Predictive value of DCE-MRI quantitative parameters for long-term postoperative prognosis in brain glioma patients

To investigate the value of DCE-MRI quantitative parameters in predicting long-term postoperative prognosis in brain glioma patients. A retrospective cohort study was conducted, including 165 brain glioma patients who underwent surgical treatment between January 2021 and December 2022, all of whom completed preoperative DCE-MRI examination. The Extended Tofts model was used to analyze DCE-MRI quantitative parameters, including Ktrans, Kep, Ve, Vp, and AUC. Absolute and relative values of each parameter were calculated (using contralateral normal white matter as reference). ROC curves were used to determine optimal cutoff values. Kaplan-Meier method and Cox regression analysis were applied to evaluate the relationship between DCE-MRI parameters and postoperative OS and PFS, and prognostic prediction models were constructed. The median follow-up time was 42.8 months, with 63 patients (38.2%) deceased and 98 patients (59.4%) experiencing tumor recurrence. All DCE-MRI quantitative parameters in the deceased group were significantly higher than those in the survival group (P<0.05 after FDR correction), with relative Ktrans showing the most significant difference (3.7 vs 2.5, P<0.001, FDR-corrected P<0.001). ROC analysis showed that relative Ktrans had the highest efficacy in predicting 3-year survival (AUC=0.758), with an optimal cutoff value of 3.1. Multivariate Cox regression analysis identified WHO grade, IDH1/2 mutation status, relative Ktrans, Ki-67 proliferation index, and MGMT methylation status as independent prognostic factors for overall survival. Patients with relative Ktrans≥3.1 had a 94% increased risk of death compared to those with lower values (HR=1.94, 95%CI: 1.14-3.30, P=0.014). The prognostic prediction model incorporating DCE-MRI parameters had a C-index of 0.795, significantly superior to the base model (NRI=0.248, P<0.001). DCE-MRI quantitative parameters, particularly relative Ktrans, are independent predictors of postoperative long-term prognosis in glioma patients. DCE-MRI parameters can non-invasively assess tumor vascular functional status and, when combined with traditional pathological molecular markers, can significantly improve prognostic prediction accuracy, providing important imaging biomarkers for clinical individualized treatment decision-making.