MRI interpretation of suspected sacroiliitis varies greatly in children
There is a significant amount of variation and discordance among interpretations of inflammatory and structural lesions in sacroiliac joints among children, according to findings published in Arthritis Care & Research.
“This is important clinically because pediatric rheumatologists base therapeutic decisions on imaging results, among other factors,” Pamela F. Weiss, MD, MSCE, of the Children’s Hospital of Philadelphia, told Healio Rheumatology. “In our study, local radiology reports indicating active inflammation resulted in treatment changes by the rheumatologist for more than 75% of cases, and almost half of these were interpreted as not consistent with active inflammation by the central review team.”
“It is important to note, however, this study focused on assessment of the reliability of MRI interpretations, not the accuracy of MRI findings in terms of diagnosis and clinical outcomes,” she added.
To assess the agreement between local radiology reports and central imaging reviewers for active inflammation and structural damage at the sacroiliac joints among children, Weiss and colleagues conducted a multicenter study of hospital data across North America. In all, eight hospitals each contributed 20 cases of consecutively imaged children and adolescents with juvenile idiopathic arthritis and suspected sacroiliitis. Among these, 120 patients were ultimately evaluated in the final analysis.
Three experienced musculoskeletal pediatric radiologists independently reviewed all studies. Two of the researchers coded local assessments of global impression and lesions from the local radiology reports. They then calculated test properties of local reports using the central imaging team’s majority as the reference standard.
According to the researchers, sensitivity, at 93.5% (95% CI, 78.6% to 99.2%), and specificity, at 69.7% (95% CI, 59% to 79%), of local reports for inflammation were high and moderate, respectively. However, positive predictive value was low, at 51.8% (95% CI, 38% to 65.3%).
Twenty-three percent of cases had active inflammation reported locally but rated normal centrally, with 70% having subsequent medication changes. Further, sensitivity of local reports detecting structural damage, at 45.7% (95% CI, 28.8% to 63.4%), and positive predictive value, at 61.5% (95% CI, 40.6% to 79.8%), were low, while specificity was high, at 88.2% (95% CI, 79.4% to 94.2%).
“Most of the discordance was perceived to be error differentiating normal physiologic metaphyseal equivalent signal from pathologic subchondral marrow edema,” Weiss said. “Additional training for both radiologists and rheumatologists, including adult and pediatric providers, about the maturational changes on MRI in the maturing sacroiliac joint may improve the reliability of sacroiliac joint MRI interpretations. This would in turn increase the utility of MRI in management of pediatric spondyloarthritis.” – by Jason Laday
Disclosures: Weiss reports consulting fees from Lilly and Novartis.