Rationale and Objectives: The authors investigate the accuracy of computed tomography linear measurement of femora with titanium stem, and the effect of the stem on these measurements. Materials and Methods: Two embedded cadaveric femora, one of them containing a titanium stem, and two cortical bone parallelepipeds were scanned. Thirty-six cross-sections were studied, each measured in two linear directions by the profile window technique. A half-maximum height method was used to determine the cortex-gap-titanium sizes from the computed tomography (CT) images. The accuracy of the measurements from the parallelepipeds, femora, and titanium stem taken from the CT was compared with those taken by a digital caliper of anatomical sections at the same level of the same bone. Results: Computed tomography measurements of the parallelepipeds were similar to the anatomical size (mean relative error 0.04% +- 0.63%). The mean error and mean relative error of the cadaveric femora CT with and without the stem were similar to the control parallelepipeds. Higher values of error were found for the titanium stem. Conclusions: The half-maximum height method in the profile window provides an accurate measurement of the femoral cortex and the titanium stem. The presence of the titanium stem in the medullary cavity of the femur did not interfere with the measurements of cortical dimensions.