Background and purpose: To evaluate the impact of deformation magnitude and image modality on deformable-image-registration (DIR) accuracy using Halcyon megavoltage cone beam CT images (MVCBCT).Materials and methods: Planning CT images of an anthropomorphic Head phantom were aligned rigidly with MVCBCT and re-sampled to achieve the same resolution, denoted as pCT. MVCBCT was warped with twenty simulated pre-known virtual deformation fields (T-i, i = 1-20) with increasing deformation magnitudes, yielding warped CBCT (wCBCT). The pCT and MVCBCT were registered to wCBCT respectively (Multi-modality and Unimodality DIR), generating deformation vector fields V-i and V-i' (i = 1-20). V-i and V-i' were compared with Ti respectively to assess the DIR accuracy geometrically. In addition, V-i, T-i, and V-i' were applied to pCT, generating deformed CT (dCT(i)), ground-truth CT (G(i)) and deformed CT' (dCT(i)') respectively. The Hounsfield Unit (HU) on these virtual CT images were also compared.Results: The mean errors of vector displacement increased with the deformation magnitude. For deformation magnitudes between 2.82 mm and 7.71 mm, the errors of uni-modality DIR were 1.16 mm similar to 1.73 mm smaller than that of multi-modality (p = 0.0001, Wilcoxon signed rank test). DIR could reduce the maximum signed and absolute HU deviations from 70.8 HU to 11.4 HU and 208 HU to 46.2 HU respectively.Conclusions: As deformation magnitude increases, DIR accuracy continues to deteriorate and uni-modality DIR consistently outperformed multi-modality DIR. DIR-based adaptive radiotherapy utilizing the noisy MVCBCT images is only conditionally applicable with caution.