Prognostic value of electronic health records-based frailty measures for all-cause mortality in older patients with non-small cell lung cancer

Introduction: Frailty screening is important to guide treatment decisions for older patients with non-small cell lung cancer (NSCLC). However, the performance of frailty measures (FMs) remains unclear. This study aimed to evaluate the prognostic value of FMs based on electronic health records (EHR) data in clinical settings for allcause mortality in older patients with NSCLC. Materials and Methods: We retrospectively analyzed 4253 patients aged >= 65 years, newly diagnosed with NSCLC (2007-2018) using EHR data from the National Cancer Center, Korea. Frailty was measured by either laboratory tests (frailty index based on routine laboratory tests [FI-Lab]), comorbidities and performance status (electronic Frailty index [eFI]), or both (combined frailty index [FI-combined]). Patients were categorized as frail or nonfrail. Cox proportional hazards models and C-index were used to estimate the predictive ability of FMs for allcause mortality in 1 year, 3 years, and 5 years post-diagnosis, adjusting for age, sex, and SEER stage. Results: EHR-based FMs could enhance the prognostic ability to predict the survival of older patients with NSCLC. In the total population, FI-Lab showed the largest predictive value, especially for 1-year mortality with an adjusted hazard ratio for frail vs. non-frail groups of 2.25 (95 % CI 2.02-2.51) and C-index of 0.74 compared to 0.72 in the base model (p-value<0.001). FI-Lab could improve the prognostic ability for 1-year mortality in patients with regional and distant SEER stages and those receiving systemic therapy, whereas FI-combined could improve the prediction of 3-year and 5-year mortality in patients with localized disease and receiving surgery. Discussion: Easy-to-use FMs derived from EHR data can enhance the prediction of all-cause mortality in older patients with NSCLC. Oncologists can utilize comprehensive FMs comprising comorbidities, functional status, and subclinical tests or FI-Lab, depending on the patient's medical condition, to facilitate shared cancer care planning.