AimTo investigate the outcomes of cabozantinib in patients with unresectable hepatocellular carcinoma (uHCC), focusing on dose setting and modification.MethodsWe retrospectively analyzed 34 Japanese patients who received cabozantinib for uHCC. Trough concentrations (Ctrough) of cabozantinib were also measured weekly for 6 weeks in the 18 patients.ResultsSixteen patients received ≥40 mg (high‐dose group), and 18 patients received 20 mg (low‐dose group). Dose escalations were performed in 27.8% of the patients in the low‐dose group. Although median duration of the first dose reduction or interruption in the low‐dose group was twice that in the high‐dose group (28 vs. 14 days, p < 0.001), there were no significant differences in the relative dose intensity (RDI) during 6 weeks, progression free survival (PFS), and overall survival (p = 0.162, p = 0.950, p = 0.817, respectively) between the two groups. Patients who received RDI during 6 weeks ≥33.4% showed a trend toward longer median PFS (p = 0.054). Each serum aldolase value during the 6 weeks was significantly correlated with the Ctrough at any point (r = 0.500, p < 0.001). In multivariate analyses, aldolase ≥8.7 U/L within 2 weeks was significantly associated with the very early dose reduction or interruption (odds ratio 20.0, p = 0.002).ConclusionsAn initial dose of 20 mg cabozantinib could be a safe option in Japanese patients. The serum aldolase value could be useful for making appropriate dose modifications of cabozantinib.