Methylation Profiles in Bone and Soft Tissue Tumors: Do They Help Classify the Unclassifiable?

Abstract

Methylation profiling offers a promising avenue for improving the diagnosis and classification of bone and soft tissue tumors, particularly in cases where traditional methods fall short. Through examining tissue- and lineage-specific DNA methylation patterns, this approach can augment the classification of morphologically similar tumors with different genetics (genetic heterogeneity) or tumors that share genetic drivers but diverge phenotypically (phenotypic heterogeneity). This tool can also help clarify previously unclassifiable, poorly differentiated or transdifferentiated nonmesenchymal tumors that mimic sarcomas. By adopting an unbiased approach to grouping and subtyping using unsupervised clustering on methylomes, methylation profiling could potentially revise how we classify sarcomas. There is a need clinically to develop a methylation classifier that accurately predicts sarcoma class based on methylation profile, specifically in sarcomas that are challenging to differentiate due to overlapping morphologic or molecular features. However, current classifiers are limited by the diversity and size of their reference cohorts, dilution of signal by nonneoplastic tissue, and alteration of the methylomic landscape by histone modifications and oncometabolite-related mutations. Key considerations include the importance of quality control for tissue-based methylation profiling, transparent and reproducible pipelines, and open data sharing. Future advancements include continued refinement of methylation-based classifiers with a broader spectrum of rare and ultrarare tumor types and adopting new biological insights from methylome-driven analyses and integration of multiomic approaches. Although methylation profiling emerges as a valuable adjunct to existing diagnostic modalities, clinicopathologic considerations should always be incorporated for a more nuanced, management-based tumor classification system.