Abstract 6699: What are the characteristics of mutations to use as targets in tumor-informed ctDNA analysis

Introduction: Tumor-informed, mutation-based approaches for the detection of circulating tumor DNA (ctDNA) are among the most common and widely used ctDNA detection approaches. To identify the mutations to be used as ctDNA targets, tumor tissue is often mutation-profiled using broad panel sequencing strategies. Optimally, the strategy for mutation profiling should enable assessment of clonality and copy number of the identified mutations. Hereby, it will be possible to rank the mutations and select those that are clonal and present in multiple copies per cancer cell. Theoretically, this should increase the ctDNA detection likelihood. However, there is no consensus on how best to identify and select the targets. Here, we propose a bioinformatic pipeline for an easy and robust selection of target mutations for ctDNA detection from tumor whole-exome sequencing (WES) profiles. Methods: The pipeline consists of Mutect2 and PureCN, which were both applied to paired tumor and normal WES profiles. The output from the pipeline is a list of somatic mutations, each with estimated measures of clonality, in the form of a cancer cell fraction (CCF), and multiplicity (local mutation copy number). Mutations with CCF above 0.9 were considered clonal. The performance of the pipeline was assessed using matched WES and ctDNA data from 98 colorectal cancer (CRC) and 30 bladder cancer patients. For CRC patients, 1-5 mutations per patient within a pre-defined CRC panel were used as ctDNA targets, while an average of 41 selected mutations per patient were used for bladder cancer patients. The performance measure was the mutation detection rate in pre-operative plasma samples, stratified for clonality (CCF), multiplicity, and the combination of CCF and multiplicity. Results: For CRC patients, 51% of mutations with CCF > 0.9 were detected compared to only 17% when CCF < 0.9 (p= 4.3e-09). In bladder cancer patients 21% of mutations with CCF > 0.9 were detected compared to 2 % when CCF < 0.9 (p=1.2e-09). For CRC patients the ctDNA detection rate was 58% when mutation multiplicity was higher than 1 and 31% when multiplicity was 1 (p= 9.175e-06). For bladder cancer, the same numbers were 22% and 16% (p=4.635e-03). Mutations with CCF > 0.9 and multiplicity >1 had significantly increased ctDNA detection rate compared with mutations with CCF < 0.9 and multiplicity 1, in both CRC (p= 1.486e-09) and bladder cancer (p=1.077e-09). Conclusion: The CCF and multiplicity measures provided by the pipeline are highly useful when selecting ctDNA targets. By selecting mutations with CCF> 0.9 and multiplicity >1 the likelihood of detecting ctDNA in plasma cfDNA is significantly improved. Citation Format: Marijana Nesic, Mads H. Rasmussen, Tenna V. Henriksen, Christina Demuth, Amanda Frydendahl, Iver Nordentoft, Lars Dyrskjøt, Claus L. Andersen. What are the characteristics of mutations to use as targets in tumor-informed ctDNA analysis [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 6699.