Classification of structural lesions in magnetic resonance imaging. Surgical implications in drug-resistant epilepsy patients
*Corresponding author:Cristina V. Torres Díaz, MD. Division of Neurosurgery. University Hospital La Princesa. Diego de León, 32. E-28006 Madrid (Spain).
E-mail: cristinatorresdiaz@yahoo.es
Introduction: The presence of a structural lesion in the preoperative magnetic resonance imaging (MRI) of drug-resistant epilepsy patients has been usually associated with a favourable surgical outcome. We present our experience in our Epilepsy Surgery Unit.
Patients and methods: Clinical records from 265 patients, operated on from 1990-2010 in our institution, were reviewed. Patients were classified, according to MRI findings, into three groups: surgical lesion (SL), tumors or vascular malformations requiring surgery 'per se'; orientative lesion (OL), dysplasia, atrophy or mesial temporal sclerosis; and (NL) group, with normal MRI. Seizure outcomes were analysed in relation to this classification.
Results: Period 1990-2000, 151 patients: 87% of SL, 65% of OL and 57% of NL patients were in Engel class I or II at the two-year follow-up. Among temporal lobe epilepsy cases (TLE), 87% of SL, 67% of OL and 56% of NL patients achieved seizure control. Differences were statistically significative. Period 2001-2010, 114 patients: 100% of SL, 90% of OL, and 81% of NL patients were in Engel's class I or II. Both TLE and extratemporal (ETLE) SL patients obtained a 100% seizure control. Among the OL patients, 95% with TLE and 43% of ETLE achieved seizure control. In the NL group, the percentages were 88% in TLE, and 50% in ETLE.
Conclusions: In our series, SL was a predictor of a favorable outcome. In TLE patients, good results were achieved despite normal MRI. Patients with ETLE and NL did not have a worse outcome than those with OL. A classification in SL, OL and NL seems more helpful for predicting the surgical outcome than the traditional classification lesion versus non-lesion MRI. Radiological findings must be carefully evaluated in the context of a complete epilepsy surgery evaluation.
Pacientes y métodos Se estudiaron retrospectivamente los pacientes intervenidos por epilepsia farmacorresistente. Distinguimos dos períodos: 1990-2000 (RM de 0,5 T) y 2001-2008 (RM de 1,5 T). La RM preoperatoria se clasificó en tres grupos: RM con lesión quirúrgica (LQ), RM orientativa (LO) y RM normal (NL). También se efectuó una clasificación anatomopatológica similar. Se correlacionaron las distintas clasificaciones y los resultados quirúrgicos.
Resultados Período 1990-2000: 151 pacientes. El 70% quedó en las clases de Engel I o II. Según la RM, los resultados fueron: LQ, 87%; LO, 65%; y NL, 57%. Las diferencias fueron estadísticamente significativas. Período 2001-2008: 114 pacientes. El 89% quedó en las clases de Engel I o II. Según la RM: LQ, 100%; LO, 90%; y NL, 81%. Las diferencias fueron estadísticamente significativas. Los pacientes con epilepsia del lóbulo temporal y extratemporal con LQ tuvieron un 100% de control; con LO, el 95% con epilepsia del lóbulo temporal y el 43% con estado epiléptico; en aquellos pacientes sin lesión (NL), el 88% con epilepsia del lóbulo temporal se controló frente al 50% con estado epiléptico.
Conclusiones La RM es una herramienta eficaz en la selección de candidatos quirúrgicos en la epilepsia. La LQ asocia muy buen pronóstico. En la epilepsia del lóbulo temporal se pueden obtener muy buenos resultados (80-90% de control) a pesar de una RM normal. En el estado epiléptico, las LO pueden tener peor resultado que la NL en la RM.
