Objectives: To compare nilotinib concentrations obtained by venous blood sampling and dried blood spot (DBS) in patients with chronic myeloid leukaemia (CML). It was investigated how to predict nilotinib plasma levels on the basis of DBS. Methods: Forty duplicate DBS and venous blood samples were collected from 20 patients. Capillary blood was obtained by finger prick and spotted on DMPK-C Whatman sampling paper, simultaneously with venous blood sampling. Plasma concentrations were predicted from DBS concentrations using three methods: (1) individual and (2) mean haematocrit correction and (3) the bias between plasma and DBS concentrations. Results were compared using Deming regression and Bland–Altman analysis. Key findings: Nilotinib plasma concentrations ranged from 376 to 2663 μg/l. DBS concentrations ranged from 144 to 1518 μg/l. The slope was 0.56 (95% CI, 0.51 to 0.61) with an intercept of −41.68 μg/l (95% CI, −93.78 to 10.42). Mean differences between calculated and measured plasma concentrations were −14.3% (method 1), −14.0% (method 2) and −0.6% (method 3); differences were within 20% of the mean in 73%, 85% and 80% of the samples, respectively. The slopes were respectively 0.96 (95% CI, 0.86 to 1.06), 0.95 (95% CI, 0.86 to 1.03) and 1.00 (95% CI, 0.91 to 1.09). Conclusions: Plasma concentrations of nilotinib could be predicted on the basis of DBS. DBS sampling to assess nilotinib concentrations in CML patients seems a suitable alternative for venous sampling.