Background: In head and neck radiotherapy, immobilization devices can affect dose delivery. In this study, a comprehensive end-to-end test was developed to evaluate the accuracy of radiotherapy treatment. Materials and Methods: An Alderson Radiation Therapy (ART) anthropomorphic phantom with EBT3 film was used to mimic the actual patient treatment process. Ten patients treated for nasopharyngeal carcinomas with IMRT were retrospectively selected. For each patient, the treatment plan, as well as the targets and OARs was transplanted onto the phantom, and the IMRT plan was subsequently recalculated to the phantom with EBT3 film. Two quality assurance (QA) plans were generated, namely “Plan-with” wherein the immobilization device was contoured and “Plan-without” wherein it was omitted. EBT3 measurements were compared with the results of the TPS calculation. Results: With different gamma calculation criteria applied, the results obtained for Plan-with were closer to the dose measured with the EBT3 film. Moreover, 1.8% deviation was observed in the posterior neck skin dose for Plan-with when compared to the film measurements while the value was 33.1% lower for Plan-without. When compared to Plan-without, each target volume in Plan-with exhibited a 1–4% reduction in the maximum dose (D_2%), minimum dose (D_98%) and mean dose (D_mean). Conclusion: Immobilization devices decrease the radiation dose to target volumes while increasing the skin dose and should be included within the body contour to ensure an accurate planning dose. The end-to-end IMRT test using an ART anthropomorphic phantom is a valuable tool to identify discrepancies between calculated and delivered radiation doses.