Abstract
Background and purpose: The World Health Organization (WHO) subdivided medulloblastoma into genetic and histopathological groups, each with a specific therapeutic intervention and different clinical outcomes. These subtypes may present with distinct imaging features. Therefore, the current study aimed to identify magnetic resonance imaging (MRI) biomarkers to predict the precise pathological characteristics of medulloblastoma. Methods: This study included 28 patients with a first diagnosis of medulloblastoma who underwent preoperative brain MRI with subsequent surgical resection and histopathological confirmation at our hospital between 2010 and 2022. Conventional MRI parameters, including apparent diffusion coefficient (ADC) mean values, were correlated with molecular subtypes to identify distinct MRI biomarkers. Results: Out of 28 tumors, two (7.1%) tumors exhibited wingless (WNT) activation, thirteen (46.4%) exhibited sonic hedgehog (SHH) activation, and thirteen (46.4%) exhibited non-WNT/non-SHH activation (Group 3 or 4). Statistical analysis revealed a significant association of SHH-activated tumors with paramidline/cerebellar location and the presence of peritumoral edema (p value = <0.0001). No significant correlations were found between the genetic subtypes and the other MRI parameters. A distinctive distribution of the ADC-mean values among the various genetic subtypes with recognizable tendencies was identified. However, it was statistically insignificant. Conclusion: Conventional MRI features of the paramidline/hemispheric location and the presence of peritumoral edema were significantly correlated with the SHH activated pathway and hence can be used to facilitate the preoperative implementation of SHH-targeted therapeutic intervention. Although the ADC-mean measurements were not statistically significant, a recognizable distribution of values among the various genetic subtypes was identified.