Abstract LB108: Specific plasma DNA end-motifs distinguish patients with lung adenocarcinoma versus benign nodules

Abstract

Abstract Background: Recent studies have identified blood plasma DNA as a rich source of biomarkers for non-invasive cancer detection. In particular, analysis of shorter DNA fragments or “fragmentomics” have shown promising results with numerous competing approaches such as those based on mutation calls, length distributions, or preferred fragmentation patterns. In this study, we assessed the potential of fragment end motif distributions in discriminating patients with lung adenocarcinoma versus benign tumors. Methods: Plasma samples from 37 patients (20 benign, 17 malignant) were subject to whole genome sequencing. For each sample, DNA fragments were grouped as short (30-99bp) and long (100-269bp) and end motif distributions were calculated by counting occurrences of 4-9bp sequences at either the 5’ or 3’ end of each fragment. A proprietary machine learning algorithm was employed to identify a minimal set of motifs capable of accurately distinguishing benign from malignant samples. Results: Just two 5’ end motifs in long fragments, CCAT and GACA, were sufficient for distinguishing malignant from benign samples in a two-step decision tree (89% accuracy with all samples, 86% accuracy in leave-one-out cross validation). Accuracy was significantly higher when compared to random guessing by shuffling labels (p=1.6*10−6, FWER < 0.05, Bonferroni correction, 351 tests). CCAT motif frequency was also significantly associated with cancer stage (p=0.04, Kruskal-Wallis test) and specifically elevated in early stage samples (I, II). Conclusions: These results suggest that plasma DNA end motif distributions are informative in discriminating patients with lung adenocarcinoma versus benign tumors. The two-motif model suggests that specific, rather than systemic changes in end motif distributions are more strongly associated with lung cancer patients and may point to specific mechanisms linked to cancer progression. Citation Format: Jason C. Hyun, Shun H. Yip, Neeti Swarup, Raghuraman Ramamurthy, Jordan C. Cheng, Irene Choi, Edmund Wong, Akanksha Arora, Denise Aberle, David T. Wong, Cheuk Y. Tang. Specific plasma DNA end-motifs distinguish patients with lung adenocarcinoma versus benign nodules [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB108.