Search

Purpose: Comparing mpMRI and 68Ga-PSMA-PET/MRI in primary staging of PCa and investigating the value of quantitative mpMRI-measurements for prediction of extracapsular extension and N-metastases.Methods: Patients with PCa undergoing 68Ga-PSMA-PET/MRI and mpMRI during January 2016 to February 2019 were retrospectively included. Two readers each on 68Ga-PSMA-PET/MRI or mpMRI rated extraprostatic extension (≥T3) and regional lymph-node-metastasis (N1) on a Likert-scale. A fifth reader measured tumor volume, maximum diameter, and capsular contact length on mpMRI. Probability of lymph-node-metastasis was additionally calculated using the 2018 Briganti model. Interobserver-agreement was assessed by squared Cohen's kappa, and diagnostic accuracy was determined using radical prostatectomy (n = 35/49) as reference standard.Results: 49 patients (median age 66 years [IQR: 61-72 years]) were evaluated. Interobserver-agreement for mpMRI and 68Ga-PSMA-PET/MRI was: ≥T3: κ = 0.58/0.47; N1: κ = 0.55/0.92. Diagnostic accuracy for mpMRI vs 68Ga-PSMA-PET/MRI readers for ≥ T3 was AUC: 0.72, 0.62 vs 0.71, 0.72 (p > 0.38) and for N1 was AUC: 0.39, 0.55 vs 0.72, 0.78 (p < 0.01). Quantitative parameters delivered diagnostic accuracies of: AUC: 0.70-0.72 for ≥ T3. The 2018 Briganti model achieved an AUC of 0.89 for N1.Conclusions: Interreader-agreement regarding ≥ T3 was similar for mpMRI and 68Ga-PSMA-PET/MRI while for N1 it was higher for 68Ga-PSMA-PET/MRI. Diagnostic accuracy was comparable for ≥ T3 while for N1 it was higher in 68Ga-PSMA-PET/MRI and the 2018 Briganti model. Combining clinical data and quantitative data from mpMRI in the 2018 Briganti model yielded the highest AUC for prediction of lymph node metastasis and may aid in selecting patients who will benefit from 68Ga-PSMA-PET/MRI for primary staging. [ABSTRACT FROM AUTHOR]

Objectives: To evaluate whether a magnetic resonance imaging (MRI) radiomics-based machine learning classifier can predict postoperative pancreatic fistula (POPF) after pancreaticoduodenectomy (PD) and to compare its performance to T1 signal intensity ratio (T1 SIratio).Methods: Sixty-two patients who underwent 3 T MRI before PD between 2008 and 2018 were retrospectively analyzed. POPF was graded and split into clinically relevant POPF (CR-POPF) vs. biochemical leak or no POPF. On T1- and T2-weighted images, 2 regions of interest were placed in the pancreatic corpus and cauda. 173 radiomics features were extracted using pyRadiomics. Additionally, the pancreas-to-muscle T1 SIratio was measured. The dataset was augmented and split into training (70 %) and test sets (30 %). A Boruta algorithm was used for feature reduction. For prediction of CR-POPF models were built using a gradient-boosted tree (GBT) and logistic regression from the radiomics features, T1 SIratio and a combination of the two. Diagnostic accuracy of the models was compared using areas under the receiver operating characteristics curve (AUCs).Results: Five most important radiomics features were identified for prediction of CR-POPF. A GBT using these features achieved an AUC of 0.82 (95 % Confidence Interval [CI]: 0.74 - 0.89) when applied on the original (non-augmented) dataset. Using T1 SIratio, a GBT model resulted in an AUC of 0.75 (CI: 0.63 - 0.84) and a logistic regression model delivered an AUC of 0.75 (CI: 0.63 - 0.84). A GBT model combining radiomics features and T1 SIratio resulted in an AUC of 0.90 (CI 0.84 - 0.95).Conclusion: MRI-radiomics with routine sequences provides promising prediction of CR-POPF. [ABSTRACT FROM AUTHOR]