Cross-lineage anomaly detection of microRNA expression: a framework for deriving exploratory candidate biomarkers in sarcoma

Introduction Sarcomas are rare, heterogeneous malignancies of mesenchymal origin whose diagnosis and molecular classification remain challenging. We introduced a oneclass anomaly detection framework to microRNA expression profiles to test whether sarcoma-associated miRNA features are distinguishable from those of epithelial, haematological, and neurological cancers and if these features carry any biological relevance. Methods Candidate miRNAs meeting cross-lineage significance criteria were further evaluated through differential expression analysis against subtype-matched normal mesenchymal reference tissue, pathway enrichment, and survival analysis. A five-miRNA index (“Index 5”), derived by Fisher’s linear discriminant analysis with leave-one-out cross-validation in a TCGA-SARC training cohort, was applied without refitting to an independent external cohort (GSE36982). Results The five-miRNA index (”Index 5”) achieved an AUC of 0.93 and, applied without refitting to an independent external cohort (GSE36982), an AUC of 0.97. After Benjamini–Hochberg correction, one candidate miRNA (miR-1305) remained significantly associated with overall survival. Discussion These findings identify a computationally derived, externally tested candidate miRNA signature associated with sarcoma. While this study lacks experimental and clinical validation, the broader aim of this study was to come up with a framework that would assist and speedup the exploratory candidate biomarker discovery process in rare malignancies.