In this study, 300 consecutive patients suspected of axSpA were enrolled. All had chronic back pain of more than 3 months’ duration. While highly experienced radiologists were asked to diagnose or rule out a diagnosis of axSpA on the basis of the MRI blinded to other clinical information, experienced rheumatologists evaluated the patients with access to all clinical, laboratory, and imaging data.
A diagnosis of axSpA was reached in 131 patients by the rheumatologists. The remaining 169 were determined not to have axSpA. Although the radiologists agreed on those with or without axSpA in 86.3% of cases, there were 31 cases (28.1%) in which rheumatologists diagnosed axSpA but radiologists did not.
In an analysis of which MRI features were considered critical by radiologists when there was agreement, they identified bone marrow edema in seven cases (7.2%). In 30 cases (30.9%), the radiologists considered the presence of chronic lesions to be critical to their diagnosis. In the remaining 69.9% of cases, radiologists were confident in their diagnosis only when both bone edema and chronic lesions were present.
Not surprisingly, the presence of chronic lesions and more pronounced bone marrow edema permitted both radiologists and rheumatologists to increase their confidence when discriminating between axSpA and non-axSpA patients.
“The combination of structural changes and bone marrow edema as assessed by MRI performed best in the process of diagnosing or ruling out axSpA in this real-life setting at our center,” Dr. Baraliakos said.
However, when only one or two features are considered, trade-offs of lower sensitivity for higher specificity or higher sensitivity for lower specificity occur. For example, although the specificity for a diagnosis of axSpA reached 99.4% when both bone marrow edema and ankylosis are present, the sensitivity of this finding was only 5.3%, according to data provided by Dr. Baraliakos. Conversely, the presence of sclerosis had a sensitivity of 81.7% but a specificity of only 43.2%.
One lesson from this analysis is that there is “increasing insecurity of only including bone marrow edema of the sacroiliac joint as the major criterion for diagnosing axSpA,” Dr. Baraliakos said. However, the larger point in the context of the earlier expert comments is that MRI findings should be considered important but insufficient for the evaluation of axSpA.
SOURCE: Baraliakos X et al. Ann Rheum Dis. Jun 2019;78(Suppl 2):255-6. Abstract OPO344, doi: 10.1136/annrheumdis-2019-eular.5027