Abstract PR19: Monitoring of melanoma progression utilizing multiplatform biomarkers of blood cell-free DNA

Abstract

Abstract Blood biopsies utilizing cell-free DNA (cfDNA) represent an alternative, minimally invasive source that permits repeated sampling to enable melanoma patient monitoring during treatment and follow-up. However, the sensitivity of cfDNA biomarker detection is limited to tumor sampling availability, which can vary significantly. This encourages the use of multiplatform biomarkers to improve sensitivity in assessing melanoma progression. Here we explored the detection of different cfDNA biomarker types, DNA mutations and copy number amplification (CNA), in blood of melanoma patients. To assess the utility of these cfDNA biomarker types, we utilized a highly sensitive multigene (54-gene cancer panel) digital next-generation sequencing (NGS) assay containing 0.1% sensitivity or multiplex droplet digital PCR ddPCR. cfDNA mutations were assessed by a multigene digital NGS assay in a cohort of preoperative stage III/IV patients rendered disease-free (n=44), whereby 75% of patients had detectable cfDNA genomic aberrations. Patients with high tumor mutation burden (TMB) (≥3 mutations) were associated with worse overall survival (OS) (p=0.048) compared to those with <3 mutations. Next, we assessed 98 serial bleeds obtained from a cohort of preoperative stage IIIA-C melanoma patients rendered disease-free followed by adjuvant immunotherapy with follow-up until stage IV distant metastasis (n=12), where cfDNA mutation burden (CMB) levels correlated with TMB (p=0.019). This enabled earlier detection of stage IV disease recurrence compared to imaging (p<0.01), and captured tumor heterogeneity, and rising CMB levels prerecurrence clinical detection. An expanded 70-gene digital NGS panel was used to assess melanoma brain metastasis (MBM) in 26 preoperative MBM patients who had paired resected MBM tumors-cfDNA. Interestingly, tumor-identified MBM mutations were detected in the paired cfDNA at a concordance of 48%, 54%, and 58% at mutant allele frequency cut-offs of 0%, 0.5%, and 1%, respectively. Encouragingly, cfDNA-tumor SNV concordance was 88% in known melanoma driver genes (BRAF/NRAS). We then assessed a novel 1q21.3 CNA during melanoma progression in patients receiving checkpoint inhibitor immunotherapy (CII). Using a multiplex ddPCR assay targeting four 1q21.3 gene targets, we identified 77% (n=31) of patients at the time of progressive disease while on CII had 1q21.3 CNA in plasma whereby patients with stable disease, partial response, or disease free had low or no 1q21.3 CNA cfDNA. These results demonstrate the ability to use different multiplatform cfDNA biomarkers to monitor melanoma disease progression for recurrence and treatment response. The study demonstrates that cfDNA as mutation detection can be improved in melanoma using other genomic aberrations in monitoring immunotherapy in melanoma patients. Citation Format: David S. Hoon, Selena Y. Lin, Rebecca Gentry, Steve O’Day. Monitoring of melanoma progression utilizing multiplatform biomarkers of blood cell-free DNA [abstract]. In: Proceedings of the AACR Special Conference on Melanoma: From Biology to Target; 2019 Jan 15-18; Houston, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(19 Suppl):Abstract nr PR19.