Purpose or Case Report: Hypoxic ischemic injury (HII) is a leading cause of neonatal encephalopathy and resultant cerebral palsy. Most studies focus on early magnetic resonance imaging (MRI) findings and patterns of injury, with few studies evaluating patterns in delayed MRI of patients with cerebral palsy.

We aim to categorise the delayed imaging MRI findings and distribution of abnormalities in a large cohort of children with cerebral palsy and obtained for medicolegal purposes to determine the causes of cerebral palsy, and to map the frequency of anatomic involvement in those with hypoxic ischemic injury.
Methods & Materials: 252 MRI studies were read by 3 radiologists who were blinded to each other and to clinical history. Specific focus was made on areas typically involved in HII with specific imaging patterns, and signs elicited. The responses were then collated and analysed for frequencies and distributions of areas of involvement with linear regression and co-variate analysis to determine interrelationships.
Results: Average results were obtained for Multicystic encephalomalacia (11,3%), Focal cysts (14%) and Ulegyria (37,2%). Regional involvement was of the Peri-Rolandic (43,7%), Corticospinal tracts (68%), posterior limb of the internal capsules (61,8%), Peri-Sylvian (29,8%), Para-Falx (17,1%), Occipital (49%), Anterior inter-vascular watershed (IVWS) (35,2%), Posterior IVWS (40,3%) & Corpus callosum (43%).

Pituitary bright spot was absent in 9,3%. Ventriculomegaly in 61,9% of patients. Thalamic (78,5%), Putamina (40,7%), Pallidus & Caudate (20,2%), Hippocampal (47,5%), Superior vermis of cerebellum (15%), Inferior vermis (6,7%), Right cerebellar hemisphere (8%) and Left cerebellar hemisphere (8%) involvement was also determined.

Final diagnoses were multicystic encephalomalacia (11,3%), term acute profound (32.3%), term partial prolonged (10%), term combined (35,7%), premature periventricular leukomalacia (3,3%), hypoglycemic injury (29,7%) and other (10%). Inter-reader concordance was high with a Fleiss kappa of 0,55 overall and with higher scores for individual selections/ data points. Linear regression and covariate analysis has been performed to analyse inter-relationships and dependencies.
Conclusions: We have demonstrated the frequency of pathology and signs in various areas of the brain in children with cerebral palsy and determined the patterns of injury with frequency of final diagnoses (especially with regards to timing of injury) as well as inter-relationships between the various patterns of abnormality.