Rheumatoid arthritis (RA) is a chronic inflammatory disease with symmetrical polyarthritis as the major feature. Persistent inflammation leads to largely irreversible joint damage which can be seen radiographically. Radiographs depict the progression of joint damage and alterations, which are one of the major parameters of RA evolution. Delayed radiologic progression is a good indicator of the success/failure of long-term drug treatment, but the quantitative analysis of changes over time and the reliability of scoring systems remain difficult steps. This paper focuses of the main current scoring systems, with an emphasis on the following four: Larsen method and its modification by Kayle, Sharp method and its modification by van der Heijde. The scoring method--be it grading, counting, or weighted counting--did not appear to influence reliability or repeatability, while the radiologist's training and the film reading technique were critical to identify disease progression accurately. Our data consistently suggest the paired reading method as the most suitable for radiologic progression assessment in RA. The radiologic studies of 62 patients with early RA after 7 years' follow-up showed joint damage in 82% of patients. The average annual progression rate of the total radiologic score obtained with Sharp method, summing erosions and joint space abnormalities, was faster in the earlier years of the disease (8.8 units/year) than later on (4.9 units/year) (p < 0.01). From the trialist's point of view, these results imply that disease duration is a critical feature for RA treatment outcome.