Personalizing Veterans' Lung Cancer Screening and Diagnosis

Veterans Health Administration, Decatur PA

Investigators

Abstract

Personalizing Veteran's Lung Cancer Screening and Diagnosis Lung cancer is the leading cause of cancer-specific mortality in United States, including among Veterans. Annual lung cancer screening with low-dose chest computed tomography (LDCT) in high-risk individuals has been demonstrated to save lives in two large, randomized-controlled trials and is recommended by the US Preventive Services Task Force (USPSTF) for individuals aged 50 to 80 with at least 20 pack years of smoking and who currently smoke or quit in the past 15 years. Yet, individuals eligible under these criteria would only capture 45- 65% of all lung cancers. In this proposal, we seek to personalize lung cancer screening and diagnosis. Veterans are at great risk for lung cancer given the higher prevalence of tobacco smoking and service-connected exposures, including agent orange, asbestos, burn pits, and ionizing radiation. These hazardous materials and exposures, frequently encountered in the military, are associated with lung cancer but are not part of current recommended eligibility criteria. Furthermore, lung cancer screening results in an increase in the detection of indeterminant pulmonary nodules. Our teams' data suggests that two candidate biomarkers could mitigate sending Veterans with benign nodules for invasive procedures and shorten the time to lung cancer diagnosis. First, an artificial intelligence model (Lung Cancer Predictor Convolutional Neural Network, LCP-CNN) significantly improves discrimination between benign and malignant nodules on LDCTs. Our LCP-CNN algorithm generates accurate reclassification of up to 30% of intermediate risk nodules (10-70% risk for cancer) to lower or higher risk levels, thus, decreasing the number of invasive procedures and shortening the time to diagnosis. The other biomarker is CYFRA 21-1, a serum protein cancer biomarker for which we developed a high sensitivity assay. Based on these preliminary data, we propose to answer the following clinical question: Can we safely offer lung cancer screening to Veterans who are at greater risk beyond their age and smoking history but due to their personal and service-related exposures while improving the rate of early and non- invasive detection of lung cancer? Aim 1: In a prospective cohort of 800 Veterans at Nashville, Denver and Louisville between 50-80 years of age who ever smoked cigarettes, we will evaluate if a new, personalized lung cancer screening eligibility criteria that incorporates military exposures, COPD, personal history of tobacco-related cancer, and family history of lung cancer results in at least the same number of lung cancers detected compared to the standard of care USPSTF screening criteria in use by our Lung Precision Oncology Program lung cancer screening programs. Aim 2: Among a cohort of Nashville Veterans with positive screenings (n=520), we will test the hypothesis that the LCP-CNN in combination with hs-CYFRA 21-1 reduces the number of invasive procedures and shortens the time to lung cancer diagnosis versus the Brock Model. We will manage positive screens from Aim 1 and test the combined biomarker approach proposed in Aim 2 to rule in cancer early and rule out benign lung nodules. Impact: We will safely personalize lung cancer screening in the Veteran community to improve the number of cancers detected early, reduce the rate of invasive procedures, and avoid a missed chance for cure for early detected lung cancers. This project will provide proof of principle for personalizing lung cancer screening criteria among Veterans and test the clinical utility of a combined biomarker-based approach to manage screen-detected indeterminate pulmonary nodules. The proposed research will provide actionable evidence for Veterans with military exposures and would impact VHA policy. Our team is poised to conduct a rigorous and necessary evaluation and the findings will inform the development of a larger, national randomized-controlled CSP trial to test if LDCT screening among Veterans meeting expanded and personalized criteria results in improved survival.