Abstract 6423: Multiomic functional biomarkers for predicting the transition from inflammation to cancer

Abstract Background: Adulthood cancer results from (epi)genomic cellular changes associated with chronic inflammatory processes and maladaptive immune responses. Identifying health-to-disease transition sensors leading to early carcinogenesis beyond cell-autonomous signals is an unmet medical need. A holistic view could help defining a high-risk model based on functional biomarkers amenable to personalized screening interventions and interceptive measures. Machine learning algorithms modelling high-dimensional data could unveil the mechanisms underlying the transition from inflammation to cancer in smokers with cardiovascular disease (CVD), facilitating screening programs. Methods: We conducted a multiomics-based translational research on retrospective (FLEMENGHO, S67011), enrolling 97 smokers with CVD, of whom 38 developed cancer during the follow-up. This omics-based study included analytes related to inflammation, immunity, metabolism, gut barrier, proteomics, and clonal hematopoiesis (CHIP). To validate our hypotheses, we investigated a prospective study (PREVALUNG, NCT03976804) in 508 smokers with CVD aimed at predicting the 17 tobacco-associated cancers beyond the NLST, NELSON, PLCOm2012 screening scores. After inclusion, patients were scheduled for a low-dose chest CT-scan, and blood and feces samples were collected concomitantly. We continued our validation in a third independent cohort of patients with germline TP53 mutation (Li-Fraumeni syndrome) presenting a higher incidence of various cancers compared to the general population (LIFSCREEN, NCT01464086), and performed the same omics assessment on biological sample taken before cancer detection. Results: We performed multiomics related to inflammation, immunity, metabolism, microbiota and CHIP mutations on retrospective (FLEMENGHO) and prospective (PREVALUNG) cohorts enrolling 97 and 86 smokers with CVD, respectively, to validate a plasma fingerprint of cancer risk with AUROC=0.85 [CI95% 0.82; 0.87] in the test and AUROC=0.63 [CI95% [0.49; 0.79] for prediction at 1 year in the validation cohorts. This cancer risk prediction was confirmed in patients with TP53 germline mutations with AUROC=0.71 [CI95% 0.51; 0.90] at 2 years. Finally, a prospective blinded analysis validated this cancer risk score. We found that a set of 27 plasma-based analytes and CHIP mutations revealing key drivers of carcinogenesis grouped at-risk patients into three subgroups potentially amenable to distinct interceptive measures. Conclusion: The transition from inflammation to cancer remains an open conundrum. Multi-omics integrating organism dysfunctions (inflammation, immunity, metabolism, microbiota and clonal hematopoiesis) should be used to develop diagnosis tools. Here, we validate a set of 27 plasma biomarkers predicting time to cancer diagnosis in two cohorts of CVD tobacco users and one on Li-Fraumeni syndrome. Citation Format: Marine Fidelle, Deborah Suissa, Aline Renneville, Tarek Ben-Ahmed, Sylvere Durand, Olivier Caron, Olaf Mercier, Lisa Derosa, Birgit Geoerger, Guido Kroemer, Damien Drubay, Suzette Delaloge, Tatiana Kuznetsova, David Boulate, Laurence Zitvogel. Multiomic functional biomarkers for predicting the transition from inflammation to cancer [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 6423.