This research study aims to investigate methods for enhancing lung cancer screening. The study will investigate whether an artificial intelligence (AI) tool, known as Sybil, can aid in predicting the risk of lung cancer. The investigators will also examine whether expanding the screening criteria (based on the guidelines of the Potter and American Cancer Society (ACS)) can help identify individuals at risk who are not currently included in the U.S. Preventive Services Task Force (USPSTF) guidelines.
Age 50-80 years at the time of consent
USPSTF: ≥20 pack-years, currently smoke or quit ≤15 years ago.
Potter: 20 years of smoking, regardless of intensity
ACS: ≥20 pack-years, no restriction on quit time
Inability to undergo LDCT
Current diagnosis or history of lung cancer < 5 years prior to study enrollment.
Life expectancy <1 year
Active lung infection requiring systemic therapy
Parallel
Sybil Artificial Intelligence (AI) screening
Diagnostic testLow-dose CT scans will be analyzed using the Sybil Artificial Intelligence (AI) screening tool
Expanded screening eligibility with Sybil AI risk scoring
To assess eligibility classification using USPSTF versus expanded criteria (Potter and American Cancer Society) and Sybil AI lung cancer risk scores calculated for all participants, including overlap between eligibility groups.
Sybil AI performance in USPSTF-eligible participants
To evaluate Sybil AI lung cancer risk prediction performance among USPSTF-eligible participants, assessed by discrimination and calibration metrics including AUC, sensitivity, specificity, and observed lung cancer incidence.
Combined biomarker, Sybil AI, and Brock model risk stratification
To assess risk stratification performance of integrated models incorporating immunometabolic biomarkers, Sybil AI risk scores, and the Brock model, assessed by AUC and risk reclassification measures.
Sybil AI performance across eligibility cohorts
To evaluate Sybil AI lung cancer risk prediction performance stratified by eligibility cohort (USPSTF vs expanded criteria), assessed by AUC, sensitivity, specificity, and calibration
Participant comprehension and acceptability of Sybil AI risk scores
To evaluate participant-reported comprehension, trust, and acceptability of Sybil AI risk scores measured using standardized survey instruments and summarized as scale scores and proportions
Clinical outcomes across eligibility groups
To evaluate lung cancer detection rate, stage at diagnosis, and low-dose CT appointment no-show rates compared across eligibility groups using clinical and imaging records
Lung cancer biorepository development
To evaluate number and characteristics of biospecimens collected, including biospecimen type, participant demographics, eligibility group, and linkage to clinical and imaging data
Evaluating blood-based immunometabolic biomarker levels
To evaluate blood-based immunometabolic biomarker levels measured and analyzed in relation to Sybil AI lung cancer risk scores and confirmed lung cancer diagnoses
Evaluating predictive performance
To evaluate predictive performance of models incorporating immunometabolic biomarkers and the Brock model assessed using discrimination metrics including AUC and risk reclassification
Click on the lead sponsor to view all of their trials.
University of Illinois at Chicago
Lead sponsor
Eli Lilly and Company
Collaborator
