Abstract 5113: Liquid biopsy-based multiomics profiling using low-pass whole genome sequencing and proteomics with computational modeling reveals molecular correlates of disease severity in EGFR/ALK wild type NSCLC patients

Zheng Feng,Danyi Wang,Francesco Vallania,Neil Smith,Anupriya Tripathi,Juergen Scheuenpflug

DOI: https://doi.org/10.1158/1538-7445.am2022-5113

IF: 11.2

2022-06-17

Cancer Research

Abstract:Background: Molecular features of disease severity provide a basis for patient stratification, and may hint at previously unrecognized drug targets. Liquid biopsy-based multiomics profiling of plasma coupled with computational modeling provides a promising approach to interrogate both tumor and non-tumor derived signals, to support patient-centric personalized precision oncology. Here, we utilized a blood-based multiomics platform to identify, refine, and combine signals to categorize late-stage, treatment-naive, EGFR/ALK wild type NSCLC. Methods: Cell-free DNA (cfDNA) was extracted from 34 late-stage, treatment-naive, EGFR/ALK wild type NSCLC patients and low pass whole-genome sequencing was performed to characterize cfDNA fragments, which reflect nucleosome protection and chromatin state. Gene activation for protein-coding genes (GAP scores) was modeled from fragment distribution around transcription start sites. The abundances of 644 plasma proteins including markers of the immune response, inflammation, cancer, and DNA repair were measured. Partial least-squares discriminant analysis (PLS-DA) was used for marker selection and sample separation by smoking status with either genomics or proteomics data. Results: By statistically modeling clinical data, we identified a significant association showing never-smokers being enriched in the highest N stage, (P = 7.95E-3). Using PLS-DA, we identified proteomic (top 3 proteins: ST1A1, AXIN1, STAMPB) and GAP score (top 3 genes: CHD3, FANCF, METTL3) markers that separated patients by smoking status, suggesting underlying molecular differences. By performing pathway analysis on GAP scores, we identified 108 significant pathways enriched in the never-smoker group (FDR < 5%), 18 of which were associated with immune cell signatures, specifically myeloid cells and macrophages. Based on our previous association of IL1RN GAP scores with disease progression in NSCLC, we examined whether these results are confirmed. In this cohort, patients with high IL1RN GAP scores were significantly associated with worse lymph-node staging (N) (P = 6E-3), which is indicative of worse prognosis. Pathway analysis comparing patients with high vs low IL1RN scores revealed 102 significant pathways (FDR < 5%), including leukocyte activity and regulation of inflammatory response. Conclusion: The liquid biopsy-based multiomics platform suggests an association between never-smoker status and pathways of inflammatory response, resulting in more severe lymph-node staging. Future applications of this multiomics platform provide promise for stratification of NSCLC patients for patient-centric precision oncology. Citation Format: Zheng Feng, Danyi Wang, Francesco Vallania, Neil Smith, Anupriya Tripathi, Juergen Scheuenpflug. Liquid biopsy-based multiomics profiling using low-pass whole genome sequencing and proteomics with computational modeling reveals molecular correlates of disease severity in EGFR/ALK wild type NSCLC patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl) nr 5113.

oncology