No abstract

Motivation Cell-free nucleic acid (cfNA) sequencing data require improvements to existing fusion detection methods along multiple axes: high depth of sequencing, low allele fractions, short fragment lengths and specialized barcodes, such as unique molecular identifiers. Results AF4 was developed to address these challenges. It uses a novel alignment-free kmer-based method to detect candidate fusion fragments with high sensitivity and orders of magnitude faster than existing tools. Candidate fragments are then filtered using a max-cover criterion that significantly reduces spurious matches while retaining authentic fusion fragments. This efficient first stage reduces the data sufficiently that commonly used criteria can process the remaining information, or sophisticated filtering policies that may not scale to the raw reads can be used. AF4 provides both targeted and de novo fusion detection modes. We demonstrate both modes in benchmark simulated and real RNA-seq data as well as clinical and cell-line cfNA data. Availability and implementation AF4 is open sourced, licensed under Apache License 2.0, and is available at: https://github.com/grailbio/bio/tree/master/fusion.

Introduction: Cancer detection blood tests have shown clinical utility after cancer diagnosis, but many evaluate only single cancers and require tumor tissue, thereby limiting their utility. We developed a versatile, tissue-free (ie, no tumor tissue required), multi-cancer detection test (“Post-Diagnosis Cancer Research Solution”) based on methylation sequencing of cfDNA from blood. This technology solution can be used to evaluate cfDNA applications in cancer research, including treatment evaluation, recurrence monitoring, and prognostic guidance. We report the analytical validation of this Post-Diagnosis Cancer Research Solution, characterizing sensitivity, specificity, precision, and input range. Methods: cfDNA samples from cancer and non-cancer donors were analyzed. Analytical sensitivity (LoD95, limit of detection with ≥95% probability) was determined as a function of methyl variant allele fraction (MVAF), a measure of circulating tumor allele fraction. A total of 6 ng of cfDNA was used (cancer cfDNA titrated into a background of non-cancer cfDNA). LoD95 was defined as the lowest observed MVAF with ≥95% cancer signal detection across tested replicates or was estimated using probit regression for eligible samples. Analytical specificity was the rate of non-cancer classification among samples from 128 non-cancer donors. Precision (defined as concordance with the expected cancer/non-cancer result) was evaluated for 15 cancer donors tested near sample LoD95 and 8 non-cancer donors, with ≥18 replicates/donor. Reliability of classification was evaluated as a function of cfDNA input mass (0.25-100 ng total cfDNA, 14 cancer and 2 non-cancer donors). Results: A total of 12 different solid cancer types from 22 individuals with cancer were assessed in the LoD95 analysis. Median LoD95 at 6 ng total cfDNA input was 0.023% (10th percentile 0.0037%; 90th percentile 0.04%) MVAF. LoD95 estimates from in silico titration analyses of >200 clinical samples across a subset of the 12 cancer types were consistent with experimental LoD95 values. Analytical specificity was determined to be 98.47% (95% CI: 94.60-99.58%). Median precision across individuals was 100% (10th percentile 87%; 90th percentile 100%). Classification performance was accurate and consistent across a wide range of cfDNA input mass (100% correct cancer/non-cancer classification from 1.5-100 ng). Conclusions: Results demonstrate that this multi-cancer Post-Diagnosis Cancer Research Solution has high analytical sensitivity, specificity, and precision, with reliable performance across a broad cfDNA input range. This technology does not require a tumor sample and provides a cancer signal estimate in terms of MVAF. These features may enable understanding of cfDNA dynamics in a wide variety of cancers for research studies. Citation Format: Mohini Desai, Svetlana Rakhmanova Shchegrov, Shoujie Chai, Yifan Zhou, Tracy Nguyen, Yirang Cho, Collin Melton, Eric Scott, Manami Roychowdhury-Saha, Pei-Yun Chang, Rita Shaknovich, Byoungsok Jung. Analytical validation of a tissue-free, multi-cancer, post-diagnosis cancer research test that uses cell-free DNA methylation profiling [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB297.