Visual-semiquantitative and quantitative magnetic resonance imaging (MRI) data may predict outcome in congenital cytomegalovirus infection

Abstract

AIM

Develop a combined magnetic resonance imaging (MRI) scoring system to categorise brain abnormalities in symptomatic newborns with congenital cytomegalovirus (cCMV) infection and evaluate its usefulness in predicting long-term neurological outcomes.

MATERIALS AND METHODS

cCMV infection onset during pregnancy and clinical severity at birth were assessed; the overall 3-year clinical outcome for language, motor, and cognitive impairments was determined based on domain scores.

We developed an MRI score combining visual-semiquantitative findings with volumetric data from brain regions to assess overall brain damage, including cortex abnormalities, white matter changes, calcifications, hippocampal dysplasia, brain volume reductions, and increased ventricular volume. The final cumulative score was the sum of all regional subscores, with higher scores indicating more severe damage.

RESULTS

Twenty-eight symptomatic cCMV newborns from cytomegalovirus (CMV)–infected pregnant women were studied, with 21 infected in the first two trimesters (early infection) and 7 in the third trimester (late infection). At diagnosis, 21.4% had mild, 7.1% had moderate, and 71.4% had severe symptoms.

Long-term neurological outcomes were observed in 82.1% of the children. An association between clinical severity of cCMV and infection timing (early/late) was found (P<0.001). The MRI score correlated with clinical outcomes (P=0.047; r=0.378) and identified infants at a high risk for severe impairment (P=0.033, odds ratio: 1.488) with a cut-off score of 4 distinguishing between higher and lower impairment levels (Area Under the Curve (AUC): 0.753, P=0.006).

CONCLUSION

The combined MRI score provides a reproducible tool to predict early neurodevelopmental impairments in cCMV patients, aiding clinicians in management and counselling.