Prediction of Prognosis in Glioblastoma Using Radiomics Features of Dynamic Contrast-Enhanced MRI

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Glioblastoma is an aggressive brain tumor with poor prognosis.
• Accurate prognostic markers are crucial for guiding treatment decisions.
• Dynamic contrast-enhanced (DCE) MRI offers quantitative imaging biomarkers.

Purpose of the Study:

• To develop and validate a radiomics risk score using DCE MRI for predicting prognosis in glioblastoma patients.
• To assess the score's ability to stratify patients into high- and low-risk groups.
• To determine if the radiomics risk score is an independent predictor of progression-free survival.

Main Methods:

• Extracted 642 radiomic features from DCE MRI (Ktrans, Vp, Ve) in 150 glioblastoma patients.
• Selected salient features from non-enhancing T2 hyperintense areas using feature selection algorithms.
• Developed a radiomics risk score in a discovery set (n=105) and validated it in a validation set (n=45).
• Performed multivariable Cox regression analysis for progression-free survival.

Main Results:

• 16 radiomic features from non-enhancing areas were selected.
• The radiomics risk score effectively stratified patients into high- and low-risk groups (p < 0.001 in both sets).
• The score showed independent associations with progression-free survival, along with IDH mutation status.

Conclusions:

• A validated radiomics risk score derived from DCE MRI non-enhancing areas can stratify glioblastoma patients.
• This score is an independent prognostic factor for progression-free survival, irrespective of IDH mutation status.