To investigate the diagnostic value of an extracellular volume (ECV) model integrating intratumoral(I) and peritumoral (P) regions (ECV-IP) for prostate cancer (PCa), and to validate the mechanism by which peritumoral ECV (ECV-P) indirectly drives malignant progression via modulation of the intratumoral microenvironment.

A retrospective study was conducted on 117 patients with prostatic disease, 63 cases of benign prostatic hyperplasia (BPH), 54 cases of PCa, the age of patients was (69.67±8.64) years. Imaging parameters including intratumoral apparent diffusion coefficient (ADC-I), peritumoral apparent diffusion coefficient (ADC-P), intratumoral extracellular volume (ECV-I), ECV-P, ECV-IP were compared. Diagnostic performance was evaluated using binary logistic regression modeling, receiver operating characteristic (ROC) curves with DeLong's test, 5-fold cross-validation, and the underlying mechanism was assessed via bootstrap-mediated effect analysis (5 000 resamples).

The ECV-IP model achieved an AUC of 0.883 (95%CI: 0.800-0.941) for PCa diagnosis, with a sensitivity of 87.2% and a specificity of 83.0%. It demonstrated significantly superior diagnostic performance compared to single-region parameters (ECV-I: AUC=0.854, P=0.017; ADC-I: AUC=0.797, P=0.032), PSA (AUC=0.775, P=0.002), and the ADC-IP model (AUC=0.810, P<0.001). Cross-validation demonstrated robust performance: training set (AUC=0.865, accuracy=85.6%); test set (AUC=0.889, accuracy=70.6%). ECV-P was significantly lower in PCa than BPH [0.28 (0.17, 0.38) vs 0.36 (0.25, 0.49), P=0.045] and showed a positive correlation with intratumoral ADC-I (r=0.278, P=0.007). A 1-unit increase in ECV-P increased ADC-I by 256.945×10^-6mm²/s (P=0.027). A decrease of 100×10^-6 mm²/s in ADC-I increased malignant risk (β=-0.001, P<0.001).The indirect effect of ECV-P on PCa progression mediated by ADC-I was significant (β=-0.294, 95%CI: -0.626 to -0.049, P=0.048; proportion mediated: 59.2%). The direct effect of ECV-P on PCa was non-significant (P=0.366). This validates that ECV-P indirectly drives malignant progression of tumors by down-regulating the diffusion of water molecules.

The ECV-IP model, as a dual-region approach, significantly enhances PCa diagnostic accuracy over conventional single-region parameters and PSA.This study provides the first imaging-based validation of the "ECV-P leads to the malignant progression of PCa by influencing ADC-I" indirect regulatory mechanism, establishing ECV-IP as a novel non-invasive biomarker for targeting the prostate cancer tumor microenvironment.