Patient-specific verification of intensity-modulated radiation therapy (IMRT) plans can be clone by dosimetric measurements or by independent dose or monitor unit calculations. The aim of this study was the clinical evaluation of IMRT verification based on a fast Monte Carlo (MC) program with regard to possible benefits compared to commonly used film dosimetry. 25 head-and-neck IMRT plans were recalculated by a pencil bean based treatment planning system (TPS) using an appropriate quality assurance (QA) phantom. All plans were verified both by film and diode dosimetry, and compared to MC simulations. The irradiated films, the results of diode measurements and the computed dose distributions were evaluated, and the data were compared on the basis of gamma maps and dose-difference histograms. Average deviations in the high-dose region between diode measurements and point dose calculations performed with the TPs and MC program. were 0.7 +/- 2.7% and 1.2 +/- 3.1%, respectively For film measurements, the mean gamma values with 3% dose difference and 3 MM. distance-to-agreement were 0.74 +/- 0.28 (TPS as reference) with dose deviations up to 10%. Corresponding values were significantly reduced to 0.34 +/- 0.09 for MC dose calculation. The total time needed for both verification procedures is comparable, however, by far less labor intensive in the case of MC simulations. The presented study showed that independent dose calculation verification of IMRT plans with a fist MC program has the potential to eclipse film dosimetry more and more in the near future. Thus, the linac-specific QA part will necessarily become more important. In combination with MC simulations and due to the simple set-up, point-dose measurements for dosimetric plausibility checks are recommended at least in the IMRT introduction phase.