To investigate dosimetric consequences of patient metal prosthesis in prostate radiotherapy, considering the body heterogeneity, metal artifacts, prostate daily displacements and target delineation. Three prostate patients with metal prosthesis were simulated with CT images and 3D ultrasound images. Conformal RT treatment plans were generated based on target volumes delineated on CT images without artifact correction (PTVa). The patients' treatments were aided with an ultrasound (US) localization system for daily PTV setup correction. We retrospectively reevaluated the delineation of PTVs by 1) using simulation CT images artifact-corrected by using an algorithm developed by M. Bazalova et al (2007) (PTVc) and 2) using the 3D ultrasound simulation images as guide (PTVus). Daily setup corrections to the PTVs were incorporated to calculate composite delivered dose by using XVMC simulation on the patient phantom derived from the artifact-corrected CT images. DVHs and dose distributions for different PTVs were then compared with the reference treatment plans (XVMC-calculated on the artifact-degraded CT images). The PTVa volume was the largest, about 1.3% larger than artifact-corrected PTVc and 5.2% larger than PTVus. Adapting artifact-corrected CT images can improve the DVH curves of PTVc and increase the D95% and V95% for PTVc by more than 5% while D50% and V50% for rectum and bladder are raised by up to 41.6%. DVH analysis on PTVa and PTVus shows a small difference in the changes of their DVH indices, less than 4% for the studied cases. The strike artifacts from metal prosthesis will increase the volume of PTVa, and affect the patient dose calculation. Original patient plan did not accurately predict the dose degradation. Artifact correction may be necessary for some cases having severe metal artifacts. Using US images to help delineating PTV makes a negligible clinical significance.