Measurements of cortical thickness in experimentally created endosteal bone lesions: a comparison of radiography, CT, MR imaging, and anatomic sections.

OBJECTIVE The purpose of this study was to evaluate the accuracy of cortical measurements of experimentally created endosteal cortical lesions and to assess the sensitivity of radiography, CT, and MR imaging in the detection and measurement of such lesions. MATERIALS AND METHODS Thirty-six cortical lesions were created in three fresh cadaveric femurs. After performing radiography, CT, and MR imaging, we sectioned the specimens in the axial plane. We then measured the remaining cortex at the lesions and the normal cortex adjacent to the lesions on all images and corresponding anatomic sections. The measurements of thickness of the cortex as seen with the different imaging methods and the anatomic sections were compared. Measurements were repeated to evaluate the influence of different window settings on the MR imaging measurements. RESULTS When measured on radiographs, cortical thickness was overestimated in 58% of lesions. With CT, cortical thickness was overestimated by 0-15% in 94% of all lesions. With MR imaging, cortical thickness was uniformly underestimated by 3-17%. Measurements made on MR images varied according to different window settings. The proton density-weighted sequence yielded the highest sensitivity in the detection of shallow cortical lesions; the T1-weighted spin-echo sequence was the least sensitive of the MR sequences. CONCLUSION In our cadaveric study, cortical thickness in the presence of endosteal lesions was overestimated on radiographs and CT scans and underestimated on MR images. Measurements derived from MR imaging are strongly influenced by the window setting. MR imaging with the proton density-weighted sequence is the most sensitive for detection of shallow cortical lesions and is more sensitive than CT.