Dysembryoplastic neuroepithelial tumors: an MRI-based scheme for epilepsy surgery.

Objective: To determine optimal resections in the 3 dysembryoplastic neuroepithelial tumor (DNT) histologic subtypes (simple, complex, and nonspecific) based on MRI features.
Methods: In 78 consecutive epilepsy patients operated for DNT, MRI features were classified as follows: type 1 (cystic/polycystic-like, well-delineated, strongly hypointense T1), type 2 (nodular-like, heterogeneous), or type 3 (dysplastic-like, iso/hyposignal T1, poor delineation, gray-white matter blurring). Correlations between histology, neurophysiologic findings, and surgical outcome were established for each MRI subtype.
Results: Type 1 MRI (25 cases, in temporal and extratemporal areas) always corresponded to simple or complex DNTs. Type 2 MRI (25 cases, predominantly in neocortical areas) and type 3 MRI (28 cases, mainly in the mesial temporal lobe) corresponded to nonspecific forms. The epileptogenic zone (EZ) differed significantly according to the MRI subtype (p = 0.0029). It colocalized with the tumor in type 1 MRI, included perilesional cortex in type 2 MRI, and involved extensive areas in type 3 MRI. Cortical dysplasia was predominantly found in type 3 MRI (p < 0.0001). The main prognostic factors for seizure-free outcome (83%) were complete tumor (p < 0.0001) and EZ (p = 0.0115) removal. Other factors favorably influencing the outcome were a short epilepsy duration (p = 0.013) and absence of cortical-subcortical damage at the resection site (p = 0.053). Age at surgery was not related to outcome; however, cortical-subcortical damage was correlated with old age (p = 0.021). Treatment discontinuation was correlated with young age at surgery (p = 0.004) and short epilepsy duration (p = 0.001).
Conclusion: We propose that resection might be restricted to the tumor in type 1 MRI and be more extensive in other MRI subtypes, especially in type 3 MRI. Early surgery and clean surgical margins are crucial for curing epilepsy.