IntroductionIn adult patients with epilepsy, predictive models have been developed and validated for anticipating a favorable response to immunotherapy. However, no such model has been evaluated in children. MethodsThis retrospective cohort study intended to assess the performance of a pediatric adaptation of the Response to Immunotherapy in Epilepsy (RITE2) score: P-RITE2 score and Antibody Prevalence in Epilepsy (APE2) score: P-APE2 score in patients aged 1–18 years. We included data of those patients who had epilepsy duration of not more than 12 months, no other known etiology (e.g., genetic, metabolic, neoplastic, or structural causes), and tested for neural-specific antibody in cerebrospinal fluid or serum for P-APE2 score and only those who received immunotherapy for P-RITE2 score. We added cognitive dysfunction, speech dysfunction, sleep disturbance, and movement disorder to the original scores to increase specificity for pediatric autoimmune epilepsy. We assumed at least a 50% reduction in seizure frequency at 6 months as a favorable response to immunotherapy. Cut-offs were chosen for both scores to maximize true positives and minimize false negatives using ROC curves. ResultsWe included data from a total of 237 patients with epilepsy (10.4 ± 2.5 years, 129 boys, 54%), out of which, 25 (10.5%, 13 girls, 52%) tested positive for autoantibodies. The median P-APE2 score in the subgroup with and without antibody positivity were 7 (IQR: 5–11) and 2 (IQR: 1–5), respectively (p<0.0001). ROC analysis of the P-APE2 score determined an AUC of 0.96. The sensitivity and specificity values of the P-APE2 score ≥6 were 94% and 92%, respectively. A total of 162 patients (10.3 ± 2.5 years, 88 boys, 54%) received immunotherapy, out of which, 101 had a favorable response at 6 months. The median P-RITE2 score in the subgroup with and without favorable response following a trial of immunotherapy were 10 (IQR: 6–17) and 3 (IQR: 1–6), respectively (p<0.0001). ROC analysis of the P-RITE2 score determined an AUC of 0.96. The sensitivity and specificity values of P-RITE2 score ≥8 were 95% and 93%, respectively. The AUC of both these ROCs was significantly higher than the AUC of ROCs for original scores in our cohort. ConclusionThe P-RITE2 and P-APE2 scores can be used to predict the response to immunotherapy and predict autoantibody positivity in children with epilepsy with/without encephalopathy or cognitive dysfunction.