Introduction CrUS is an important prognostic variable in prematurity, although with some limitations. Quantitative CrUS techniques might have the potential to overcome some of those limitations.
Objectives To investigate whether a semi-quantitative method of CrUS analysis correlated with developmental outcome at 2 years in a cohort of preterm infants, and to compare this method with clinical variables and CrUS qualitative evaluation.
Methods Cohort of 88 <33 weeks gestational aged (GA) infants underwent several CrUS scans from birth to term. This last scan was analysed using a quantitative texture method (LBR), eventually obtaining 3 Quantitative Scores (QS). Outcome variables included abnormal Bayley Scales of Infant Development (3rd edition): Motor, Cognitive, and Language composite scores (<1SD); and Any Developmental Impairment (any abnormal previous result or a vision or hearing impairment).
Results Mean GA was 29,0 (SD 2,2). Hearing and vision deficits were present in 3 patients. Abnormal scores occurred 9.1% in cognitive, 11.4% motor, 18.2% language and 23.9% NDI. QS significantly correlated to all outcome variables (p values: cognitive = 0.003, motor = 0.011, language = 0.015, NDI = 0.036). ROC analysis is shown in table. When clinical information and CrUS abnormalities were added in logistic regression analysis, QS added significant information (R2 Nagelkerke) in all but motor outcome (Figure).