Abstract
Non-invasive sampling of tumor-derived genetic material in circulation through liquid biopsy may be very beneficial for an accurate diagnosis and evaluation of response to treatment in patients with malignant and benign soft tissue tumors. We previously showed that tumor-derived genomic aberrations can be detected in plasma of patients with leiomyosarcoma (LMS) and leiomyoma (LM). In LMS patients, we also showed that the levels of circulating tumor DNA (ctDNA) correspond with response to treatment. We developed an approach tailored to genomic profile of LMS (characterized by intermediate levels of point mutations and copy number alterations, CNAs). Based on TCGA data, we designed a panel of 89 most frequently mutated genes in LMS, which we profiled in plasma DNA by deep sequencing. In parallel, plasma samples were analyzed by shallow whole genome sequencing for detection of CNAs. With this approach, we detected ctDNA in 71% (20/28) of samples from 6/7 patients with advanced disease with >98% specificity. The combination approach for orthogonal profiling of point mutations and CNAs proved to increase the sensitivity of ctDNA detection. Currently, we seek to further improve the sensitivity of ctDNA detection by refining our capture panel and tracking LMS-specific DNA methylation markers in circulation, in addition to point mutations and CNAs. The ultimate goals of our ctDNA studies are 1) to develop a highly sensitive assay for evaluation of response to therapy and long-term surveillance for patients with LMS, and 2) to develop a blood-based test for accurate pre-operative distinction between LMS and LM.
To identify LMS-specific DNA methylation markers, we analyzed a test cohort of 76 LM, 35 uterine LMS and 31 extra-uterine LMS by Illumina Infinium EPIC arrays. We identified differentially methylated CpGs between LM and uterine LMS, and between LM and all LMS using a newly developed custom pipeline in R. The results of this analysis are currently being validated in a new dataset of 41 LM and 153 LMS generated by our group. Recently published (PMID: 34301934) genomic data from new 53 LMS samples are used to refine the panel of the most frequently mutated genes that we identified previously in the LMS TCGA data.
Our preliminary analysis of test cohort revealed >270 differentially methylated CpGs between LM and uterine LMS, and >1000 differentially methylated CpGs between LM and all LMS. The preliminary analysis of genomic data shows that the initial panel of 89 frequently mutated genes could be substantially narrowed down to cover only selected tumor suppressor genes. Once validated, these results will be used to refine the ctDNA assay for LMS and LM.
Our results point to multiple epigenetic markers that could be used for ctDNA profiling, in addition to point mutations or CNAs. Further validation will allow us to select the most reliable LMS- and LM-specific DNA methylation markers and the most frequently mutated regions across independent datasets, and these markers will be incorporated into our new ctDNA test for a concurrent detection of point mutations, CNAs and DNA methylation markers in circulation.
