Clinical utility of sarcoma methylation classifiers in soft tissue sarcoma diagnosis

Abstract Purpose Soft tissue sarcomas are rare malignant tumors of mesenchymal origin characterized by their heterogeneity and diagnosis is challenging despite integrated histopathology and molecular testing. DNA methylation profiling offers an objective approach for tumor classification, but clinical implementation data in soft tissue sarcoma remain limited. Methods In this single-center study, 40 tumor samples from 34 patients were evaluated by routine pathology and array-based DNA methylation profiling and classified using two versions of the sarcoma methylation classifier (V12.3 and V13.1). Results With version V12.3, 16 samples (40%) achieved a high calibrated score ≥ 0.9, among which 87.5% (14/16) were consistent with the histopathologic diagnosis. With V13.1, 29 samples (72.5%) reached a high calibrated score (≥ 0.9) and the classifier’s diagnosis aligned with the histopathologic diagnosis in 28 cases (96.6%). Conclusion Overall, DNA methylation-based classification is a valuable adjunct diagnostic tool to confirm or refine the histopathologic diagnosis. V13.1 substantially improved classification confidence and agreement compared with V12.3. However, a considerable subset of specimens remained below the high-confidence threshold, underscoring the need for further technical and workflow optimization before routine clinical implementation.