney U-test)

Our results highlight the potential benefit of whole-prostate radiomics for prediction of all the examined pathological features of prostate cancer, with AUC values in the 0.80-88 range. Literature models have used baseline clinical and mpMRI-based variables to predict cancer aggressiveness. The potential of a radiomic plus clinical feature model to better predict pathological features of prostate cancer, and in particular extraprostatic extension and pelvic lymph node involvement, is of considerable interest for guiding the clinical decision-making process and can provide valuable information for personalizing therapy. These preliminary but promising results will be validated in the larger cohort of 1500 patients.