Making the Most of Brain MRI Through Computer-Assisted Morphometric Assessment

By Z. Irene Wang, PhD

Interpretation of MRI data can be markedly improved by computer-assisted methods designed to pull out information not easily seen by visual analysis. At Cleveland Clinic’s Epilepsy Center, scientists and clinicians are spearheading efforts to use computer-assisted morphometric assessment of brain MRIs to improve noninvasive epilepsy localization in clinical practice. This strategy offers a novel field of analysis that goes beyond expert visual inspection of the MRI and allows for rater-independent imaging correlates to be generated for each individual patient.

Clinical application already underway

The new technology has already made an impact on many patients with epilepsy evaluated at Cleveland Clinic, particularly those whose MRIs were negative by conventional visual analysis. These “MRI-negative” or “nonlesional” patients, approximately 40 percent of the total epilepsy surgery population, typically require expensive and invasive intracranial electroencephalography (EEG) to identify the epileptogenic zone.

Several years ago, our Epilepsy Center research group published a study indicating that the main pathology in nonlesional epilepsy is focal cortical dysplasia (FCD).1 These lesions are usually characterized by subtle MRI features and can be difficult to identify by visual analysis of 3T MRI scans. Our group performed well-designed studies to validate the usefulness of a voxel-based morphometric analysis program (MAP) for detection of subtle FCD.2,3 Reliable performance has been seen on MRI scans with different field strengths (1.5T/3T) and from different scanners (Siemens, Philips and GE). In a retrospective study of a cohort of 150 MRI-negative surgical patients,3 MAP-guided MRI interpretation showed positive findings in 43 percent, and complete resection of the MAP-positive region correlated positively with seizure-free outcome. The false-positive rate was 2 percent in a group of 52 healthy controls.

An illustrative case

Figure 1 presents images from a 37-year-old right-handed man evaluated for surgery for drug-refractory focal epilepsy. Seizure semiology suggested a perisylvian/opercular onset of unclear laterality. During three previous video-EEG monitoring evaluations, no interictal spikes were captured, and ictal events were lateralized only to the right hemisphere. This patient had several previous MRIs that were reported as normal. Other noninvasive modalities did not provide additional localization.

MRI post-processing using MAP showed a subtle lesion represented by gray-white blurring in the ventral aspect of the depth of the right central sulcus, which was concordant with the seizure semiology. This finding prompted a very focal design of stereoelectroencephalographic (SEEG) electrode implantation, which confirmed the epileptogenicity of the subtle lesion. The patient became seizure-free after focal resection, which was confirmed to contain the lesion.

Figure 1. Images from an illustrative case in which computer-assisted morphometric assessment of MRI (MAP) provided essential information. A subtle lesion was detected by MRI post-processing using the MAP method in the depth of the central sulcus (top row of images). This finding was missed on multiple previous MRI scans and only became apparent based on MAP-guided analysis of the MRI. The first image in the second row shows the ictal onset location from invasive evaluation using SEEG (onset electrodes marked with red; other electrodes marked with green). A typical seizure captured during the invasive monitoring is shown in the bottom panel. Seizure onset at R5-8 is marked by the red arrow; dashed arrows denote the quick spread to Q7-10. Ictal onset was in the exact location of the detected subtle abnormality. The postoperative MRI in the second row of images shows complete resection of the abnormality. Pathological examination showed type IIb focal cortical dysplasia in the surgical specimen, characterized by balloon cells (lower right panel; HE stain, ×200).

Our published studies provided validation of the morphometric analysis methods, which enabled us to integrate them in the routine presurgical evaluation of patients with MRI-negative medically intractable epilepsies at Cleveland Clinic’s Epilepsy Center. With minimal extra cost and no additional tests or risks to the patient, this process has shed light on many difficult cases like the one profiled in Figure 1.

Next frontier: Application to 7T

Moreover, the MAP method is now being applied to 7T MRI. Figure 2 shows images from a patient with histopathologically confirmed FCD type IIb whose 3T MRI was previously interpreted as normal but in whom 7T MRI revealed a transmantle abnormality in the right frontal lobe. The presence and extent of the abnormality in the cortical mantle was further illustrated by MAP: The 7T MAP results showed markedly enhanced visualization of blurring in the gray-white boundary radially connecting the cortex to the ventricle, whereas the 3T MAP results also highlighted the region but only showed the component adjacent to the cortex. These findings strongly suggest that higher field strength combined with MRI morphometric analysis can assist in delineation of epileptogenic pathologies.

Figure 2. Comparison of the application of MAP to 3T and 7T MRIs in a patient with histopathologically confirmed FCD type IIb. The top row shows a 3T T1w sequence followed by a MAP gray-white junction feature map showing the subtle FCD lesion with a voxel size of 1 mm3. The bottom row shows a 7T T1w sequence followed by a MAP gray-white junction feature map highlighting the same lesion with a voxel size of 0.5 mm3.

Dr. Wang will speak on voxel-based morphometry and other MRI post-processing techniques in epilepsy at the CME/MOC-certified Cleveland Clinic Epilepsy Update & Review Course, to be held in Cleveland Sept. 22-24, 2018. For information and registration, visit ccfcme.org/epilepsyupdate18.

References

1. Wang ZI, Alexopoulos AV, Jones SE, Jaisani Z, Najm IM, Prayson RA. The pathology of magnetic-resonance-imaging-negative epilepsy. Mod Pathol. 2013;26:1051-1058.
2. Wang ZI, Alexopoulos AV, Jones SE, et al. Linking MRI postprocessing with magnetic source imaging in MRI-negative epilepsy. Ann Neurol. 2014;75:759-770.
3. Wang ZI, Jones SE, Jaisani Z, et al. Voxel-based morphometric magnetic resonance imaging (MRI) postprocessing in MRI-negative epilepsies. Ann Neurol. 2015;77:1060-1075.

Dr. Wang is a staff scientist in Cleveland Clinic’s Epilepsy Center and joint staff in the Department of Biomedical Engineering, Cleveland Clinic Lerner Research Institute.