Abstract
Background/Objectives: This study investigates whether a patient's clinical status influences the evaluation of sacroiliac joint (SIJ) radiographs. Methods: The study involved analysing SIJ radiographs of patients diagnosed with spondyloarthritis (SpA) at our clinic. Two rheumatologists working at another centre evaluated the images independently. Three months were allowed to elapse so that the rheumatologists would forget the results of the first evaluation. The radiographs were then re-evaluated in a different order by the same rheumatologists. However, during the second evaluation, the evaluators were also provided with general clinical information about the patients. Inter- and intra-observer agreement were assessed. Results: In the first blinded evaluation of our study, we found moderate-to-substantial agreement between rheumatologists (right κ: 0.534; p < 0.001; left κ: 0.609; p < 0.001) and statistically significant interpretation agreement. In the second evaluation, we observed an increase in agreement (κ increased from 0.534 to 0.774 for the right SIJ and from 0.609 to 0.855 for the left SIJ), and these metrics were also significant. Conclusions: The interpretation of SIJ radiographs showed notable inter-observer variability in the absence of clinical information. Incorporating clinical context significantly improved the agreement between readers. As conventional radiography remains central to spondyloarthritis classification, these findings highlight the value of clinical data in enhancing the reliability of radiographic assessment. To our knowledge, this is the first study to systematically demonstrate the impact of clinical information on inter-observer agreement in SIJ radiograph interpretation.