Varied interpretation of MRI reports by neurologists may lead to differences in MS care

SAN DIEGO — Variations in satisfaction and confidence in MRI reports may lead to significant variability in the clinical decisions made by neurologists specializing in MS care, according to a poster from ACTRIMS 2023.

“The goal of this study is to understand the impact of MRI on clinical decisions,” Kelly M. Leyden, senior product lead of MRI insights at Octave Bioscience, told Healio in an interview. “We wanted to understand the underlying variability of neuroradiology reports and how different neurologists interpret those results in their clinical practice.”

Researchers conducted a retrospective study of 90 individuals with MS from the University Hospital Basel in Switzerland, who had two MRIs taken a median of 1 year apart. Three neuroradiologists were asked to interpret MRI results and provide standard-of-care reports for each. The reports were randomized so each neurologist received 90 individual reports from the three neuroradiologists.

Leyden and colleagues assessed intra-rater variability by repeating 30 cases within each set to generate a total of 120 reports. They determined inter-rater variability by randomly assigning six neurologists to three groups which received matching reports.

After reviewing the reports and accompanying clinical summaries, the neurologists completed a survey, which rated the quality of the MRI report and clinical impressions and decision-making based on the patient’s disease status, level of physician concern and treatment plan. Results from five neurologists were complete at analysis.

Researchers reported that intra-rater agreement of satisfaction with MRI reports ranged from slight to substantial (50%-96.7%, K = –0.06 to 0.77), while agreement on clinical decision-making varied from slight to near perfect (50-100%, K = 0.01-1).

Data also demonstrated variable inter-rater reliability between neurologist pairs for MRI report satisfaction (slight to moderate; 7.8%-66.7%; K = 0.01-0.48).

Further, inter-rater agreement for clinical decision-making varied from slight to substantial between pairs. For example, one pair had fair agreement regarding “neurologist’s level of concern” (30%; K = 0.25), while another had moderate agreement (58.9%; K = 0.56).

“There is a lot of variability in neurologists’ interpretation of these reports, and what this ultimately leads to is potentially different actions taken by different neurologists in treatment of patients with MS,” Leyden said.