Rationale:
Congenital brain malformations are a major cause of epilepsy and developmental delay. Fetal MRI is increasingly used to detect these anomalies, offering greater anatomical resolution than ultrasound, particularly for cortical anomalies. However, its accuracy and utility in predicting neonatal seizures and long-term epilepsy outcome remain uncertain. This study examines the predictive value of fetal MRI in identifying seizure risk and epilepsy outcomes, to inform early neonatal interventions and enhance prenatal counseling by providing families with clearer prognostic information. Methods:
We conducted a retrospective review of 34 patients with brain lesions identified on fetal MRI. Eight were excluded due to fetal demise, termination, stillbirth, early neonatal death, or loss to follow-up. The remaining 26 patients were analyzed, with a median follow-up age of 48 months, focusing on clinical history, neuroimaging, postnatal course, EEG findings, and epilepsy outcomes. Results:
Fetal MRI was performed at a median gestational age of 31 weeks, with delivery occurring at a median of 38 weeks. Among the cohort, 20 patients had cortical anomalies and 6 had non-cortical anomalies, including holoprosencephaly and encephalocele, confirmed on postnatal MRI. Fetal MRI detected brain anomalies in 88% of cases (23/26). Cortical anomalies were identified in 85% (17/20) of patients, with exact lesion subtypes correctly classified in 40% (8/20). EEG monitoring was performed in 12 patients, typically on the first day of life. Five underwent routine screening with no seizures detected. The remaining seven had long-term EEGs prompted by clinical concerns—three due to apnea/desaturations and four for suspected seizures. Seizures were confirmed in three patients (11.5%): two had clinical events, and one presented with apnea. All three subsequently developed refractory epilepsy. Epilepsy was diagnosed in 18 patients (69%), with a median onset at 6 months. Early-onset epilepsy (≤ 6 months) was observed in patients with lissencephaly, TSC, heterotopia, and cortical dysplasia, while later onset occurred in cases of schizencephaly and polymicrogyria.
Conclusions: Fetal MRI is highly effective in detecting cortical malformations but demonstrates only moderate accuracy in identifying specific lesion subtypes, regardless of the gestational age at which it is performed. Although limited in predicting neonatal seizures without clinical signs, it remains a valuable tool for risk stratification based on lesion type and is strongly associated with epilepsy severity and long-term outcomes. Neonatal EEG is most useful when guided by clinical symptoms, particularly in cases of cortical dysplasia, heterotopia, and lissencephaly, where seizures often precede intractable epilepsy. Early-onset epilepsy and infantile spasms are most often linked to lissencephaly, heterotopia, and TSC, supporting the need for close postnatal monitoring. Developmental delay was also commonly observed across these high-risk groups.
Funding: No funding received for this research