BACKGROUND: Lung cancer is the leading cause of cancer-related deaths in the United States. Several anaplastic lymphoma kinase (ALK) rearrangement inhibitors have been approved for the treatment of metastatic ALK-positive non-small cell lung cancer (NSCLC). Effective disease management requires an understanding of how these treatments are used in clinical practice, since low treatment adherence and/or early discontinuation have been associated with poor patient outcomes. Owing to the recency of approvals, real-world data on the use of ALK inhibitors in patients with ALKpositive NSCLC are currently limited; this represents a notable gap in our understanding of ALK treatment use. OBJECTIVE: To assess real-world adherence and persistence with ALK inhibitors in patients with ALK-positive NSCLC. METHODS: This retrospective observational study used US commercial claims for patients aged at least 18 years with lung cancer receiving ALK inhibitors (alectinib, brigatinib, ceritinib, crizotinib) between July 1, 2015, and December 31, 2018. Patients' first and any subsequent ALK inhibitor uses were categorized into ALK inhibitor-naive and ALK inhibitor-pretreated cohorts, respectively. Adherence was measured by medication possession ratio and persistence by time from treatment initiation to discontinuation (earliest of a treatment switch or greater than a 60-day gap). Descriptive statistics were used to summarize patient characteristics. Cohort comparisons were made using chi-square tests and t-tests. Persistence and time to next ALK inhibitor were analyzed using Kaplan-Meier methods and the log-rank test. Poisson and Cox regression models of adherence and persistence, respectively, were applied to compare ALK inhibitors. RESULTS: We identified 1,482 patients treated with alectinib (n = 445) or crizotinib (n = 1,037) in the ALK inhibitor-naive cohort; 604, 142, and 134 patients received alectinib, brigatinib, or ceritinib in the ALK inhibitor-pretreated cohort. Adherence during the treatment period (95%-97%) and the proportion of patients with a medication possession ratio of at least 0.8 (92%-95%) were similar for all ALK inhibitors. In the ALK inhibitor-naive cohort, median time to treatment discontinuation with alectinib and crizotinib was 27.1 and 8.8 months, respectively; patients receiving alectinib were 46% less likely to discontinue than patients receiving crizotinib (adjusted hazard ratio [aHR] [95% CI]: 0.54 [0.44-0.65]; P < 0.0001). In the ALK inhibitor-pretreated cohort, the discontinuation risk for alectinib was 64% lower than for ceritinib (aHR [95% CI]: 0.36 [0.27-0.49]; P < 0.0001) and 34% lower than for brigatinib (aHR [95% CI]: 0.66 [0.42-1.02]; P = 0.062). CONCLUSIONS: To our knowledge, this study is the first to address a current research gap by assessing real-world adherence and persistence with ALK inhibitors among patients with ALK-positive NSCLC in real-world clinical practice. Alectinib was associated with longer real-world persistence than other ALK inhibitors, despite similar adherence. Further research with more patients and longer follow-up is needed to link persistence to real-world clinical outcomes. DISCLOSURES: This study was funded by Genentech Inc. Ganti has received research support from Takeda and has provided consulting services to Genentech Inc., AstraZeneca, Flagship Biosciences, Cardinal Health, BioGene, Mirati Therapeutics, Blueprint Medicines, and G1 Therapeutics. Lin, Wong, and Ogale are employees of Genentech Inc. and may own stock in F. Hoffmann-La Roche. Yang was employed by Genentech Inc. at the time of this study. Part of the study findings were presented as a poster at the NCCN 2020 Virtual Annual Conference, April 9, 2020.