The main clinical and MRI features driving therapeutic choices are not as clear for pediatric multiple sclerosis (MS) patients as for adults.
We aimed at assessing early predictors of long-term clinically-relevant outcomes in a large cohort of pediatric MS patients.
Clinical and MRI assessment was obtained at disease onset and after 1, 2 and 3 years, in a cohort of 123 pediatric MS patients. The longest clinical follow-up (mean 9.33 +/- 3.45 years) was considered for long-term outcomes. Cox proportional hazards models were used to assess predictors of time to first relapse, while multivariable logistic and linear regression models identified clinical and MRI predictors of long-term outcomes.
Across baseline features, optic nerve involvement predicted a shorter time to first relapse (hazard ratio=1.9, p=0.03). Predictors of annualized relapse rate (ARR) were: at baseline, presence of cerebellar (b=-0.16, p=0.00) and number of cervical cord lesions (b=0.14, p=0.01); considering short-term predictors, the same baseline variables together with time to first relapse (2-year: b=-0.12, p=0.01; 3-year: b=-0.08, p=0.00) and the number of relapses (1-year: b=0.14, p=0.00; 2-year: b=0.06, p=0.02). Baseline predictors of 10-year disability worsening were: at baseline, presence of optic nerve [odds ratio(OR)=6.45, p=0.01] and brainstem lesions (OR=6.17, p=0.04); considering short-term predictors, Expanded Disability Status Scale (EDSS) changes at 1 (OR=26.05, p=0.00) and 2 (OR= 16.38, p=0.02) years and the detection of at least two new T2-lesions in 2 years (2-year: OR=4.91, p=0.02; 3-year: OR=5.49, p=0.09). Predictors of higher 10-year EDSS score were: at baseline, EDSS score (b=0.58, p<0.001), presence of brainstem (b=0.31, p=0.04) and number of cervical cord lesions (b=0.22, p=0.05); considering short-term predictors, EDSS changes (1-year: b=0.82, p<0.001; 2-year: b=0.79, p<0.001, 3-year: b=0.27, p=0.04 ), together with the detection of at least two new T2-lesions at 1 (b=0.28, p=0.03) and 2 (b=0.35, p=0.01) years.
In conclusion, baseline spinal cord, brainstem and optic nerve lesions have a major role in predicting long-term outcomes, both in term of disease activity and of disability worsening. In addition, an accurate clinical and MRI monitoring during the first 2 years of disease has proven to represent a powerful tool for counseling patients about long-term prognosis and personalizing treatment strategies.