Background: The accurate stratification of infants with congenital cytomegalovirus (CMV) infection at risk for more severe outcome may help in the management of patients. Aim of this study was to investigate the ability of a comprehensive neuroimaging investigation in predicting the long-term neurodevelopmental outcome in patients with congenital CMV. We analyzed the prognostic accuracy of a traditional score and a recently proposed scale applied to head ultrasound (HUS), computed tomography (CT) and magnetic resonance imaging (MRI).
Methods: All consecutive neonates born from 2002 to 2015 with congenital CMV infection were considered eligible for the study. Neuroimaging findings were scored according to both scores.
Results: One hundred seventy infants were included (112 symptomatic patients). One-hundred eighteen infants received both HUS, CT and MRI. CT and MRI were normal in all 56 asymptomatic patients, while 32% of them presented an abnormal HUS. The prevalence of abnormal findings differed according to the neuroimaging study. The sensitivity of the new neuroimaging score in detecting patients at risk for poor neurologic outcome was higher than the traditional one for all neuroimaging examinations. CT and MRI showed higher positive predictive value compared with HUS. No neuroimaging examination showed a negative predictive value equal to 100%.
Conclusions: Although HUS is the safest neuroimaging technique, it performs less well in detecting some brain abnormalities that can be associated with a poor neurodevelopmental outcome. A comprehensive neuroimaging evaluation is mandatory in infants with congenital CMV infection to decide for treatment and make a prognostic evaluation.