A next-generation sequencing methylation assay for sub-typing small cell lung cancer in cell-free DNA.

Abstract

e15021 Background: Liquid biopsy provides an easy access and minimally invasive technique to inform cancer patient management in clinics as well as clinical trials. Detecting genomic alterations in circulating tumor DNA (ctDNA) has demonstrated clinical utility and actionability in therapy selection and longitudinal monitoring. The development of a comprehensive assay that provides multi-omic information is imperative to expand the utility of liquid biopsy. Methods: Agilent Resolution ctDNA diagnostic assays can detect all four major types of actionable genomic alterations for cancer patients: SNVs, Indels, fusions, and CNVs. The Agilent Resolution ctDx FIRST assay has recently been approved by the FDA as a companion diagnostic assay for non-small cell lung cancer (NSCLC). Applying the same core technology, we developed an additional assay that enables methylation profiling in cfDNA. The custom methylation-specific panel targets informative regions to detect tumor signals and potentially profile the disease by subtype. Recent reports suggest that small cell lung cancer (SCLC) may be stratified into subtypes based on gene expression, including that of ASCL-1, NEUROD1, and POU2F3. We used the proposed SCLC subtyping model to test the feasibility of subtyping SCLC by cfDNA methylation profiles. A baseline methylation profile was established using cell-free DNA from healthy donors. Test samples included genomic DNA from SCLC cell lines and plasma samples of SCLC patients with matching tissue, where available. Results: We applied a methylation-specific bioinformatic pipeline that generates error-corrected unique molecules. The methylation workflow showed unique coverage comparable to our ctDx assay and produced consistent results from input DNA as low as 1ng. We found consistent methylation patterns in cfDNA among healthy donors. SCLC cell lines showed distinct profiles that could be classified into sub-types. When spiked into healthy donor cfDNA background, the tumor signals were distinguishable down to ~0.5%. Plasma samples from SCLC patients showed profiles consistent with that of SCLC sub-types defined by cell lines. Matching plasma and tissue samples produced concordant sub-typing results, supporting the feasibility of disease subtyping in both sample types. Samples were also tested by a custom SCLC panel to analyze genomic alterations. Prevalent variants in SCLC were detected, including SNVs in TP53, PIK3CA, RB1, and MYC amplifications. Conclusions: We showed a cfDNA methylation profiling platform that could support disease sub-typing for SCLC. This platform has great potential in disease stratification and patient management.