Exploring the association of smoking status with clinical outcomes in patients with mPDAC treated on a clinical trial of maintenance niraparib plus immune checkpoint blockade.

Abstract

e16302 Background: Previous research has hinted at a positive association between smoking history and outcomes in non-small cell lung cancer patients being treated with immune checkpoint blockade (Wang L. et al., 2021)(Cortellini A. et al., 2021). The PARPVAX trial (NCT03404960) assessed the efficacy of maintenance niraparib plus either nivolumab (Arm A) or ipilimumab (Arm B) in platinum-sensitive patients with mPDAC (Reiss KA. et al., 2022). We investigated whether there was an association between patient smoking history and outcomes on this trial. Methods: Patients treated on NCT0340960 were categorized by smoking status (binary; never/any smoking history). We performed multivariable Cox proportional hazards models for PFS and OS with a pre-specified set of covariates that underwent negative backward selection until the ratio of events to degrees of freedom in the model was less than 10 to avoid overfitting. Models were also performed within each arm in an exploratory fashion to determine if differences were observed between treatment groups. Results: Overall, 42% (Arm A) and 58% (Arm B) of patients had any history of smoking. Any smoking history was significantly associated with improved OS in the multivariate model (HR = 0.45 [0.24 – 0.86]; P = 0.015). The multivariate PFS analysis demonstrated a 37% reduction in the hazard of progression or death, but the result was not statistically significant (HR = 0.63 [0.37 – 1.09]; P = 0.10). When separated by treatment arm, any smoking history had a significant association with PFS (HR = 0.32 [0.14 – 0.73]; P = 0.007) and OS (HR = 0.30 [0.11 – 0.80]; P = 0.016) in Arm B but not Arm A. Conclusions: Our ad hoc analysis of outcomes by smoking status in a population of patients with advanced PDAC undergoing experimental maintenance treatment with PARPi and ICB suggests that a smoking history may be associated with improved outcomes, particularly in patients who received anti-CTLA4 therapy. Sample size greatly limited adjustment for potential confounders and these results are hypothesis generating only. [Table: see text]