Differentiation of Molecular Glioblastomas and Lower-grade Diffuse Astrocytomas Using MRI Perfusion and Diffusion...

Area of Science:

• Neuroradiology
• Neuro-oncology
• Medical imaging

Background:

• Distinguishing molecular glioblastomas from lower-grade diffuse astrocytomas (grades 2-3) is challenging using standard MRI sequences (T1-gadolinium, T2/T2 FLAIR) without T2/T2 FLAIR mismatch.
• Molecular glioblastomas are aggressive brain tumors requiring accurate preoperative characterization.

Purpose of the Study:

• To evaluate the diagnostic performance of apparent diffusion coefficient (ADC) from diffusion-weighted imaging (DWI) and relative cerebral blood volume (rCBV) from dynamic susceptibility contrast (DSC) perfusion imaging.
• To differentiate molecular glioblastomas from lower-grade diffuse astrocytomas using these quantitative MRI parameters.

Main Methods:

• Included 14 patients with molecular glioblastomas (IDH-wildtype with TERT promoter mutation, EGFR amplification, or +7/-10 alterations) and 13 patients with lower-grade diffuse astrocytomas (grades 2-3).
• Measured ADC values in a 7 mm² ROI in the lowest ADC region and normalized DSC-rCBV values in a 10 mm² ROI in the highest perfused region.
• Compared ADC and rCBV values between groups using T-test and Wilcoxon rank-sum test, and analyzed combinations using binary logistic regression, ROC analysis, and Firth regression.

Main Results:

• Molecular glioblastomas showed significantly lower minimum, mean, and maximum ADC values, and higher ADC standard deviation compared to lower-grade astrocytomas (p < 0.05).
• Molecular glioblastomas exhibited significantly higher normalized median, average, and minimum rCBV ratios than lower-grade astrocytomas (p < 0.01).
• A combined ROC model using ADC and rCBV achieved an area under the curve (AUC) of 0.93 (95% CI: 0.82-1.00).

Conclusions:

• ADC and rCBV measurements effectively differentiate molecular glioblastomas from lower-grade diffuse astrocytomas lacking T2/T2 FLAIR mismatch.
• These quantitative imaging biomarkers can assist in preoperative tumor characterization, surgical planning, and prognosis assessment.
• This approach enhances diagnostic accuracy for challenging brain tumor cases.