Abstract 1953: Detection of clinically actionable somatic variants in post-operative samples from patients with molecular residual disease

Abstract Introduction: Molecular residual disease (MRD) detection via ctDNA analysis refines recurrence risk estimation. Patients with MRD have substantial risk for metastatic disease and may have genomic biomarkers that could inform treatment. We describe detection rates of clinically actionable genomic biomarkers in plasma based on ctDNA fraction (TF) in the post-operative MRD setting. Methods: Positive Guardant Reveal samples were tested using Guardant360. Reveal is a tissue-free epigenomic assay for post-operative MRD detection and quantitation in stage II-III colorectal (CRC), breast, and lung cancers; Guardant360 is intended for comprehensive epigenomic and genomic profiling (739 genes). Actionable variants were defined based on associations with targeted or immunotherapy benefit referenced in guidelines and those suggestive of another primary cancer (e.g., ESR1 resistance mutations in lung or CRC). We evaluated the association between TF and likelihood of actionable variant detection. Results: 4, 203 Reveal positive samples (median TF 0.17%, IQR 0.05-1%) from 4, 165 patients were sequenced on Guardant360. Actionable variants were detected in 28% (1% were suggestive for another primary cancer). Increasing TF was associated with higher likelihood of clinically actionable variant detection. Across all three cancer types, TFs >0.05% had high yield of actionable genomic findings (∼36%) and 97% of all actionable variants found were above this threshold (Table 1 Table 1. Rate of actionable variant detection by MRD TF and cancer type CRC Breast Lung Overall No. samples 2, 769 781 653 4, 203 % with actionable variant 29% 29% 24% 8% % >0.05% TF 77% 79% 63% 75% % with actionable finding by TF <0.05% 2% 6% 5% 3% >0.05% 37% 35% 36% 36% Proportion of all detected actionable variants by TF <0.05% 2% 4% 8% 3% >0.05% 98% 96% 92% 97% Top five actionable genes/variants detected in samples with >0.05% TF KRAS (26%), BRAF V600E (5%), MSI-H (4%), NRAS (2%), ESR1 (1%) PIK3CA (18%), ESR1 (12%), TMB-H (7%), PTEN (3%), AKT1 (2%) KRAS (16%), EGFR (10%), ERBB2 (3%), ESR1 (3%), MET (2%) ). Samples <0.05% TF rarely had detectable actionable variants. Conclusion: These data suggest profiling with Guardant360 is reasonable in MRD-positive samples with TF >0.05%. Reflex testing may have utility for rapid return of genomic data to facilitate treatment planning in these high-risk patients. This paradigm may be particularly beneficial for patients with limited tumor specimens or who acquired resistance mutations to targeted therapy in the (neo)adjuvant setting. Citation Format: Thereasa A. Rich, Jun Zhao, Christopher Pretz, Yupeng He, Derek Dustin, Katie Quinn, Tingting Jiang, Errin Lagow, Krysten Shipley, Kristin Price, Leslie Bucheit, Kimberly Banks. Detection of clinically actionable somatic variants in post-operative samples from patients with molecular residual disease [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1953.