Radiomics Strategy for Molecular Subtype Stratification of Lower-Grade Glioma: Detecting IDH and TP53 Mutations Based...

Area of Science:

• Radiology and Medical Imaging
• Oncology
• Genetics

Background:

• Noninvasive detection of isocitrate dehydrogenase (IDH) and TP53 mutations is crucial for molecular stratification of lower-grade gliomas (LGG).
• Accurate molecular subtyping aids in personalized treatment strategies for LGG patients.

Purpose of the Study:

• To identify magnetic resonance imaging (MRI) features indicative of IDH and TP53 mutations in LGG.
• To develop and validate a radiomics strategy for noninvasive detection of these mutations.

Main Methods:

• A retrospective analysis of 103 LGG patients using T1-weighted, T2-weighted, and fluid-attenuation inversion recovery MRI sequences.
• Support vector machine-based recursive feature elimination (SVM-RFE) was employed to select optimal radiomics features.
• Support vector machine (SVM) models were trained and validated for mutation detection.

Main Results:

• SVM models achieved high accuracies for IDH and TP53 mutation detection in both development (up to 92.0%) and validation (up to 85.0%) cohorts.
• Significant differences in histopathologic subtypes were observed among molecular subgroups (P=0.017).
• T2-weighted image features were found to be more critical than those from other sequences.

Conclusions:

• A radiomics approach integrating multimodal MRI features and SVM models effectively stratifies LGG patients based on IDH and TP53 mutations.
• The proposed method demonstrates promising efficiency for noninvasive molecular subtyping of LGG.