Abstract
Medulloblastoma is an embryonal tumor of the central nervous system that is divided into 3 molecular subgroups: WNT-activated, SHH-activated, and non-WNT/non-SHH-activated. The latter comprises group 3 and group 4. Given significant prognostic differences, distinguishing molecular subgroups is important. Although a panel of immunohistochemical markers can be used for grouping, verification of the results with a molecular technique is highly recommended. The aim of this study is to evaluate the performance of immunohistochemistry (IHC)-based molecular grouping with DNA methylation analysis data. The study included 44 cases overall with a histologic diagnosis of medulloblastoma and with optimal-quality extracted tumor DNA. The tumors were grouped into 3 molecular subgroups using IHC for beta-catenin, YAP1, GAB1, and p53. The extracted DNA samples were analyzed using methylation profiling to perform DNA methylation-based molecular grouping. Of 44 cases, methylation profiles matched the defined profiles of medulloblastoma molecular subgroups in 31 cases. Four cases (12.9%) showed a WNT-activated profile, 8 (25.8%) showed an SHH-activated profile, 12 (38.7%) were group 3, and 6 (19.3%) were group 4. In 1 case, differentiation between group 3 and group 4 could not be made. In comparison with IHC-based molecular classification, a discrepancy was found in 7 cases (22.5%), which were initially classified as SHH-activated by IHC but were revealed as WNT-activated, group 3, or group 4 by methylation analysis. We conclude that IHC-based molecular classification of medulloblastomas may yield results that differ from those obtained through methylation profiling-based classification. This discrepancy arises from the lack of a universal cutoff for nuclear beta-catenin positivity and can occur in the SHH-activated group with nondesmoplastic histology, showing focal, weak GAB1 or focal YAP1 immunopositivity. Validated IHC interpretation protocols may help ensure greater accuracy across these 2 techniques.