Purpose Anaplastic lymphoma kinase (ALK) rearrangement–positive non–small-cell lung cancers can be effectively treated with an ALK tyrosine kinase inhibitor (TKI) such as crizotinib, but the response magnitude and duration are heterogeneous. Several ALK variants have been identified, but few studies have focused on the effects of different ALK variants on the efficacy of crizotinib. Patients and Methods Among 55 patients treated with crizotinib as the initial ALK-TKI between January 2007 and December 2014, we identified 35 patients with tumor specimens that could be evaluated for ALK variants by reverse transcription polymerase chain reaction. We retrospectively evaluated the efficacy of crizotinib on the basis of the objective response rate and progression-free survival (PFS) according to the ALK variants. Results The most frequent ALK variant was variant 1 in 19 patients (54%), followed by variant 2 in five patients (14%), variant 3a/3b in four patients (12%), and other variants in seven patients (20%). Objective response rate was 69% in all patients, whereas it was 74% and 63% in the variant 1 and non–variant 1 groups, respectively. The median PFS time was significantly longer in patients with variant 1 than in those with non–variant 1 (median PFS, 11.0 months [95% CI, 6.5 to 43.0 months] v 4.2 months [95% CI, 1.6 to 10.2 months], respectively; P < .05). Multivariable analysis identified two significant factors associated with PFS duration, ALK variant 1 (hazard ratio, 0.350; 95% CI, 0.128 to 0.929; P < .05) and advanced stage (hazard ratio, 4.646; 95% CI, 1.381 to 21.750; P < .05). Conclusion Our results indicate the better efficacy of crizotinib in patients with ALK variant 1 versus non–variant 1. The ALK variant status might affect the efficacy of ALK-TKIs.