MRI Tool Helps to Predict Conversion to Pediatric MS



AMSTERDAM – An MRI scoring tool helped to identify children with acute central nervous system demyelination who later developed pediatric multiple sclerosis in a prospective study of 284 children.

The tool found that the risk of an MS diagnosis dramatically increased in these children when a baseline MRI showed the presence of one or more T1-weighted hypointense lesions, as well as when scans revealed T2-weighted periventricular lesions.

"Several studies have described the MRI characteristics of pediatric patients with established MS, and how they compare to that of adults matched for disease duration," said Leonard Verhey, a doctoral student working with Dr. Brenda Banwell at the Hospital for Sick Children at the University of Toronto.

"However, less is known about the MRI features that predict MS in children with acute demyelination, and how these features might compare with the predictive criteria for adult-onset MS."

Mr. Verhey and his associates used a standardized 14-item MRI scoring tool to identify MRI parameters that might predict MS in an unselected population of children with acute CNS demyelination. The tool was originally developed based on data from 61 children with relapsing-remitting MS (RRMS), monophasic acute disseminated encephalomyelitis, and non-demyelinating CNS inflammation, none of whom were included in the current prospective study.

The investigators recruited children younger than 16 years for the current study at 23 centers in Canada. During a 5-year follow-up period, they obtained MRI scans at baseline, 3 months, 6 months, and 1 year. The study’s funding did not allow serial MRIs to be taken between years 2 and 5 of follow-up, but the children were assessed clinically every year until the end of the study, and some had additional MRIs taken if required.

"Less is known about the MRI features that predict MS in children with acute demyelination."

Two experts, blinded to the clinical findings, scored all MRI scans with the standardized 14-item tool. So far, 57 (20%) children have been given a confirmed diagnosis of MS, while the remainder (n = 227) have monophasic acute acquired demyelination (ADS). Those with MS had been followed for a mean of 4.3 years, while those with monophasic ADS had been observed for a mean of 3.9 years, Mr. Verhey reported at the joint triennial congress of the European and Americas Committees for Treatment and Research in Multiple Sclerosis (ECTRIMS/ACTRIMS).

Patients with pediatric MS were significantly more likely than were those with monophasic ADS to have an older age of onset (12.8 vs. 8.8 years) and to be female (67% vs. 48%).

The most common abnormality seen on baseline MRI was the presence of T2-weighted lesions (about 60% of the entire cohort). Other lesions affected the brainstem, periventricular area, or cerebral white matter in about 40% of patients. T1 hypointense lesions also were seen in about 30% of scans.

A univariate analysis identified nine MRI characteristics that were indicative of an MS diagnosis. "T1 hypointense lesions showed the strongest association with MS outcome," noted Mr. Verhey, who won one of the two best oral presentation awards given to young investigators at the congress.

Periventricular lesions also were strongly associated with outcome, and "interestingly, although only 22% of the patients had contrast-enhancing lesions, this was still associated with a 10-fold increased risk of MS diagnosis," he said. Another interesting finding was that the presence of thalamic lesions appeared to confer a 61% decreased risk of MS diagnosis.

A multivariate analysis that adjusted for age and gender revealed only two significant and independent predictors of an MS diagnosis: the presence of one or more T1 hypointense lesions (hazard ratio, 20.3), which were found in scans of 80 (28%) patients; and one or more T2 periventricular lesions, which occurred in scans of 112 (39%) patients (HR, 3.3). If both of these parameters were present at baseline, the risk of MS diagnosis was greatly increased (HR, 34.3).

Mr. Verhey and his colleagues found that those two predictive factors had a more favorable combination of sensitivity (84%), specificity (93%), positive-predictive value (76%), and negative-predictive value (96%) than did other predictive models in published studies.

Another 100 children will be enrolled in a validation cohort. Other centers may be able to validate the results because of the tool’s high inter-rater reliability and its accompanying MRI atlas, he said.

Mr. Verhey noted that the study has been accepted for publication in the Lancet Neurology.

The study was supported by a grant from the Multiple Sclerosis Scientific Research Foundation. Additional support was provided by the Multiple Sclerosis Society of Canada, the Canadian Institutes of Health Research, and the Hospital for Sick Children. None of the investigators had relevant disclosures.

Next Article: