Noninvasive Prediction of IDH1 Mutation and ATRX Expression Loss in Low-Grade Gliomas Using Multiparametric MR...

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Accurate molecular stratification of low-grade gliomas (LGGs) is crucial for patient management.
• Noninvasive detection of isocitrate dehydrogenase 1 mutation (IDH1(+)) and alpha thalassemia/mental retardation syndrome X-linked (ATRX(-)) expression is clinically significant.

Purpose of the Study:

• To investigate a radiomic approach utilizing multiparametric MRI for noninvasive determination of IDH1(+) and ATRX(-) status in LGG patients.

Main Methods:

• A retrospective study involving 57 LGG patients.
• Extracted 265 radiomic features from T2 FLAIR, 3D-ASL (CBF), DWI (ADC, eADC) sequences.
• Utilized Support Vector Machine with Recursive Feature Elimination (SVM-RFE) for feature selection and Receiver Operating Characteristic (ROC) analysis for model evaluation.

Main Results:

• Achieved high accuracy in predicting IDH1(+) status (94.74% accuracy, 0.931 AUC) and ATRX(-) status in IDH1(+) LGGs (91.67% accuracy, 0.926 AUC).
• Optimal predictive models were established using selected radiomic features from multiparametric MRI.

Conclusions:

• Multiparametric MRI-based radiomics offers a promising, noninvasive strategy for predicting IDH1 and ATRX molecular subtypes in LGGs.
• This approach can aid in the molecular stratification of LGGs, potentially improving treatment decisions.