https://meetinglibrary.asco.org/record/158492/abstract
Authors:
Background:Poor performance status (PS) is associated with worse clinical outcomes for pts with mNSCLC. However, in the real-world PS is not routinely captured in electronic health records (EHRs). Pts with poor PS may also have cachexia and low BMI. As such, underweight BMI may be a proxy for deteriorated real-world PS when PS score is missing. We have shown that female gender and positive EGFR/ALK mutation status predict longer OS, but not age or other factors. We explore the association of baseline BMI with OS in real world mNSCLC pts treated with N or P.Methods:We conducted a retrospective analysis of mNSCLC pts treated with N or P using de-identified real-world data (RWD) from the Flatiron Health network. Index date was start of first single agent N or P. Baseline BMI (kg/m2) was calculated from the most recent weight and height recorded within 30 days prior to index date and categorized as: Underweight ( < 18.5), Normal (18.5 – 24.99), Overweight (25 – 29.99), and Obese (30+). Association of baseline BMI with OS was assessed using multivariate Cox proportional hazards model adjusted for gender, age, and EGFR/ALK status prior to N or P start. Eligible pts had valid values for all parameters.Results:703 pts met inclusion criteria (Table). For OS, gender was significant overall, EGFR/ALK mutation was significant overall and for women.Conclusions:We observed BMI-based differences in OS for real world mNSCLC pts treated with N or P. Underweight BMI is associated with shorter OS and obese BMI is associated with longer OS. In pts with cancer, declining BMI and PS may signal biologic processes indicative of progressive disease that negatively affect OS. Since PS data is commonly missing in EHRs, our results suggest that BMI can potentially be used as a proxy for poor PS in RWD studies. Investigating the impact of changes in BMI on OS and correlation of BMI with PS would further verify this assumption.
Sources:
ASCO Annual Meeting