A new study published in JAMA Pediatrics revealed that use of fetal magnetic resonance imaging (MRI) and machine learning techniques can predict the need for postnatal cerebrospinal fluid (CSF) diversion among patients with fetal ventriculomegaly.

A total of 253 patients diagnosed with fetal ventriculomegaly from January 2008 through December 2014 were identified and recruited from an institutional database from the Children's Hospital of Philadelphia (CHOP) and analyzed by researchers from January 2008 through December 2015.

Among 253 patients recruited for the study, 25 patients required postnatal CSF diversion were selected and matched by gestational age with 25 patients with fetal ventriculomegaly who did not require CSF diversion. 

The model was applied to a sample of 24 consecutive patients with fetal ventriculomegaly who were evaluated at a separate institution (the replication cohort) from January 1998, through December 2007. Data were analyzed from January 1998, through December 2009.

The main study outcomes included accuracy, sensitivity, and specificity of the model to classify patients who were in need of postnatal CSF diversion correctly.

The researchers generated a fetal MRI based model stemmed from four notable measurements extracted from the fetal MRI images: linear measurements, area, volume, and morphologic features.

Measurements were then inputted into a machine learning algorithm that determined the combination of features to predict whether each patient required postnatal CSF diversion.

A total of 74 patients (41 girls and 33 boys; mean gestational age, 27.0±5.6] months) were included from both cohorts in the analysis. In the discovery cohort, a median time to CSF diversion was 6 days (interquartile range [IQR], 2-51 days).

The patients with fetal ventriculomegaly who did not develop symptoms were followed up for a median of 29 months (IQR, 9-46 months).

The discovery cohort model correctly classified patients who required CSF diversion with 82% accuracy, 80% sensitivity, and 84% specificity. In the replication cohort, on the other hand, the model attained 91% accuracy, 75% sensitivity, and 95% specificity.

The researchers concluded that an MRI-based predictive model with high levels of accuracy and generalizability may provide clinicians with prenatal prognostic information and help guide postnatal management of infants with fetal ventriculomegaly, and may also be helpful in the selection of candidates for potential fetal intervention.