RNA-based next-generation sequencing strategy for screening MDM2-amplified sarcomas

The accurate diagnosis of soft tissue sarcomas and bone tumors poses persistent challenges in pathological practice. While targeted RNA-based next-generation sequencing (NGS) has enhanced diagnostic precision through fusion gene detection, its clinical utility remains limited for dedifferentiated liposarcoma (DDLPS) due to the inability to identify MDM2 copy number variations - a critical diagnostic feature. This limitation is particularly clinically relevant as morphologically atypical DDLPS cases may undergo targeted RNA NGS testing. Through retrospective analysis of 150 patients undergoing RNA NGS and validation with TCGA-SARC data, we developed a novel screening algorithm for DDLPS using RNA NGS data, specifically the number of fusion events within the 12q13-15 chromosomal region and MDM2 transcript levels. Both metrics were significantly elevated in DDLPS compared to non-DDLPS cases in our institutional cohort (p < 0.001). ROC analysis established optimal diagnostic thresholds: ≥3 fusions in 12q13-15 region achieved 100% sensitivity (95% CI: [73.53%-100%]; 12/12) and 95.65% specificity (95% CI: [90.78%-98.39%]; 132/138). Similarly, MDM2 RNA expression ≥ 100 TPM showed 100% sensitivity (95% CI: [63.1%-100%]; 8/8) and 97.37% specificity (95% CI: [90.8%-99.7%]; 74/76). The algorithm also performed robustly in the TCGA-SARC data, although with a slightly different optimal cutoff, likely due to differences in gene panels, depth of sequencing, gene coverage, and bioinformatics pipelines. This study presents an effective strategy for screening DDLPS and potentially other MDM2-amplified sarcomas.