Background: Imatinib is a competitive inhibitor of protein tyrosine kinase Bcr-Abl and is currently used for the treatment of chronic myeloid leukaemia (CML) and other digestive malignant pathologies. Trough imatinib plasma levels are associated with major molecular response in CML. Due to important interpatient variability, monitoring of imatinib plasma levels can be very useful especially in the case of treatment failure. The aim of this work was to determine intra-individual variability in imatinib plasma concentrations to a better knowledge of therapeutic drug monitoring (TDM) of imatinib in the real life. Methods: Imatinib plasma concentrations were determined by highperformance liquid chromatography with UV detection at 262nm after liquidsolid extraction. Limit of quantification was set at 200 ng/mL. Blood samples were collected at steady-state (trough values before drug administration). Samples were collected and analysed over a period of 3 years with an interval of at least 3 months between each sample. Dosage regimens ranged from 200mg/day to 600mg/day. Results were expressed as mean ± standard deviation and variability in plasma concentrations were presented as a CV expressed in percentage. A minimum of three measurements was needed to calculate CV%. Results: 21 patients were evaluated with a mean number of 5 measurements per patients (3−10) and a total of 108 samples. Sex-ratio (M/F) was 0.48, mean age and weight were respectively 55±16 years and 76±19 kg. Concerning the 400mg/day group of patients (61 samples), mean imatinib plasma concentration was 1130±640 ng/mL, ranging from 250 to 2800 ng/mL(total CV of 57%) with 18 subtherapeutic concentrations (a plasma threshold of 1000 ng/mL is associated with major molecular response). For all dosage regimens (21 patients), mean intra-individual variability of imatinib was 31%, ranging from 9.7% to 70%. Possible causes for pharmacokinetic variability are multiple: drug/drug interactions (due to metabolism of imatinib with CYP3A4 and transport by Pgp), poor compliance, genetic polymorphisms. A value of 31% for intra-individual variability can be considered low. Conclusion: The high interindividual and relatively low values of intrapatient variability in plasma support therapeutic drug monitoring. Nevertheless mean plasma concentrations were often closed to therapeutic threshold and our data suggest the need of regular TDM measurement.