Résumé : The dose distribution resulting from partially overlapping intensity modulated beams (IMBs) assigned to different isocenters for the treatment of the same planning target volume (PTV) was evaluated. These partially overlapping IMBs are used in static intensity modulated radiation therapy (IMRT) treatments with the Novalis® system using the mini-MultiLeaf Collimator (mini-MLC) in Dynamic MultiLeaf Collimation (DMLC) mode. The resultant dose distribution was verified dosimetrically for a cylindrical target defined in a homogeneous cubic phantom. The phantom positioning can introduce dose nonuniformities in the resultant dose distribution by nonperfect positioning of the isocenters in accordance with each other. The dose inhomogeneities are quantified mathematically by summation of the dose profiles of the used IMBs and experimentally by measurement of the resulting dose profiles with radiographic film and thermoluminescent detectors (TLD). The mathematical estimation of the resulting dose profile of the treatment with a perfect positioning of the isocenters showed a good agreement with the planned dose profile. The magnitude of the maximum dose inhomogeneities introduced by the simulated supplementary shifts between the isocenters decreases by - 8.54% mm-1 as the shift changes from -0.30±0.10 cm to +0.30±0.10 cm. The TLD measurements showed a similar variation of the magnitude of the maximum dose inhomogeneities: - 8.77% mm-1. The amount of dose variation was underestimated with the radiographic film measurements, which showed a variation of - 7.17% mm-1. The film measurements demonstrated that the magnitude of the introduced maximum dose inhomogeneities did not alter significantly throughout the PTV. The approach of using partially overlapping IMBs assigned to different isocenters to enlarge the treatment region introduces smaller dose inhomogeneities in the resultant dose distribution than when abutting treatment fields are used. The resultant dose distribution of this treatment technique is less sensitive to positioning errors of the used treatment isocenters. © 2003 American Association of Physicists in Medicine.