Abstract 2334: Computational methods for optimizing marker selection, clonal lineage reconstruction, and longitudinal tracking of clonal dynamics via circulating tumor DNA (ctDNA)

Xuecong Fu,Yueqian Deng,Zhicheng Luo,William Laframboise,David Bartlett,Russell Schwartz
DOI: https://doi.org/10.1158/1538-7445.am2024-2334
IF: 11.2
2024-03-24
Cancer Research
Abstract:Background: Technologies for blood-based profiling of tumor DNA ("liquid biopsy") have offered great prospects for non-invasive early cancer diagnosis, treatment monitoring, and clinical guidance, but further advances in computational models and data analysis to optimize monitoring protocols and develop liquid biopsy as a robust quantitative assay of tumor clonal evolution. Methods: We propose new computational methods to improve our ability to characterize tumor clonal dynamics from circulating tumor DNA (ctDNA). Our methods address how to apply ctDNA assays most effectively to two distinct questions in profiling tumor clonal dynamics: 1) How to apply longitudinal ctDNA data to refine phylogenetic tree models of clonal evolution, and 2) how to most effectively characterize changes in clonal frequencies that may be indicative of treatment response or tumor progression. We pose these questions computationally in terms of a probabilistic framework for optimally identifying maximum likelihood markers for the preceding tasks and applying measured marker concentrations to their solution. We apply this framework by first evaluating the distribution of plausible clonal lineage models using bootstrap samples over pre-treatment tissue-based genetic variation data, then refining these lineage models and the clonal frequencies we can infer from them over successive longitudinal samples of selected markers. Results: We tested our method on synthetic data under various model assumptions and showed the method to be effective at refining distribution tree models and clonal frequencies over longitudinal samples so as to minimize measures of tree distance relative to the ground truth. We further apply the methods to a real ovarian cancer case and show their ability to develop and refine a clonal lineage model and assess clonal frequencies. Conclusion: Our methods show the power of ctDNA assays in conjunction with computational optimization to lead to improved marker selection, clonal lineage reconstruction, and tracking of clonal dynamics, with potential for more precise and quantitative profiling of tumor progression enabling improved clinical decision-making. Citation Format: Xuecong Fu, Yueqian Deng, Zhicheng Luo, William Laframboise, David Bartlett, Russell Schwartz. Computational methods for optimizing marker selection, clonal lineage reconstruction, and longitudinal tracking of clonal dynamics via circulating tumor DNA (ctDNA) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular s); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl) nr 2334.
oncology
What problem does this paper attempt to address?