Abstracts

Rationale: Low-grade epilepsy-associated neuroepithelial tumors (LEAT) present an important group of central nervous system (CNS) neoplasm in children and young adults. We investigated the association of molecular classification with neuroimaging features in LEAT.

Methods: Patients were identified retrospectively from the neuropathology database. We found 93 patients of LEAT. Genetic analyses of the surgical specimens were performed in 66 patients through Multiplex ligation-dependent probe amplification and next-generation sequencing of LEAT-related genes. Nineteen patients with gene panel-negative tumors, two patients without complete data of preoperative MRI, three patients with concomitant hypomyelination on MRI, and one patient with coexisting diffuse white matter disease on MRI were further excluded. Consequently, 41 patients (male: 21, female: 20) were included. The mean age was 13.5 (1-34) years old. DNA methylation analyses were performed in 21 cases. Preoperative neuroimaging studies were evaluated independently by three neuroradiologists. Hierarchical cluster analysis was performed to identify groups of patients with similar neuroimaging characteristics. Histopathological diagnoses were determined by three neuropathologists following the 2021 WHO Classification of CNS tumors. Finally, the association of molecular types with imaging groups was examined.

Results: Genotypes of our study subjects included 27 cases with BRAF V600E mutation, six 6 cases with FGFR1-TKD duplication, two cases with FGFR1 point mutation, and the remaining six cases had other different genotypes, respectively. The imaging features were classified into three groups: Group1, indistinct tumor border and slightly high or high T2-weighted signal intensities without a diffuse mass effect; Group2, sharp tumor border and very high T2-weighted signal intensities with a diffuse mass effect and exophytic growth; Group3, high or slightly high T2-weighted signal intensities with a diffuse mass effect. The neuroimaging features of Group1 tumors were associated with BRAF V600E mutation with 92.6% sensitivity and 100% specificity, and Group2 tumors were associated with FGFR1 mutation with 100% sensitivity and specificity. Group3 tumors were associated with various genetic mutations except for FGFR1. Two cases of BRAF V600E mutant tumors were classified into Group3. The DNA methylation properties of these two tumors clustered differently from the other BRAF V600E mutant tumors by unsupervised clustering. The pathological diagnosis of diffuse low-grade glioma, MAPK pathway-altered (n=23), was not associated with specific imaging groups. Ganglioglioma (n = 8) was 81.3% specific and 20.0% sensitive to Group1. Polymorphous low-grade neuroepithelial tumor of the young (n = 5) was 100% specific and 20.0% sensitive to Group1.

Conclusions: We found genotype-specific neuroimaging features with BRAF V600E or FGFR1 mutations. Our findings will likely contribute to establishing a neuroimaging classification of LEAT having clinical relevance, which needs to be confirmed by future randomized controlled trials.

Funding: This study was supported by a grant from the Japan Agency for Medical Research and Development (no.JP20ek0109374).