Abstract 1036: Detection of circulating tumor DNA by whole genome sequencing enables prediction of recurrence in stage III colorectal cancer patients with great inter-lab reproducibility

Abstract Background: Detection of circulating tumor DNA (ctDNA) in plasma indicates the presence of cancer. In patients with very few copies of tumor DNA circulating in the blood, the blood sampling process becomes a potentially limiting factor. For ctDNA methods based on a single or a few genomic targets, it becomes stochastic whether the collected sample contains any tumor DNA fragments from targeted regions. To overcome this sampling limitation, we developed C2inform; a whole genome sequencing (WGS) approach, which detects ctDNA using a cumulative patient-specific signal from thousands of mutations throughout the entire genome. Aim: Here, we aim to I) evaluate the performance of C2inform in patients with UICC stage III colorectal cancer (CRC) using samples collected after the end of curatively intended treatment and serially during surveillance, II) assess the inter-lab reproducibility of C2inform and, III) investigate the potential for using the serial WGS data to track tumor genomic evolution in recurrence patients. Methods: From a cohort of 146 stage III CRC patients, including 37 patients with recurrence, 2 mL plasma samples were collected serially for up to three years (n = 1309, median 10 samples per patient). By WGS of tumor and blood-derived normal DNA, a mutational signature was established for each patient. Enhanced by an AI-based error suppression model, this signature was used to screen 20x WGS plasma cfDNA profiles for the presence of ctDNA. To evaluate the reproducibility, paired samples (n = 2 x 187 samples) were processed and sequenced at two independent laboratories while the bioinformatics processing of samples was identical. Using coverage, split-read, and discordant read-pair information, genomic structural alterations were identified in serial plasma samples from 19 recurrence patients and compared to tissue WGS from subsequent recurrence metastasis. Results: Preliminary results showed that detection of ctDNA after the end of treatment was associated with poor prognosis (HR = 7.1, 95% CI: 3.2-15.9). The presence of ctDNA during surveillance was a predictor of recurrence (HR = 9.5, 95% CI: 3.6-25.5) and enabled the detection of recurrence up to 21 months (median: 9.5 months) before detection by radiological imaging. We found indication of tumor evolution by searching serially collected plasma samples for novel genomic changes, which were confirmed by WGS of the metastatic tissue. Analysis of paired samples showed great reproducibility of C2inform with a high agreement between both ctDNA detection (Cohens Kappa = 0.9) and the estimated ctDNA level (r² = 0.99). Conclusion: C2inform showed great inter-lab reproducibility and enabled prediction of recurrence after treatment and during surveillance. Thus, C2inform has the potential to guide clinical decision-making during the postoperative management of CRC patients. Citation Format: Amanda Frydendahl, Jesper Nors, Mads Heilskov, Ester Ellegaard Sørensen, Thomas Reinert, Jesper Bramsen, Danielle Afterman, Tomer Lauterman, Maja Kuzman, Santiago Gonzalez, Dunja Glavas, James Samdbeck, Dillon Maloney, Jurica Levatic, Sunil Deochand, Michael Yahalom, Ryan Ptashkin, Iman Tavassoly, Zohar Donenhirsh, Eric White, Ravi Kandasamy, Ury Alon, Paz Polak, Anders Husted Madsen, Uffe Schou Løve, Per Vadgaard Andersen, Ole Thorlacius-Ussing, Lene Hjerrild Iversen, Kåre Andersson Gotschalck, Boris Oklander, Asaf Zviran, Claus Lindbjerg Andersen. Detection of circulating tumor DNA by whole genome sequencing enables prediction of recurrence in stage III colorectal cancer patients with great inter-lab reproducibility [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1036.