Extensive genetic analysis of WHO grade 2 and 3 gliomas has recently been conducted ^1,2. These investigations revealed that WHO grade 2 and 3 comprise of several disease subtypes with different genetic or molecular backgrounds and a novel concept of "lower grade glioma" has been introduced ¹. Among several truncal mutations, IDH1/2 and TERT promoter mutations were identified to stratify 4 lower grade glioma subtypes ^1,2. Both IDH and TERT promoter muted tumors are considered as classical oligodendroglial tumors while IDH mutated but TERT promoter wild type tumors as astrocytic. The present investigation was conducted to elucidate the differences revealed on MR images including textures and locations of the tumors according to genetic mutation status of lower grade gliomas.

Patient selection and genetic analysis for lower grade glioma

43 lower grade gliomas (19 WHO grade and 24 grade 3) were retrospectively collected. Genetic analysis was conducted in 41 cases with IDH1/2 and TERT promotor mutation status being determined. Patients were then grouped according to IDH mutation +/- and TERT promoter mutation +/- status.

MR based location and texture analysis

T2-high intensity lesions were extracted in 3-dimension for all 43 cases. Entropy of the extracted T2-high intensity lesion was calculated using the following equation.

$$H(A)=-\sum_{0}^{255}plogp$$

Entropy is considered to represent the magnitude of “heterogeneity” within the lesion. Therefore, the calculated values were compared to a binary visual assessment of image, that is heterogenous or homogenous image, by a board certified diagnostic radiologist. Entropy was then compared between genetic mutation status. Subsequently, T2-weighted images were registered to a 1.0-mm isotropic, high-resolution T1-weighted brain atlas provided by the Montreal Neurological Institute (MNI152 database) using a mutual information algorithm with a 12-degree of freedom transformation using FSL-FLIRT ³. The segmented T2-high intensity lesions were then registered to and mapped on MNI152 to visualize tumor location according to genetic mutation status.

Genetic mutation status of the cohort

There were 11 IDH mt / TERT wt, 14 IDH mt / TERT mt, 14 IDH wt / TERT wt, and 2 IDH wt / TERT mt.

T2-entropy and tumor heterogeneity

T2-entropy was statistically significantly higher in tumors that were grouped as “heterogenous” compared to those grouped “homogenous” (p = 0.02 and Figure 1). ROC analysis for using T2-entropy as surrogate for tumor heterogeneity was also statistically significant (p = 0.04, AUC 0.7 and Figure 2).

T2-entropy and tumor grade / genetic mutation status

While T2-entropy did not differ between WHO grades (p = 0.11 and Figure 3), T2-entropy was statistically significantly different between genetic mutation status (p = 0.003, one-way ANOVA and Figure 4). Post-hoc analysis revealed that IDH mt / TERT mt tumors showed higher T2-entropy compared to IDH mt / TERT wt tumors (p < 0.05, Tukey’s test and Figure 4).

Tumor location according to genetic mutation status

When tumor location was compared between IDH mt / TERT mt and IDH mt / TERT wt tumors, there was a clear difference between these two groups with IDH mt / TERT mt type tumors arising at mid-based frontal lobe (Figure 5).

This investigation successfully revealed that lower grade gliomas present different MR imaging features depending on their genetic mutation status. T2-entropy, a newly introduced image texture metric was capable of discriminating heterogenous versus homogenous textures of the tumor on T2-weighted images. Using this novel metric combined with genetic mutation status of the tumor, we were able to reveal that tumor heterogeneity represented on T2-weighted images are different depending on their genetic status. More specifically, both IDH and TERT promoter mutated tumors (classic oligodendroglial tumor) present more heterogeneously on T2-weighted images than IDH muted but TERT promoter wild type tumors (classic astrocytic tumor). Another significant finding was that different genetic mutation backgrounds lead to different locations of the tumor to arise within the brain. Classic oligodendroglial tumors located at the mid-base frontal lobe while astrocytic tumors occupied much lateral side of the brain. These observations suggest that these two biologically different lower grade gliomas indeed arise from different locations of the brain, implying that type of tumor initiating cells could be different. As this was a successful investigation in terms of proof-of-concept for combing genetic mutation status and MR imaging features, a more extensive analysis using a larger cohort is desired.The current investigation was able to cross correlate genetic mutation status and MR imaging features focusing on image texture and tumor location of lower grade gliomas revealing that differences in genetic backgrounds lead to different presentations of the tumor on MRI.