Fluid and White Matter Suppression Imaging and Voxel-Based Morphometric Analysis in Conventional Magnetic Resonance Imaging-Negative Epilepsy

Abstract

Purpose: Delineation of subtle lesions in magnetic resonance imaging (MRI)-negative patients is of great importance in preoperative epilepsy evaluation. The aim of our study was to explore the diagnostic value of the novel fluid and white matter suppression (FLAWS) sequence in comparison with a voxel-based MRI postprocessing morphometric analysis program (MAP) in a consecutive cohort of non-lesional patients. Methods: Surgical candidates with a negative finding on an official neuroradiology report were enrolled. High-resolution FLAWS image and MAP maps generated based on high-resolution three-dimensional (3D) T1 image were visually inspected for each patient. The findings of FLAWS or MAP-positive (FLAWS/MAP+) regions were compared with the surgical resection cavity in correlation with surgical outcome and pathology. Results: Forty-five patients were enrolled; the pathological examination revealed focal cortical dysplasia (FCD) in 32 patients and other findings in 13 patients. The positive rate, sensitivity, and specificity were 48.9%, 0.43, and 0.87, respectively, for FLAWS and 64.4%, 0.57, and 0.8, respectively, for MAP. Concordance between surgical resection and FLAWS+ or MAP+ regions was significantly associated with a seizure-free outcome (FLAWS: p = 0.002; MAP: p = 0.0003). A positive finding in FLAWS and MAP together with abnormalities in the same gyrus (FLAWS-MAP gyral+) was detected in 31.1% of patients. FLAWS+ only and MAP+ only were found in 7 (15.5%) and 14 (31.1%) patients, respectively. Conclusions: FLAWS showed a promising value for identifying subtle epileptogenic lesions and can be used as a complement to current MAP in patients with MRI-negative epilepsy.

Keywords: MRI post-processing; epilepsy; flaws; focal cortical dysplasia; non-lesional; voxel-based morphometric.

Figure 1

Flowchart of patient enrollment. FLAWS unperformed due to organizational problems: the time between presurgical evaluation and invasive implantation or resective surgery was too close to allow for a FLAWS scan.

Figure 2

Examples of patients having FLAWS–MAP gyral+ region concordant with the surgical resection cavity. Both patients remained seizure-free at 12 months. The crosshairs indicate the same coordinate between different images. The red arrows indicate the FLAWS+ or MAP+ region. Upper, left: registered FLAWS image; right: junction map of MAP. Lower, left: T1-weighted magnetization-prepared rapid acquisition with gradient-echo (MPRAGE) image; middle: registered T2-weighted fluid-attenuated inversion recovery (FLAIR) image; right: registered postsurgical computerized tomography (CT) indicating the site and extent of resection. (A) belongs to the “subtly lesional” subgroup, FLAWS–MAP gyral+ region in the parietal lobe; FLAWS image shows blurring of gray matter and white matter transition as well as abnormal subcortical white matter signal. The junction file shows the suprathreshold region located in the same gyrus with FLAWS image. (B) belongs to the “subtly lesional” subgroup, FLAWS–MAP gyral+ region in the frontal lobe; FLAWS image shows abnormal cortical thickness, blurring of gray matter and white matter transition, and transmantle sign; the junction map shows a half ring-shaped suprathreshold region, which was located in the same sulcus with FLAWS image. Pathology: (A) focal cortical dysplasia (FCD) IIb; (B) FCD IIa.

Figure 3

Examples of patients with only FLAWS+ region concordant with the surgical resection cavity. Both patients were seizure-free after surgery at 12 months. The crosshairs indicate the same coordinate between different images. The red arrows indicate the FLAWS+ region. Upper, left: registered FLAWS image; right: junction map of MAP. Lower, left T1-weighted magnetization-prepared rapid acquisition with gradient-echo (MPRAGE) image; middle: registered T2-weighted fluid-attenuated inversion recovery (FLAIR) image; right: registered postsurgical CT indicating the site and extent of resection. (A) belongs to the “strictly nonlesional” subgroup, FLAWS+ region in the frontal lobe; FLAWS image shows blurring of gray matter and white matter transition as well as abnormal cortical thickness. The junction file shows no suprathreshold region. (B) belongs to the “strictly nonlesional” subgroup, FLAWS+ region in the temporal lobe; FLAWS image shows blurring of gray matter and white matter transition as well as abnormal subcortical white matter signal; the junction map shows no suprathreshold region. Pathology: (A) FCD IIb; (B) gliosis.

Figure 4

Examples of patients who had subtle FLAWS+ region revealed only under the guidance of the MAP+ region. Both patients were seizure-free after surgery. The FLAWS+ and MAP+ region was concordance with resection. The crosshairs indicate the same coordinate between different images. The red arrows indicate the FLAWS+ or MAP+ region. Upper, left: registered FLAWS image; right: junction map of MAP. Lower, left T1-weighted magnetization-prepared rapid acquisition with gradient-echo (MPRAGE) image; middle: registered T2-weighted fluid-attenuated inversion recovery (FLAIR) image; right: registered postsurgical CT indicating the site and extent of resection. (A) belongs to the “strictly nonlesional” subgroup, MAP guided FLAWS+ region in the temporal lobe; the junction file shows the suprathreshold region in the temporal lobe; FLAWS image shows subtle blurring of gray matter and white matter transition as well as abnormal subcortical white matter intensity. (B) belongs to the “subtly lesional” subgroup, MAP guided FLAWS+ region in the parietal lobe; the junction file shows the suprathreshold region in the parietal lobe; FLAWS image shows very subtle blurring of gray matter and white matter transition. Pathology: (A) FCD Ib; (B) FCD Ia.