Utility of additional dedicated high-resolution 3T MRI in children with medically refractory focal epilepsy

Highlights

• Additional high-resolution MRI may aid in identifying an epileptogenic lesion in focal epilepsy patients with negative 3T MR.

• Addition of high-resolution MRI identified a lesion in 66% of patients who had normal MRI on standard epilepsy protocol.

• The MRI lesions were concordant to surgical resection in all patients.

• Lesion identification guided placement of depth electrodes and/or subdural strips in 80% who underwent invasive monitoring.

Abstract

Purpose

In patients with medically refractory epilepsy and normal magnetic resonance imaging (MRI), high-resolution dedicated MRI may identify cryptic lesions. The aim of this study was to assess improvement in lesion detection and its impact on clinical management, using additional high-resolution dedicated 3T MRI in children with medically refractory epilepsy who had normal 3T epilepsy protocol MRI.

Materials and methods

Children who had resective epilepsy surgery and suspected focal cortical dysplasia (FCD) or normal 3T epilepsy protocol MRI were included. Those with other diagnosis on MRI including tumor and hippocampal sclerosis were excluded. Patients who had normal MRI on 3T epilepsy protocol underwent dedicated high-resolution 3T MRI through the epileptogenic zone, guided by video EEG, Magnetoencephalography and FDG-PET data.

Results

101 patients with at least 1 year follow-up were included. Twenty-nine of 44 (66%) patients who had normal epilepsy protocol MRI had a lesion identified on dedicated high-resolution MRI. The addition of dedicated high-resolution MRI to standard epilepsy protocol increased sensitivity from 53.1% (95%CI: 40%–66%) to 85.9% (95%CI: 75%–93%). Identified lesions were concordant to surgical resection in all patients and guided depth/strip electrode insertion in 20/25 (80%) patients who underwent staged resection. Dedicated MRI detected small deep seated lesions in 10/20 (50%), and guided depth electrodes placement, without which it would not be feasible, as the lobar location of epileptogenic zone from other non-invasive tests were not sufficiently precise.

Conclusion

Patients with non-lesional epilepsy on standard epilepsy protocol MR may benefit from high-resolution dedicated MRI to aid identification of an underlying lesion, which could impact surgical management and improve seizure control.

Introduction

Approximately 30% of patients with epilepsy become medically intractable (Berg, 2001; Engel, 1998; Farrell et al., 2006; Kwan and Brodie, 2000). Epilepsy surgery offers the potential to render these patients seizure free. However, successful outcomes in epilepsy surgery, are in part dependent upon identification of a lesion on Magnetic Resonance Imaging (MRI). Up to 20–40% of patients with refractory epilepsy have no identifiable lesions on MRI (Carne et al., 2004; Hong et al., 2002). Patients with non-lesional epilepsy, that is, no lesion seen on MRI, have poorer surgical outcomes compared to those with a lesion seen on MRI (Bien et al., 2009; Tellez-Zenteno et al., 2010; Tonini et al., 2004). Higher field magnets, such as 3T MRI, have higher yield of identifying a lesion relative to 1.5T MRI (Knake et al., 2005; Mellerio et al., 2014; Nguyen et al., 2010; Phal et al., 2008; Winston et al., 2014; Zijlmans et al., 2009) due to higher signal-to-noise ratio and improved resolution. We have previously shown that the use of an epilepsy protocol on 3T MRI, could improve surgical outcome in children with medically refractory focal epilepsy. However, some patients who have undergone an epilepsy protocol 3T MRI continue to have a normal appearing MRI. These patients may have an underlying subtle focal cortical dysplasia that is difficult to detect even on a high-resolution epilepsy protocol.

We postulate, that by targeting the epileptogenic zone with higher-resolution 3T MRI, we could potentially increase the yield of identifying a lesion on presurgical MRI studies. Our hypotheses were that first, the addition of dedicated high-resolution 3T MRI, targeting the epileptogenic zone in patients with an initial negative standard 3T epilepsy protocol MRI, would improve the diagnostic performance of MRI. Second, the lesion thus identified on the dedicated high-resolution MRI study could guide placement of invasive monitoring including depth and/or strip electrodes. The study aims were first, to evaluate the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the standard 3T epilepsy protocol MRI and for dedicated high-resolution 3T MRI following a standard epilepsy protocol 3T MRI in children with refractory epilepsy. Our second aim was to assess whether the lesion identified on the dedicated high-resolution 3T MRI guided clinical management.