Fung, Eric T.

Abstract Background The Circulating Cell-free Genome Atlas (CCGA; NCT02889978) study is a prospective, multi-center, observational, case-control study with longitudinal follow-up to support the development of a plasma cfDNA-based multi-cancer early detection assay. We previously reported that the fraction of tumor-derived cfDNA fragments in plasma (estimated tumor fraction [TF]) is associated with detection of multiple cancer types using a prototype whole-genome bisulfite sequencing (WGBS) assay in a pre-specified substudy. In addition, detectability of breast cancer varied by hormone-receptor (HR) status: HR negative (HR-) breast cancer had a higher detection rate than the overall breast cancer cohort. Here, we performed a planned in-depth analysis of the breast cancer cohort to understand the main clinical and biological determinants of detectability using our WGBS assay. Methods Blood samples were prospectively collected from participants (pts) with newly-diagnosed, untreated breast cancer. In a pre-specified CCGA substudy, 511 (31.4%) of 1,628 pts had a clinicopathologic diagnosis of breast cancer (any stage, excluding stage 0) and a WGBS result. Pre-specified biological and clinical factors were assessed for correlation with TF: clinical stage, pathologic N stage, lesion size (by imaging), T stage, histologic grade, hormone receptor (HR) status, and proliferative rate (Ki-67 by IHC [MIB-1 clone]) via univariate and multivariate analyses. Ki-67 was obtained from pathology reports provided by enrollment sites and also assessed centrally (Mayo Clinic [Rochester, MN]). Stage-specific differences in detection rates between HR positive (HR+) and HR- breast cancers due to biological differences such as histologic grade and/or proliferative rate was also assessed. Results Of 511 pts, 94 (18.4%; 48.9% HR- and 12.3% in HR+) had WGBS-detected cancer. Higher TF was significantly associated with cancer detection (p<0.001). In univariate analyses, the variables significantly associated with TF were clinical stage, pathological N stage, and histologic grade (p<0.001 each). Primary lesion size was modestly correlated with TF (all stages combined: Spearman's correlation coefficient rs=0.4, p<0.001), and when grouped by early and late stages (stages I/II: rs=0.27, p<0.01; stages III/IV: rs=0.50, p<0.01); clinical T stage was significantly associated with TF only when all stages were combined (p<0.001). Proliferative rate (Ki-67) was positively correlated with TF at stages I/II (rs=0.29, p<0.05) but not III/IV (rs=-0.24, p=0.34); Ki-67 results obtained through Mayo were similar to that from enrollment sites (Pearson’s correlation coefficient rp=0.68, p<0.001). HR- breast cancers had significantly higher TF versus HR+ breast cancers only at stage II (p<0.001). In addition, compared to HR+ breast cancers, HR- breast cancers had a significantly higher proportion of high histologic grade (p<0.001 at stages I and II) and proliferative rate at earlier stages (p<0.05 at stages I, II, and III). In the multivariate analysis, clinical stage (stage III vs I and IV vs I) and HR status remained the most significantly associated with TF (p<0.001 each). Conclusions Using a prototype methylation-based plasma cfDNA cancer detection assay, TF was found to be associated with detection, and with clinical and biological features of breast cancer. Features routinely used to assess clinical aggressiveness (eg, HR- status and histologic grade) also demonstrated stage-specific associations with TF. Given the known relationship between TF and detection, these data suggest that clinical and biological cancer features may also provide insight into the variability of cancer detection using plasma-based cfDNA tests. Citation Format: Minetta C. Liu, Jodi M. Carter, Daniel W. Visscher, Karla Kopp, Rita Shaknovich, Xiaoji Chen, Kathryn N. Kurtzman, Shilpen Patel, Jacqueline D. Brooks, Carlo R. Cosenza, Jafi A. Lipson, Donald A. Richards, Fergus J. Couch, Zhao Dong, Hai Liu, Oliver Venn, Joerg Bredno, Eric T. Fung, Anne-Renee Hartman. Blood-based cancer detection in plasma cell-free DNA (cfDNA): Evaluating clinical and pathologic tumor characteristics in participants with breast cancer [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-01-01.

3049 Background: For multi-cancer detection using cfDNA, TOO determination is critical to enable safe and efficient diagnostic follow-up. Previous array-based studies captured < 2% of genomic CpGs. Here, we report genome-wide fragment-level methylation patterns across 811 cancer cell methylomes representing 21 tumor types (97% of SEER cancer incidence), and define effects of this methylation database on TOO prediction within a machine learning framework. Methods: Genomic DNA from 655 formalin-fixed paraffin-embedded (FFPE) tumor tissues and 156 isolated cells from tumors was subjected to a prototype 30x whole-genome bisulfite sequencing (WGBS) assay, as previously reported in the Circulating Cell-free Genome Atlas (CCGA) study (NCT02889978). Two independent TOO models, one with and one without the methylation database, were fitted on training samples; each was used to predict on the test set. A WGBS classifier was used to detect cancer at 98% specificity; reported TOO results reflect percent agreement between predicted and true TOO among those detected cancers (166 cases: 81 stage I-III, 69 stage IV, 16 non-informative). Results: Genome-wide methylation data generated from this database allowed fragment-level analysis and coverage of ~30 million CpGs across the genome (~60-fold greater than array-based approaches). Incorrect TOO assignments decreased by 35% (20% to 13%) after incorporating methylation database information into TOO classification. Improvement was observed across all cancer types and was consistent in early-stage cancers (stage I-III). Respective performances in breast cancer (n = 23) were 87% vs 96%; in lung cancer (n = 32) were 85% vs 88%; in hepatobiliary (n = 10) were 70% vs 90%; and in pancreatic cancer (n = 17) were 94% vs 100%. Results using an optimized approach informed by these results in a large cohort of CCGA participants will be reported. Conclusions: Incorporating data from a large methylation database improved TOO performance in multiple cancer types. This supports feasibility of this methylation-based approach as an early cancer detection test across cancer types. Clinical trial information: NCT02889978.