Multiomics blood-based biomarkers in non-small lung cancer demonstrate specificity against smoking and high performance in early detection

SESSION TITLE: Emerging Biomarkers and Biotechnology SESSION TYPE: Rapid Fire Original Inv PRESENTED ON: 10/09/2023 12:00 pm - 12:45 pm PURPOSE: Lung cancer is the leading cause of cancer-related deaths worldwide, with 236,740 new cases and 118,830 deaths estimated in the United States in 2022, with smoking as a major driver of risk. Liquid biopsies based on peripheral blood sampling hold great promise to reduce lung cancer morbidity and mortality by enabling early detection to downstage disease at initial diagnosis, theragnostic identification and efficacy monitoring to maximize benefit and minimize harm of specific therapeutic interventions, and detection of residual disease. METHODS: We conducted a case-control study of 1031 subjects composed of 361 subjects with untreated non-small cell lung cancer (NSCLC) and 670 matched controls which included 340 subjects with an assortment of potentially confounding comorbidities including salient pulmonary and gastrointestinal conditions. To balance against important confounders, we identified a subset of 540 subjects (180 cancers and 360 non-cancer controls with and without comorbidities) balanced for age, assigned sex at birth, and (never) smoking status. Deep multi-omic profiling was performed on each subject sample across 3 dimensions: proteins, metabolites, and mRNA. All molecular data were normalized using standard methods specific to each assay. RESULTS: We detected 4423 proteins, 111152 RNA transcripts, and 1307 metabolites on average for each subject. Of note, the performance of the PrognomiQ proteomic platform based on the Seer Proteograph™ compared very favorably to the deepest published plasma proteomic studies performed at scale of ≥500 subjects which each detected <600 proteins on average per subject. The study cohort was split into a training set comprising of 324 subjects and a validation set of 216 subjects for the purposes of classifier development. The validation set demonstrated a classification AUC of 0.88 [0.83-0.93]. Across all stages, the combined multi-omic classifier had a sensitivity of 81% [70-92%] at a specificity of 83% [75-91%] on the validation set. Stage I sensitivity was 77% [55-98%] at a specificity of 83% [75-91%] and Stage II-IV sensitivity was 82% [70-94%] at a specificity of 83% [75-91%], Of note, the AUC of the trained classifier with smoking as a target was 0.54 [0.43-0.64], suggesting against smoking as a significant contributor to classification of malignancy in this population. CONCLUSIONS: This NSCLC classifier based on a multi-omic approach including unprecedented proteomic interrogative depth at scale demonstrated high performance in early and all-stage detection as well as specificity against a number of potential confounders, including smoking. CLINICAL IMPLICATIONS: These encouraging data may serve as the foundation of a multi-omics assay for peripheral blood-based liquid biopsies for the early detection of lung cancer. DISCLOSURES: Employee relationship with PrognomiQ Inc Please note: 3/29/2021-present Added 03/31/2023 by Chinmay Belthangady, source=Web Response, value=Salary No relevant relationships by Ehdieh Khaledian No relevant relationships by Brian Koh No relevant relationships by Manway Liu Seer, Inc. relationship with PrognomiQ, Inc. Please note: 9/2017 to present Added 04/03/2023 by Philip Ma, source=Web Response, value=equity Removed 04/03/2023 by Philip Ma, source=Web Response Employee relationship with PrognomiQ, Inc. Please note: 10/2020 to present Added 04/03/2023 by Philip Ma, source=Web Response, value=equity Employee relationship with Prognomiq Please note: Dec 2021-current Added 04/07/2023 by Madhuvanthi Ramaiah, source=Web Response, value=Salary No relevant relationships by Saividya Ramaswamy Employee relationship with PrognomiQ, Inc. Please note: 12/01/2020-current Added 03/31/2023 by Bruce Wilcox, source=Web Response, value=Salary Employee relationship with PrognomiQ, Incl Please note: 12/01/2020-Current Added 03/31/2023 by Bruce Wilcox, source=Web Response, value=Ownership interest Consultant relationship with Olympus Please note: 2021-present by Lonny Yarmus, value=Grant/Research Support