Conclusions
Our study demonstrates that combining miRNA markers with MRI-based radiomics improves the identification of clinically aggressive prostate cancer.
Methods
We conducted a study on prostate cancer patients, analyzing baseline blood plasma and MRI data. Exosomes were isolated from blood plasma samples to quantify miRNAs, while MRI scans provided detailed tumor morphology. Radiomics features from MRI and miRNA expression data were integrated to develop predictive models, which were evaluated using ROC curve analysis, highlighting the multivariable model's effectiveness.
Results
Our findings indicate that the univariate feature-based model with the highest Youden's index achieved average areas under the receiver operating characteristic (ROC) curve of 0.76, 0.82, and 0.84 for miRNA, MR-T2W, and MR-ADC features, respectively, in identifying clinically aggressive (Gleason grade) disease. The multivariable feature-based model yielded an average area under the curve (AUC) of 0.88 and 0.95 using combinations of miRNA markers with imaging features in MR-ADC and MR-T2W, respectively. Conclusions: Our study demonstrates that combining miRNA markers with MRI-based radiomics improves the identification of clinically aggressive prostate cancer.