Table 2.

Comparison of methods for MB classification in the clinical laboratory

Method	Description	Pros	Cons	References
Immunohistochemistry	Description: Classification based on pattern of expression of three proteins as detected by immunohistochemistry (YAP1, GAB1, and beta‐catenin)	Easy implementation Low capital expense Technology available in most clinical labs Protein is a relatively stable macromolecule Works in samples with low tumor content Works well with ffpe	Interpretation challenges in low‐volume laboratories Divergent differentiation yields indeterminant class Cannot resolve g3 and g4 tumors Cannot account for increasingly granular mb classification Limited to mb classification	(17)
	Classification: Tumors separable into WNT, SHH, Non‐WNT/non‐SHH
Transcription profiling	Description: Classification based on shared RNA expression signature as detected by transcription array, RNA sequencing, or nanostring TM	Can separate tumors into all four canonical molecular groups Works with ffpe or frozen material Moderate to high capital expense	Based on RNA, a relatively unstable macromolecule Technology not available in many clinical laboratories No supervised classification model currently available Based on RNA, a relatively unstable macromolecule Limited to MB classification More granular MB classification not currently available Requires relatively pure tumor	(12, 40, 54, 83)
	Classification: Tumors separated into WNT, SHH, G3, and G4
Methylation	Description: Classification based on methylation signature using unsupervised or supervised classification models. Typically utilize Illumina Human Infinium 450K/EPIC arrays	Can separate all four canonical methylation classes Scales well to other tumor types Large reference series available for mb and other tumor types Works with ffpe and frozen tissue Utilizes dna, relatively stable macromolecule	Not currently widely available High capital expense Not all clinical samples contain sufficient dna quantity for classification Requires high proportion of pure tumor	(10, 32)
	Classification: Tumors separated into WNT, SHH, G3, and G4 More granular class structure can be resolved
DNA Sequencing	Description: Classification based on sequencing of class specific recurrent mutations	Can be performed in most modern molecular laboratories Many commercial labs evaluate the genes recurrently mutated in mb Works on ffpe or frozen material Utilizes dna, relatively stable macromolecule	Many MB do not have contain defining mutations (ie. G3/G4) Some mutations cross class boundaries (ie. Subclonal SHH pathway mutations in WNT tumors) High capital expense	(1, 51, 68)
	Classification: Reliable for WNT class in most instances, but cannot resolve all four canonical classes