Risk-Based Lung Cancer Screening in Clinical Practice.

Abstract

Introduction: Current US lung cancer screening guidelines use only age and smoking history; however, individual risk calculators may better stratify risk.

Methods: In a referred cohort design, we implemented a multisite lung cancer screening program across 4 states. We screened patients who qualified by either the USPSTF₂₀₁₃ criteria or a PLCO_m2012 risk of >1.34%. Invasive procedures were abstracted retrospectively. We compared the incidence and prevalence of lung cancer among patients who qualified by only USPSTF₂₀₁₃ or PLCO_m2012 as well as along the continuum of prospective lung cancer risk using PLCO_m2012.

Results: Of 2471 screened patients, 114 had lung cancer. Eighty-four percent of all patients and 91% of patients who were diagnosed with cancer qualified by both criteria. Prevalence lung cancers were over 7 times higher in the 10% of the cohort with the highest prospective risk than the lowest risk 10%. Incidence cancers were higher among patients who qualified only by PLCO_m2012 (3.6 per 1000 person-years) compared to patients who qualified only by USPSTF₂₀₁₃ (0 per 1000 person-years). Seventy-four percent of screen-detected NSCLC were stage I or II. Three (4.5%) surgical resections were performed for screen-identified nodules which proved to be benign. Overall, 106 patients (4.3%) underwent an invasive intervention due to screening.

Conclusions: Most patients qualified for lung cancer screening by both UPSTSF₂₀₁₃ and PLCO_m2012 criteria. Incidence cancers were higher among patients who qualified by PLCO_m2012 but not USPSTF₂₀₁₃ criteria. Prevalence and incidence cancer identification increased with prospective risk. Invasive procedures and resections for benign disease were relatively low.