Two-Stage Training Framework Using Multicontrast MRI Radiomics for IDH Mutation Status Prediction in Glioma.

Area of Science:

• Radiology
• Oncology
• Medical Imaging Analysis

Background:

• Isocitrate dehydrogenase (IDH) mutation status is a critical prognostic indicator in glioma.
• Accurate preoperative prediction of IDH status can guide treatment decisions and improve patient outcomes.
• Current methods for determining IDH status may be invasive or time-consuming.

Purpose of the Study:

• To develop and validate a radiomics framework using preoperative MRI to predict IDH mutation status in glioma.
• To assess the framework's performance across multiple diverse datasets.

Main Methods:

• Radiomics features were extracted from preoperative MRI scans (whole tumor, non-enhancing, necrosis, edema regions).
• Boruta algorithm was used for feature selection, and random forest/XGBoost models were trained on balanced data.
• The framework was evaluated on six retrospective datasets (TCIA, UCSF, EGD, UTSW, NYU, UWM) with over 2000 patients.

Main Results:

• Models trained on the TCIA dataset achieved AUCs ranging from 0.86 to 0.94 across test sets.
• Models trained on the UTSW dataset showed slightly higher AUCs, ranging from 0.88 to 0.96.
• The framework demonstrated consistent and high performance in predicting IDH mutation status.

Conclusions:

• The developed MRI radiomics framework provides a non-invasive and accurate method for preoperative prediction of IDH mutation status in glioma.
• This approach has the potential to significantly impact clinical decision-making for glioma patients.
• Further validation in prospective studies is warranted.