AS722 - Integrating Biomarkers into Eligibility Criteria for Lung Cancer Screening

Investigator Names and Contact Information

Introduction/Intent

Lung cancer screening by low-dose computed tomography (LDCT) has revolutionized the prospect of preventing lung cancer death by allowing for detection at early stages when curative treatment can be offered. To translate screening into broad public health impact, a central challenge is that current criteria for lung screening eligibility require that participants have a heavy smoking history. People with many years of cessation are excluded, as are people who never smoked. Paradoxically, the proportion of lung cancer cases occurring among this non-eligible group is rising. The overarching aim of Project 2 is to address this challenge by integrating complementary information from blood-based biomarkers to identify people who have high lung cancer risk despite not meeting current eligibility criteria.

Using a state-of-the-arts proteomics platform, the first INTEGRAL program successfully developed a novel protein biomarker panel with excellent risk discriminatory performance that can readily be implemented to assess eligibility for lung cancer screening. Project 2 will take two key steps to translate this panel into CT-screening programs: first, to develop and validate a biomarker-informed risk prediction framework for smoking-related lung cancer; and second, to evaluate if it is practically feasible to implement the panel to improve the identification of high-risk individuals for screening. Furthermore, for people who have never smoked, predicting lung cancer using standard information on demographics and health history is difficult, and the discriminative ability of current tools is poor. Following our success in predicting smoking-related lung cancer using protein markers, we will identify and validate similar markers for lung cancer among never smokers, with the long-term goal to adapt the INTEGRAL panel to predict lung cancer risk regardless of smoking history.

We will address the following specific aims:

Develop and validate an integrated biomarker-based risk prediction framework to inform lung screening eligibility among people with a smoking history. We will leverage existing data for the INTEGRAL panel measured on 1,700 lung cancer cases and 2,900 cohort representatives sampled from 14 prospective cohorts in the Lung Cancer Cohort Consortium. We will develop a prediction model for absolute risk of lung cancer over 3 years, using 7 cohorts as a training set and the other 7 cohorts as an independent validation set. We will compare the risk-discriminatory performance of the biomarker-informed model with existing models and develop a framework to identify potential screenees who might benefit from biomarker-based risk assessment.
Evaluate the practical feasibility of using a pre-screening biomarker test to inform lung screening eligibility. This aim will assess whether it is practically feasible to collect a blood sample from potential screenees, return the biomarker results and risk estimate in a timely manner, and recall participants who become eligible after biomarker testing for screening. This aim will include qualitative studies of how participants react to biomarker-based risk assessment and whether it affects their intention to participate in screening. The study population will be nested within the St Elizabeth Healthcare Lung Cancer Screening Program and will involve risk assessment in 2,000 potential screening participants.
Identify and validate biomarkers of imminent lung cancer among people who never smoked. We will measure 3,000 protein markers using the Olink Explore platform in 616 cases and 616 matched controls, using pre-diagnostic blood samples taken at most 5 years prior to lung cancer from 10 cohorts in the Lung Cancer Cohort Consortium. The results will be used to update the INTEGRAL custom protein panel which will allow the same protein panel to be used to assess risk of imminent lung cancer regardless of smoking status.

Our long-term vision is that by using personalized information from risk-informative and readily implementable multiplex panel, lung cancer screening can be more precisely targeted to people at high risk of lung cancer. This would allow a higher fraction of lung cancer cases to be detected early without subjecting more individuals to CT-screening. We also envisage that this panel will be highly useful outside the screening context in the general practice to work up never-smoking patient who present with symptoms.