To the Editor:

The observation of breaks in the juxtaarticular cortical bone of healthy individuals in our study is indeed an unexpected and interesting finding, and warrants further investigation. Rothschild suggests that such erosions may be false-positive lesions, which are based on “averaging” artifacts of CT scans. However, this is very unlikely. So-called “averaging” artifacts are in fact partial volume artifacts, which arise from limitations in the spatial resolution of the CT scan (1). This effect can, for instance, lead to failure to detect small lesions (false-negative result) or to a blurred measurement of trabecular bone thickness. However, existing lesions, such as cortical breaks, in a 2-dimensional (2-D) scan cannot be based on partial volume artifacts. In our study, cortical breaks were assessed on 2-D sections.

This is in contrast to cortical irregularities, which do not represent cortical breaks (erosions) and can only be found in 3-D reconstructions. The size of these latter lesions is dependent upon the threshold used. The necessity of using identical threshold is illustrated in Figure 2 of our report; however, this only applies to the detection of cortical irregularities in 3-D scans and not bone erosions in 2-D scans. We can, therefore, rule out imaging artifacts as the basis for bone erosions. Cortical breaks in healthy individuals are small and always <2 mm in size. It should be noted that the CT technique allows detection of cortical breaks as small as 200 μm. Such small lesions may easily be overlooked during a macroscopic examination of bones from anatomic collections. Additionally, it should be mentioned that erosive lesions can also be detected on magnetic resonance imaging of healthy individuals (2, 3). We can confirm Dr. Rothschild's assumption that there is no inflammatory trigger of bone erosions in the healthy individuals in our cohort, since we did rule out inflammatory arthritis. However, such lesions may not require inflammation, but may instead emerge from other triggers such as trauma, microdamage, and mechanical stress from inserting tendons and ligaments.

Georg Schett MD*, * University of Erlangen–Nuremberg, Erlangen, Germany.