MAGNIMS and McDonald Criteria Have Similar Accuracy

The 2016 Magnetic Resonance Imaging in Multiple Sclerosis (MAGNIMS) criteria have accuracy similar to that of the 2010 McDonald criteria in predicting the development of clinically definite multiple sclerosis (MS), according to a retrospective study.

Massimo Filippi, MD

“Among the different modifications proposed, our results support removal of the distinction between symptomatic and asymptomatic lesions, which simplifies the clinical use of MRI criteria, and suggest that further consideration be given to increasing the number of lesions needed to define periventricular involvement from one to three, because this might slightly increase specificity,” said Massimo Filippi, MD, of the neuroimaging research unit in the division of neuroscience at San Raffaele Scientific Institute at Vita-Salute San Raffaele University in Milan, and colleagues. The report was published online ahead of print December 21, 2017, in Lancet Neurology. “Further effort is still needed to improve cortical lesion assessment, and more studies should be done to evaluate the effect of including optic nerve assessment as an additional dissemination in space [DIS] criterion.”

An Effort to Draft Revisions

In an effort to guide revisions of MS diagnostic criteria, Dr. Filippi and other members of the MAGNIMS network compared the performance of the 2010 McDonald and 2016 MAGNIMS criteria for MS in a cohort of 368 patients with clinically isolated syndrome (CIS) who were screened between June 16, 1995, and January 27, 2017. They used a time-dependent receiver operating characteristic curve analysis to evaluate MRI criteria performance for DIS, dissemination in time (DIT), and DIS plus DIT. Changes to the DIS definition contained in the 2016 MAGNIMS criteria included removal of the distinction between symptomatic and asymptomatic lesions, increasing the number of lesions needed to define periventricular involvement to three, combining cortical and juxtacortical lesions, and inclusion of optic nerve evaluation. For DIT, removal of the distinction between symptomatic and asymptomatic lesions was suggested.

Dr. Filippi and his coauthors at eight centers reported that 189 (51%) of the 368 patients had developed clinically definite MS by the last evaluation, which occurred at a median of 50 months from baseline. At 36 months, DIS alone showed high sensitivity in the 2010 McDonald and 2016 MAGNIMS criteria (91% vs 93%, respectively), similar specificity (33% vs 32%), and similar area under the curve values (AUC, 0.62 vs 0.63). Inclusion of symptomatic lesions did not alter performance. The researchers also found that requiring three periventricular lesions reduced sensitivity to 85% and increased specificity to 40%, but did not affect AUC values (AUC, 0.63). When optic nerve evaluation was included, sensitivity was similar (92%), while specificity decreased to 26%, and AUC declined to 0.59.

The 2016 MAGNIMS and 2010 McDonald criteria achieved similar sensitivity, specificity, and AUC values when compared on the performance of DIT criteria and DIS plus DIT criteria.

“For both sets of criteria, specificity was lower than that of previous studies that evaluated the diagnostic performance of the 2010 McDonald criteria,” the authors wrote. “Several factors could help explain our findings, including the different follow-up durations, the statistical methods (eg, using a time-to-event analysis in our study), and the effect of treatment, which might have delayed or prevented the occurrence of the second attack during the study period.” They acknowledged certain limitations of the study, including its retrospective design and the fact that patients were recruited in highly specialized centers, which may have resulted in the selection of patients at higher risk of conversion to clinically definite MS.

MRI Abnormalities May Not Indicate MS

As MS diagnosis evolves, revisions to existing diagnostic criteria have increased sensitivity, thus helping clinicians establish earlier diagnoses. Although the study by Dr. Filippi et al showed that for both sets of MRI criteria, sensitivity was greater than specificity for predicting clinically definite MS, the modest specificity is cause for concern, said Anne H. Cross, MD, Head of the Neuroimmunology (MS) Section, and Robert N. Naismith, MD, Associate Professor of Neurology, both at Washington University in St. Louis, in an accompanying editorial. They cited one study that emphasized the importance of not misdiagnosing other CNS diseases as MS. “In that study at four academic medical centers, 110 people seen over a period of less than 1.5 years were found to have been misdiagnosed,” said Drs. Cross and Naismith. “[Seventy percent] of the 110 individuals had received disease-modifying therapy, and 31% had unnecessary morbidity. Leading factors contributing to erroneous diagnosis in the study included overreliance on MRI abnormalities in patients with nonspecific neurologic symptoms.”

Vascular and other diseases can cause MRI abnormalities that could meet the 2016 MAGNIMS recommendations or the 2010 and 2017 McDonald MRI criteria, the authors continued. For example, patients with monophasic inflammatory and infectious diseases might have gadolinium-enhancing lesions that meet the 2017 McDonald criteria for DIT, which require only the simultaneous presence of gadolinium-enhancing and gadolinium-negative lesions in the proper locations. For patients with an atypical presentation who meet the 2010 and 2017 McDonald or 2016 MAGNIMS recommendations, clinicians should weigh all of the observed imaging features (including the number of periventricular lesions, along with lesion size, shape, and location) to improve diagnostic specificity and help to limit misdiagnoses, they concluded.

—Doug Brunk

Suggested Reading

Cross AH, Naismith RT. Refining the use of MRI to predict multiple sclerosis. Lancet Neurol. 2017 Dec 21 [Epub ahead of print].

Filippi M, Preziosa P, Meani A, et al. Prediction of a multiple sclerosis diagnosis in patients with clinically isolated syndrome using the 2016 MAGNIMS and 2010 McDonald criteria: a retrospective study. Lancet Neurol. 2017 Dec 21 [Epub ahead of print].