PurposeTo evaluate the dosimetric errors associated with the effect of the collimator angle error in volumetric-modulated arc therapy (VMAT) treatments.Methods and materialsFour patients with different planning target volume (PTV) and localisations treated using VMAT were analysed (high-risk prostate, low-risk prostate, head and neck (H&N) and holocranial with hippocampus protection) in terms of dosimetric variations when errors in the collimator angle were introduced. Original plans underwent modifications of the planned collimator angles of ±0·5°, ±1° and ±1·5°. These modified plans were re-calculated using the same original plan fluencies, and the resulting dose–volume histograms and homogeneity index (HI-ICRU) were compared.ResultsFor the high-risk prostate case, there was a noticeable loss of PTV dose coverage for collimator angle errors larger than ±1°, with HI-ICRU relative variations up to 75% in the range analysed. The low-risk prostate case did not present significant changes in organs at risk or PTV dose coverage. For the H&N case, the spinal cord presented changes around 4% forD0·1 cc. In the holocranial case, optic lens showed dose variations up to 5% for collimator angle errors larger than ±1°.ConclusionsThe effect of the collimator error in VMAT increased as the PTV increased.For selecting the position of the isocentre, one should be cautious, and whenever possible choose a position close to the geometrical centre of the PTVs in order to avoid or minimise errors from the calibration of the collimator angle.