to distinguish cases with progression from cases with repair
based on specific features considered relevant to repair, such
as sclerosis, recortication, and filling-in of erosions. In that
study the experts demonstrated good agreement on which
image showed the least damage, but not on whether the best
image was the first or second in time; in other words, whether
the case was one of progression or repair. A few explanations
for this observation were possible. First, there were quite different levels of experience among the readers, raising the possibility that the readers were not sufficiently experienced to
recognize the features of repair. It was also possible that the
images used in that study did not have a sufficient number of
features of repair, or that the repair features were not clearly
defined for technical reasons. Third, only single joints were
presented to the readers, which might have hampered the correct ordering of the images into cases of progression or repair
as change in other joints was not available to help in the decision. Most importantly, we were still not informed about the
relation between cases depicted as repair by experts and negative scores obtained with traditional scoring methods such as
the modified Sharp method [2,14].