Proteogenomic framework for precision lung cancer screening

The National Cancer Institute awarded $546,542 to the Centre International de Recherche sur le Cancer (IARC) on August 11, 2026, to develop a proteogenomic framework for precision lung cancer screening under the Cancer Cause and Prevention Research program (CFDA 93.393).

The project refines the INTEGRAL protein risk model, a 13-marker assay designed to improve lung cancer screening eligibility assessment and nodule management following low-dose computed tomography. The existing INTEGRAL model captures 85 percent of incident lung cancer cases compared to 63 percent sensitivity under current U.S. Preventive Services Task Force criteria, but the research will enhance predictive performance by incorporating genetic factors influencing protein concentration and lung cancer susceptibility. Heritable factors account for more than 20 percent of variance in most protein markers, and this work will determine whether heritable variation in non-causal protein markers introduces noise requiring model refinement.

IARC, the specialized cancer research agency of the World Health Organization, performs the work in France. The award period runs through July 31, 2031. The assistance type is a Project Grant.

Abstract

Lung cancer is the leading cause of cancer death globally, accounting for 1.8 million deaths annually. Screening by low-dose computed tomography (ldct) has revolutionized our ability to diagnose lung cancer at early stages when an effective cure can be offered. However, considering the negative consequences of screening, targeting individuals who are most likely to benefit is of outmost importance. Current US preventive services task force (uspstf) screening criteria leave 40% of ever-smoking lung cancer cases ineligible for screening, but smoking- based risk models can to some extent improve the sensitivity of eligibility criteria. However, these traditional risk models tend to favour older individuals who gain fewer life-years from screening. Research aiming to develop biomarkers that provide additional information on lung cancer risk has therefore been highlighted as a priority. The integral protein panel was developed as a fit-for-purpose assay intended to inform risk assessment when evaluating screening eligibility and nodule management following ldct. The integral risk model comprised of 13 protein markers outperforms existing lung cancer risk assessment tools when evaluated on independent pre-diagnostic samples representative of the target population. These data demonstrate that the integral risk model can improve sensitivity of the current uspstf screening criteria (63%) by capturing 85% of incident lung cancer cases. Whereas the performance of the integral panel is promising, there is ample room for further improvement and refinement by considering genetic factors that influence protein concentration and lung cancer susceptibility. Heritable factors can explain more than 20% of the variance for most of the markers included on the integral protein panel. For protein markers that are not causally influencing the risk of developing lung cancer, this heritable variation will introduce noise that decreases the performance of risk prediction models and leads to misleading risk assessment for some individuals. We have demonstrated that IT is possible to correct for this "genetically induced noise" and improve the accuracy of blood-based protein markers. We hypothesize that accounting for genetic sources of protein variation will improve both short- and long-term predictions of the protein-based integral risk model. Further integrating this with information on genetic susceptibility to lung cancer will produce a powerful tool for risk assessment. We propose a comprehensive project where we will address three specific aims: (1) optimize models to genetically correct the proteins included on the integral protein panel, (2) elucidate regulatory pathways that will inform more accurate genetic adjustment model and develop a transcriptome-based risk score for lung cancer (ptrs), and (3) genotyping of samples from a world- leading cohort resource for lung cancer risk modelling to train and validate the integral risk model 2.0 that incorporates genetically corrected proteins and lung cancer ptrs. The completion of these aims will provide robust evidence that can readily be translated into clinical practice and patient benefit.