Abstract BACKGROUND: Immune checkpoint inhibition (ICI) continues to be a promising area of research with treatments continuing to show efficacy in late-stage cancers. The detection of disease progression for patients undergoing treatment with this modality remains challenging however given the lack of reliable biomarkers for which to monitor disease. Current methods including imaging can be challenging to evaluate as they do not always capture the timing and nature of clinical responses. Liquid biopsies may offer a solution by allowing for minimally invasive tumor monitoring during therapy. We previously reported on the DELFI Tumor Fraction (DELFI-TF) algorithm, as a tumor- and mutation-naive approach to monitoring using cell-free DNA (cfDNA) fragmentomes (Alipanahi et al., AACR 2023). In this study we examine the performance of DELFI-TF in a cohort of patients with metastatic NSCLC treated with ICI. METHODS: Longitudinal blood samples (n=323) were collected from NSCLC patients (n=109) which received ICI therapy. Whole genome sequencing at low coverage (∼4x) was performed following cfDNA extraction. DELFI-TF scores, predicted using fragmentome features, were used to monitor changes in circulating tumor DNA (ctDNA) levels and associated variations in disease burden. Predictions from DELFI-TF were compared to maximum observed mutation allele frequencies (maxMAF) from a targeted sequencing gene panel for the same samples. In order to determine the prognostic performance of DELFI-TF we evaluated ctDNA changes at baseline and within 3 to 9 weeks of ICI initiation. The presence of undetectable DELFI-TF, defined as below the observed limit of blank (LOB) within 3 to 9 weeks of ICI initiation, indicated a molecular response (mR). In contrast, the persistence of DELFI-TF within this interval signified molecular disease progression (mPD). RESULTS: We observed a strong correlation between DELFI-TF predictions and maxMAF (r=0.93, p<0.001). Further, changes in ctDNA predictions from baseline to the first post-treatment timepoint showed a high correlation between DELFI-TF and maxMAF (r=0.94, p<0.001). Patients with a higher disease burden prior to immunotherapy, as predicted by DELFI-TF and maxMAF, had shorter overall survival (OS) compared to patients with a lower disease burden (DELFI-TF: 11.4 v 33.0 months, p=0.006; maxMAF: 7.82 vs low 33.0 months, log-rank p=0.001). Among 79 patients evaluable for molecular response, 28 (35%) were classified as mPD and 51 (65%) were classified in the mR category. Patients with mR attained longer progression free survival (PFS;16.04 v 2.70 months, log-rank p<0.001) compared to those in the mPD group. Furthermore, assessment of landmark PFS at 6 months revealed a significant association between ctDNA molecular response and durable clinical benefit (DCB; p < 0.001, Fisher’s exact test). CONCLUSIONS: DELFI-TF is a low-cost, tumor and mutation naive approach that accurately monitors longitudinal dynamic changes in tumor burden and captures durable therapeutic responses in patients with metastatic NSCLC receiving immunotherapy. Citation Format: Lavanya Sivapalan, Bahar Alipanahi, Jamie E Medina, Bridget Tripp, Zachary L Skidmore, Paola Ghanem, Gavin Pereira, Nisha Rao, Kavya Velliangiri, Caitlin Schonewolf, Noushin Niknafs, Bryan Chesnick, Jennifer Tom, Stephen Cristiano, Peter Bach, Nicholas C Dracopoli, Robert B Scharpf, Victor Velculescu, Lorenzo Rinaldi, Valsamo Anagnostou. Monitoring response to immunotherapy in lung cancer using cell-free DNA fragmentomes [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr B025.
Read full abstract