Kim Annink

77 Cerebral ultrasound scoring system for neonates with HIE Furthermore, the additional value of day one CUS to diagnose antenatally acquired pathology of the brain was determined. Neurodevelopmental outcome We retrospectively collected clinical parameters of all infants. Neurodevelopmental follow-up was performed with the Bayley Scales of Infant and Toddler Development, third edition (BSITD-III) at the age of two years (29) by a neonatologist and an educational therapist or child psychologist. Adverse outcome was defined as death, cerebral palsy or a cognitive/motor composite score <85 according to the BSITD-III (United States of America norms) at two years of age. Validation: inter-observer variability A neonatologist (JD) and a pediatric neuroradiologist (ML), both with an expertise in CUS, scored the day one and day three to seven CUS of cohort I independently of each other (observer one and observer three). Observer one and three did not work in the UMC Utrecht in the period that the CUS were conducted, so they were completely blinded to outcome. Another neonatologist (DV) without a special focus on CUS scored the CUS of 20 randomly selected patients to determine the inter- observer agreement in daily clinical practice (observer two). Validation of the scoring system in cohort II The CUS of cohort II were scored by two neonatologists with more than 25 years of experience in reading cerebral ultrasound scans (FG and LdV). They scored the CUS together and reached consensus about the CUS score of all infants. The observers did not work in the Erasmus Medical Center at the moment the CUS were performed, so they were completely blinded to outcome. Validation: correlation with MRI and histology Secondary outcomes were the correlation with MRI and histological findings. A correlation between the CUS scoring system and MRI was assessed using the MRI scoring system of Weeke et al. (13) in cohort I. In cohort II the diffusion weighted sequences were often of suboptimal quality; therefore, the secondary outcomes were only analyzed for cohort I. 4