Quantitative parameters on whole-body diffusion weighted MRI (WB-DWMRI) and PSMA PET/CT in ¹⁷⁷ Lu-PSMA therapy for metastatic prostate cancer (mPC).

Abstract

e17034 Background: This original study explores the utility of quantitative imaging biomarkers from WB-DWMRI and PSMA-PET/CT in predicting lesion-level response to 177 Lu-PSMA Therapy in mPC. Treatment resistance and failure may be influenced by heterogeneity in response. Methods: Twenty-eight mPC patients were analysed, each undergoing WB-DWMRI and PSMA-PET/CT within 3 months prior to 177 Lu-PSMA therapy. Data was gathered from the 5 hottest lesions identified on PSMA-PET/CT, with baseline PSMA SUV (mean and peak) and DWI metrics ADC (mean, kurtosis and volume) recorded. Response was evaluated using a modified lesion-level adaptation of PERCIST 1.1 for PSMA PET/CT and MET-RADS-P criteria for WB-DWMRI. Statistical analysis accounted for intra-patient correlation using Huber-White robust standard errors. Lesion volume was log10-transformed. Second degree fractional polynomial transformations optimised fit. Multilevel logistic regression and AUROC were used to evaluate biomarker performance. Results: Of 140 lesions reviewed, 65 bone and 30 lymph node lesions were suitable for analysis. Response rates were 68% for bone lesions and 53% for lymph nodes. SUVmean and SUVpeak were highly correlated (rank-correlation=0.97) and demonstrated strong predictive value (AUROCs: 0.74 and 0.75, respectively). Lesion volume exhibited moderate predictive performance (AUROC=0.69) and correlated moderately with SUVmean (rank-correlation=0.43). Neither ADCmean (AUROC=0.45, p=0.48) nor kurtosis (AUROC=0.49, p=0.17) were prognostic. In bone lesions, a 1-unit increase in volume raised response odds by 4.3 (95% CI 0.7–27), while in lymph nodes, odds increased by 6.2 (95% CI 0.6–67). A 1-unit SUVmean increase raised response odds by 1.5 (95% CI 1.1–2.0) in bone lesions and 1.2 (95% CI 1.0–1.4) in lymph nodes. No evidence suggested combining volume with SUVmean enhanced predictive performance over SUVmean alone (p=0.58). Conclusions: This study is the first to evaluate combined PSMA-PET/CT and WB-DWMRI parameters for predicting lesion-level responses to 177 Lu-PSMA therapy in mPC. SUVmean, SUVpeak, and DWI-derived volume are promising biomarkers, the former two demonstrating stronger predictive potential. However, ADC mean and kurtosis lacked prognostic value. Integrating metabolic and structural imaging biomarkers may improve treatment stratification and response monitoring. Future research should prioritise prospective studies with standardised imaging protocols, larger patient cohorts, and entire tumour volume analysis to improve predictive models and their clinical applicability. Modality Progressive Disease (PD) Stable Disease Response (R) PSMA PET/CT SUVpeak ↑ ≥30% and absolute ↑ ≥0.8 units Between PD and R SUVpeak ↓ ≥30% and absolute ↓ ≥0.8 units WB-DWMRI Volume ↑ >20% with ADC 600–1000 μm²/s Between PD and R ADC ↑ >25% and/or volume ↓ >30%