P16.03 Early Mortality Associated Factors With Immune Checkpoint Inhibition in Real-World Canadian NSCLC Patients

Abstract

Immune checkpoint inhibitors (ICI) have been demonstrated to improve survival in metastatic non-small cell lung cancer (mNSCLC); however, within randomized controlled trials the phenomena of early crossover of Kaplan-Meier survival curves favouring the control group suggests a subpopulation of NSCLC patients at risk for early mortality (EM), even among these carefully selected patients. Given the increased heterogeneity and often poorer functional status of real-world patients, we sought to explore the frequency of, and factors associated with, EM for unselected real-world patients receiving ICI. Clinical, treatment and outcome data for mNSCLC patients, in receipt of ICI between 2015 and 2018, were extracted from the institutional Glans-Look Lung Cancer Database. EM was defined as death from any cause within 60 days of ICI initiation. Univariate analysis was performed to compare EM/non-EM cohorts, and significant factors included in multivariate analysis. A total of 339 patients with mNSCLC receiving ICI were evaluated. 58% received nivolumab and 42% pembrolizumab, primarily in the 2L+ setting (79%), with a median post-ICI initiation survival (mICI-OS) of 8.6 months. 15% experienced EM. Compared to the non-EM cohort, the EM cohort had reduced mICI-OS (30 days vs. 11.1 months), and were more likely to have abnormal baseline lab values, age < 70 years at ICI treatment start and ECOG > 1, but were less likely to report an adverse event. 71% of the EM cohort had progressive disease (PD) documented before death, at a median 29 days after ICI-initiation, compared to 3.8 months for non-EM; 29% of the EM cohort died before response to treatment could be quantified. Binomial logistic regression identified a neutrophil-lymphocyte ratio ≥ 5 [odds ratio (OR): 3.5; 95% CI 1.2 – 9.8; p=0.02], Albumin < 33g/L [OR: 3.1; 95% CI 1.3 – 7.3, p=0.01], and lactate dehydrogenase > 234U/L [OR: 3.8; 95% CI 1.5 – 9.4; p=0.004] as significant predictors of EM. Early development of PD was common among patients with EM, who differed clinically from non-EM patients. Abnormal baseline laboratory values significantly predicted EM, but neither age nor ECOG were independent predictors of early mortality when controlling for confounders. This suggests careful consideration of baseline laboratory metrics for patients eligible to receive ICI therapy is required to identify those with low likelihood of benefit.