¹¹C-Methionine Integrated PET/MRI-Based Texture Analysis Features May Have a Potential Ability to Distinguish Oligodendroglioma (IDH-Mutant and 1p/19q-Codeleted) From Varied Gliomas

Rationale and objectives: Different histology and gene status of gliomas results in different natural history, treatment, and prognosis in different subgroups. Low-grade gliomas (LGGs) with isocitrate dehydrogenase (IDH) mutant and 1p/19q-codeleted are kind of gliomas with the most favorable outcome, reflecting operational strategy. Less invasive method for prediction of pathological type-even gene status-is desired.

Materials and methods: This study investigates the potential ability of methionine-positron emission tomography (MET-PET) to determine LGGs with IDH-mutant and 1p/19q-codeleted through a retrospective review of information of 70 glioma patients. Patients underwent preoperative MET-PET, followed by operation and histopathological analysis including Immunohistochemistry and polymerase chain reaction analysis for IDH-mutant and fluorescence capillary electrophoresis analysis for 1p/19q codeletion. Texture analysis was performed for further data mining. The t-test and receiver operating characteristic curve analysis were conducted for statistical analysis.

Results: In the whole cohort analysis, SUVmax, SUVmean and texture features (SD and median) of oligodendroglioma, IDH-mutant and 1p/19q-codeleted patients were lower than these values of other patients. In WHO grade II subgroup analysis, no statistical difference of conventional features was observed between groups. Texture analysis displayed higher diffEntropy, diffVariance, and entropy in oligodendroglioma, IDH-mutant and 1p/19q-codeleted patients. Receiver operating characteristic analysis suggested AUCs of some conventional features and texture features ranged from 0.722 to 0.892 that are effective for diagnosis, determining LGGs with IDH-mutant and 1p/19q-codeleted in this cohort and WHO II grade glioma subgroup analysis respectively.

Conclusion: ¹¹C-Methionine integrated PET/MRI based texture analysis and conventional features may be a promising noninvasive predictor for differentiating the varied gliomas.