Association Between Functional Multiparametric MRI Features and Tumour Immune Microenvironment Phenotypes in Prostate Cancer: A Retrospective Study.

OBJECTIVE

This study aimed to investigate the association between functional multiparametric magnetic resonance imaging (mpMRI) features and tumour immune microenvironment (TIME) phenotypes (immune-inflamed, immune-excluded and immune-desert) in prostate cancer (PCa) and to develop a multiclass prediction model.

METHODS

This retrospective study enrolled 228 patients with PCa who underwent mpMRI between January 2023 and October 2025. The patients were categorised into three TIME phenotypes based on immune cell infiltration patterns. Functional mpMRI features (apparent diffusion coefficient (ADC)_mean, ADC_min, Prostate Imaging Reporting and Data System (PI-RADS) score and lesion diameter) and clinicopathological indices were collected. Variables identified through univariate analysis were entered into a multinomial logistic regression model to assess independent factors associated with TIME phenotypes. Model performance was evaluated through 10-fold cross-validation, receiver operating characteristic (ROC) analysis and confusion matrix evaluation.

RESULTS

Significant differences in prostate-specific antigen (PSA) level, International Society of Urological Pathology (ISUP) grade and mpMRI feature were observed among the three TIME phenotypes (p < 0.05). The immune-inflamed phenotype was associated with higher PSA and lower ADC_mean. The immune-desert phenotype was more common in lesions with low PI-RADS scores. Elevated PSA and reduced ADC_mean were independent factors associated with the immune-inflamed phenotype, whereas a low PI-RADS score was significantly associated with the immune-desert phenotype. ROC analysis showed area under the curve values of 0.654 (95% CI: 0.557-0.691) for immune-inflamed, 0.742 (95% CI: 0.654-0.791) for immune-excluded and 0.802 (95% CI: 0.782-0.852) for immune-desert phenotypes.

CONCLUSIONS

Functional mpMRI features show significant associations with TIME phenotypes in PCa. Elevated PSA levels and low ADC_mean value may serve as indicators of the immune-inflamed phenotype, whereas a low PI-RADS score may suggest an immune-desert phenotype. These findings highlight the potential of mpMRI as a non-invasive imaging surrogate for assessing TIME phenotypes and inform precision immunotherapy in PCa.