BACKGROUND: Myxoid liposarcomas (MLS) can exhibit a disseminated metastatic pattern, necessitating extensive diagnostics during follow-up. With no tumor markers available, early diagnosis of recurrences and tumor monitoring is difficult. The detection of circulating tumor DNA (ctDNA
liquid biopsy) in MLS with the characteristic translocations t(12
16) and t(12
22) can provide an additional diagnostic. However, due to the often very low tumor fraction, distinguishing actual tumor variants from sequencing artifacts remains a key challenge. METHODS: Using MLS as a model for translocation-driven tumors, this study evaluates a refined analytical approach for detecting both single nucleotide variants (SNVs) and structural variants (SVs) with the highest possible sensitivity and specificity. Different analysis pipelines using Unique Molecular Identifiers (UMIs) were compared in dilution series of tumor DNA from MLS patients (n = 11) and a cell line. The results were validated on plasma samples (n = 36) from two MLS patients and one patient with Synovial Sarcoma (SS). RESULTS: In dilution series, the use of UMIs significantly reduced false positive events in SNV analysis while maintaining high sensitivity without significant differences. In SV analysis, the effect of UMIs was not consistently detectable, as some dilution series exhibited no false positive events even without UMI correction. Additional filter criteria further improved specificity without significantly compromising assay sensitivity. Validation on patient plasma samples confirmed these findings, demonstrating the advantages of the differentiated analytical approach. CONCLUSION: By integrating a refined analytical approach for SNVs and SVs, we achieved reliable ctDNA detection that corresponded with the clinical course of the patients' disease. This method enables non-invasive tumor detection in translocation-driven tumors with low mutational burden and can easily be adapted into routine clinical diagnostics.