Quantitative and Qualitative Parameters of DCE-MRI Predict CDKN2A/B Homozygous Deletion in Gliomas

Abstract

Rationale and objectives: Homozygous deletion (HD) of CDKN2A/B holds important prognostic value in gliomas. This study aimed to explore the predictive potential of conventional MRI characteristics combined with dynamic contrast-enhanced MRI parameters in predicting CDKN2A/B HD status in gliomas.

Materials and methods: Preoperative MRI data of 105 patients (69 without CDKN2A/B HD, and 36 with CDKN2A/B homozygous deletion) with gliomas were retrospectively collected. Conventional MRI features and dynamic contrast-enhanced-MRI qualitative parameter time-intensity curve type, quantitative parameters K^trans, K_ep, V_e, V_p, and iAUC were obtained. Logistic regression models for prediction of CDKN2A/B HD status were constructed in all types of gliomas and both subtypes of IDH-mutant and IDH-wild gliomas.

Results: Multivariate analysis for all patients demonstrated that age (OR=1.103, p = 0.002) and K^trans (OR=1.051, p < 0.001) independently predicted CDKN2A/B HD. In IDH-mutant subgroup, multivariate analysis results indicated that K^trans (OR=1.098, p = 0.031) emerged as autonomous predictors of CDKN2A/B HD. In IDH-wild subgroup, age (OR=1.111, p = 0.002) and K^trans (OR=1.032, p = 0.001) were independent predictors of CDKN2A/B HD according to the multivariate analysis. The areas under the receiver operating characteristic curve of the corresponding models were 0.90, 0.95 and 0.84, respectively.

Conclusion: K^trans can serve as valuable predictive parameters for identifying CDKN2A/B HD status in all types of gliomas and both subtypes of IDH-mutant and IDH-wild gliomas. These findings provide a foundation for precise preoperative non-invasive diagnosis and personalized treatment approaches for glioma patients.

Keywords: CDKN2A/B homozygous deletion; DCE-MRI; Gliomas; Predict.