A - 110 Cognitive Phenotypes in Children with a History of Brain Tumors.

Abstract

Research indicates children with a history of pediatric brain tumors (PBTs) are at increased risk for cognitive impairment. Cognitive phenotyping could reveal latent cognitive profiles and ascertain the impact of clinical and sociodemographic variables in this population. There is only one such study to date. We hypothesized patients with more treatment modalities, additional medical diagnoses, and diagnosis age/time since treatment completion would impact performance. K-means clustering was used to explore 17 test scores across 9 cognitive domains in 67 children with brain tumors aged 8-17years (median age = 12; 41 male; 85.1% Caucasian) in a pediatric hospital. Exclusion criteria included failed performance validity testing, IQs < 50, and active treatment. Parametric and non-parametric tests were utilized, and effect sizes were calculated. Two clusters were identified. Cluster 2 (n = 37) had lower scores than Cluster 1 (n = 30) on all neuropsychological domains assessed except for memory and mental flexibility with large effect sizes (p's < 0.005, n2 = 0.151-0.57). Clusters differed by tumor location (p = 0.021, V = 0.340), hydrocephalus (p = 0.041, V = 0.250), and treatment type (p = 0.037, V = 0.447). Results suggest multi-domain cognitive impairment in a large subgroup of children with PBTs associated with infratentorial tumors, history of hydrocephalus, and exposure to multiple treatment modalities (i.e., surgery, radiation, and chemotherapy). Other demographic, treatment (i.e., diagnosis age, time since diagnosis/treatment completion), and medical (i.e., tumor grade, seizures, genetic diagnoses) factors did not account for clusters. Future studies should include larger samples, increased age range, and additional clinical, sociodemographic, and psychosocial variables to form more generalizable models of cognitive phenotypes in youth with PBTs.