Integrated ⁶⁸Ga-PSMA PET/MRI models improve prostate cancer diagnosis and ISUP stratification, especially in the...

To evaluate the diagnostic efficacy of ⁶⁸Ga-PSMA PET/MRI-derived parameters in prostate cancer and their significance in risk stratification.

A retrospective study was performed on patients who underwent ⁶⁸Ga-PSMA PET/MRI from November 2020 to April 2024, all with histopathological confirmation obtained from biopsy or prostatectomy. Parameters analyzed included SUV_max, PSMA tumor volume at a 40% SUV_max threshold (PTV40%), mean uptake in this volume (SUV_40%mean), and uptake ratios to the liver (PLR), blood pool (PBpR), and parotid gland (PPgR). MRI was evaluated using PI-RADS v2.1 (scores ≥ 4 were defined as positive), and ADC_mean and ADC_min were recorded. Diagnostic performance of PET and MRI parameters was compared using ROC analysis. Two logistic regression model sets were constructed: one for improved diagnosis performance in the low-PSA subgroup (PSA < 20 ng/mL) and one for International Society of Urological Pathology (ISUP) Grade Group (GG) stratification (GG 1-2 vs. GG 3-5). Model performance was assessed via LOOCV, 1,000 bootstrap resamples, and decision curve analysis.

Of the 74 patients (mean age 65.9 ± 7.4 years), 52 had prostate cancer and 22 had benign lesions. PET parameters (SUV_max, PLR, PBpR, PPgR) and MRI-derived ADC values differed significantly between groups. An SUV_max cut-off of 7.665 yielded 76.9% sensitivity, 100% specificity (AUC = 0.920), while a PLR cut-off of 1.551 showed higher sensitivity (92.3%) and specificity (86.4%) (AUC = 0.907). In patients with PSA < 20 ng/mL, PLR showed higher sensitivity (84.6%) than SUV_max (53.8%). For the low-PSA subgroup, the Clinical+ADC_mean+SUV_max model yielded a LOOCV AUC of 0.795, as opposed to 0.449 for the clinical-only model and 0.733 for the clinical+ADC_mean model. For ISUP stratification, the ADC_mean+SUV_max model reached a LOOCV AUC of 0.906. DCA validated the clinical utility of the optimal models.

⁶⁸Ga-PSMA PET/MRI-derived parameters exhibit excellent diagnostic performance for prostate cancer. PLR demonstrates the highest sensitivity, especially in patients with low PSA. Combining PET and MR parameters-especially PET-guided ADC_mean with SUV_max and PLR-improves diagnosis, potentially enables non-invasive ISUP Grade Group stratification, and may facilitate personalized management.

NCT03756077. Registered 27 November 2018-Retrospectively registered, https://clinicaltrials.gov/show/NCT03756077 .