# Advanced machine learning to empower ultra-sensitive liquid biopsy in melanoma and non-small cell lung cancer > **NIH NIH K08** · SLOAN-KETTERING INST CAN RESEARCH · 2024 · $292,597 ## Abstract Research: The ability to monitor malignant tumor burden below the limit of radiographic detection remains a major unmet need. Liquid biopsy for circulating tumor DNA (ctDNA) offers promise, however, deep targeted sequencing methods – the conventional approach in the field – face a sensitivity plateau in low volume cancer due to the sparsity of ctDNA signal. Whole genome sequencing (WGS) of plasma overcomes this sensitivity barrier by expanding the number of informative sites to the thousands of somatic single nucleotide variants observed across the genome in solid tumors. We showed with our tumor-informed MRDetect framework that WGS of plasma can increase liquid biopsy sensitivity by 1-2 orders of magnitude beyond deep targeted sequencing methods. To expand applicability and overcome MRDetect’s need for matched tumor tissue, I built MRD-EDGE, a plasma-only (de novo) classifier that uses advanced machine learning to increase error suppression and amplify ctDNA signal. My preliminary data shows that MRD-EDGE can quantify ctDNA tumor burden during the nadir of response to immunotherapy in patients with advanced melanoma, and can demonstrate a response to treatment as early as 3 weeks after first infusion. MRD-EDGE therefore enables precise monitoring of malignant disease burden in response to therapy using standard WGS alone. In this proposal, I aim to first radically improve MRD-EDGE sensitivity by including epigenetic features that inform likelihood of cancer mutagenesis, which I hypothesize will allow for unprecedented plasma-only liquid biopsy sensitivity. I will then use the optimized MRD-EDGE platform to define early response or resistance to immunotherapy in metastatic melanoma, which will establish ctDNA as a biomarker that can complement or replace imaging. Finally, I will optimize MRD-EDGE for use in lung cancer and use the platform to monitor response to neoadjuvant immunotherapy and detect postoperative minimal residual disease. I expect that ultra-sensitive monitoring of ctDNA dynamics in the neoadjuvant period can guide precision adjuvant therapy in the postoperative period and thereby provide a transformative impact on patient care. Candidate: I am an Instructor of Medicine at Memorial Sloan Kettering Cancer Center (MSK) and a Visiting Fellow at the New York Genome Center (NYGC). I have outlined a 5-year career development plan to transition to an independent, tenure-track physician-scientist investigating the detection and monitoring of solid tumors through ultrasensitive liquid biopsy. I will conduct the proposed research under the mentorship of Dr. Dan Landau, an internationally recognized expert in liquid biopsy and cancer genomics. I will use the K08 award to further develop skills in next-generation sequencing methods and analysis and advanced machine learning. MSK and the NYGC are ideal environments in which to pursue my scientific and career goals. Both institutions have world class research communities and an outstanding trac... ## Key facts - **NIH application ID:** 10881697 - **Project number:** 5K08CA263301-02 - **Recipient organization:** SLOAN-KETTERING INST CAN RESEARCH - **Principal Investigator:** Adam Widman - **Activity code:** K08 (R01, R21, SBIR, etc.) - **Funding institute:** NIH - **Fiscal year:** 2024 - **Award amount:** $292,597 - **Award type:** 5 - **Project period:** 2023-07-04 → 2028-06-30 ## Primary source NIH RePORTER: https://reporter.nih.gov/project-details/10881697 ## Citation > US National Institutes of Health, RePORTER application 10881697, Advanced machine learning to empower ultra-sensitive liquid biopsy in melanoma and non-small cell lung cancer (5K08CA263301-02). Retrieved via AI Analytics 2026-05-26 from https://api.ai-analytics.org/grant/nih/10881697. Licensed CC0. --- *[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*