Biomarker-Based Eligibility for Lung Cancer Screening: Validation of the Protein-Based INTEGRAL-Risk Model

Area of Science:

• Oncology
• Biomarker Discovery
• Preventive Medicine

Background:

• Low-dose computed tomography screening reduces lung cancer mortality but excludes many high-risk smokers.
• Lung cancer incidence remains high in individuals with a smoking history not eligible for current screening protocols.

Purpose of the Study:

• To develop and validate the Integrative Analysis of Lung Cancer Risk and Etiology (INTEGRAL)-Risk model.
• To assess the protein-based model's performance in predicting absolute lung cancer risk in individuals with a smoking history.

Main Methods:

• Utilized 14 case cohorts from the Lung Cancer Cohort Consortium (3695 participants).
• Trained the INTEGRAL-Risk model using 7 cohorts (n=1951) with 13 proteins, age, and smoking history.
• Validated the model in 7 independent cohorts (n=1744) assessing discrimination (AUC) and calibration (E/O) at 1, 2, and 3 years.

Main Results:

• The INTEGRAL-Risk model demonstrated superior discrimination compared to the PLCOm2012 model at 1-year follow-up (AUC 0.88 vs 0.79).
• At 1-year, the model identified 85% of lung cancer cases, surpassing USPSTF 2021 (63%) and PLCOm2012 (70%) criteria at equivalent specificity.
• Model calibration remained good over 3 years (E/O ratio 0.87).

Conclusions:

• The protein-based INTEGRAL-Risk model significantly improves short-term lung cancer prediction in smokers.
• This model has the potential to enhance the selection of high-risk individuals for lung cancer screening.
• Protein-based risk assessment offers a promising alternative to questionnaire-based methods for lung cancer screening eligibility.