Detection of covert lesions in focal epilepsy using computational analysis of multimodal magnetic resonance imaging data

The utility of Multicentre Epilepsy Lesion Detection (MELD) algorithm in identifying epileptic activity and predicting seizure freedom in MRI lesion-negative paediatric patients

AIM

Paediatric patients with drug-resistant focal epilepsy (DRFE) who have no clear focal lesion identified on conventional structural magnetic resonance imaging (MRI) are a particularly challenging cohort to treat and form an increasing part of epilepsy surgery programs. A recently developed deep-learning-based MRI lesion detection algorithm, the Multicentre Lesion Detection (MELD) algorithm, has been shown to aid detection of focal cortical dysplasia (FCD). We applied this algorithm retrospectively to a cohort of MRI-negative children with refractory focal epilepsy who underwent stereoelectroencephalography (SEEG) to determine its accuracy in identifying unseen epileptic lesions, seizure onset zones and clinical outcomes.

METHODS

We retrospectively applied the MELD algorithm to a consecutive series of MRI-negative patients who underwent SEEG at our tertiary Paediatric Epilepsy Surgery centre. We assessed the extent to which the identified MELD cluster or lesion area corresponded with the clinical seizure hypothesis, the epileptic network, and the positron emission tomography (PET) focal hypometabolic area. In those who underwent resective surgery, we analysed whether the region of MELD abnormality corresponded with the surgical target and to what extent this was associated with seizure freedom.

RESULTS

We identified 37 SEEG studies in 28 MRI-negative children in whom we could run the MELD algorithm. Of these, 14 (50 %) children had clusters identified on MELD. Nine (32 %) children had clusters concordant with seizure hypothesis, 6 (21 %) had clusters concordant with PET imaging, and 5 (18 %) children had at least one cluster concordant with SEEG electrode placement. Overall, 4 MELD clusters in 4 separate children correctly predicted either seizure onset zone or irritative zone based on SEEG stimulation data. Sixteen children (57 %) went on to have resective or lesional surgery. Of these, only one patient (4 %) had a MELD cluster which co-localised with the resection cavity and this child had an Engel 1 A outcome.

CONCLUSIONS

In our paediatric cohort of MRI-negative patients with drug-resistant focal epilepsy, the MELD algorithm identified abnormal clusters or lesions in half of cases, and identified one radiologically occult focal cortical dysplasia. Machine-learning-based lesion detection is a promising area of research with the potential to improve seizure outcomes in this challenging cohort of radiologically occult FCD cases. However, its application should be approached with caution, especially with regards to its specificity in detecting FCD lesions, and there is still work to be done before it adds to diagnostic utility.

Causal Connectivity Network Analysis of Ictal Electrocorticogram With Temporal Lobe Epilepsy Based on Dynamic Phase Transfer Entropy

Temporallobe epilepsy (TLE) has been conceptualized as a brain network disease, which generates brain connectivity dynamics within and beyond the temporal lobe structures in seizures. The hippocampus is a representative epileptogenic focus in TLE. Understanding the causal connectivity in terms of brain network during seizures is crucial in revealing the triggering mechanism of epileptic seizures originating from the hippocampus (HPC) spread to the lateral temporal cortex (LTC) by ictal electrocorticogram (ECoG), particularly in high-frequency oscillations (HFOs) bands. In this study, we proposed the unified-epoch dynamic causality analysis method to investigate the causal influence dynamics between two brain regions (HPC and LTC) at interictal and ictal phases in the frequency range of 1-500 Hz by introducing the phase transfer entropy (PTE) out/in-ratio and sliding window. We also proposed PTE-based machine learning algorithms to identify epileptogenic zone (EZ). Nine patients with a total of 26 seizures were included in this study. We hypothesized that: 1) HPC is the focus with the stronger causal connectivity than that in LTC in the ictal state at gamma and HFOs bands. 2) Causal connectivity in the ictal phase shows significant changes compared to that in the interictal phase. 3) The PTE out/in-ratio in the HFOs band can identify the EZ with the best prediction performance.

Identification of different MRI atrophy progression trajectories in epilepsy by subtype and stage inference

Artificial intelligence (AI)-based tools are widely employed, but their use for diagnosis and prognosis of neurological disorders is still evolving. Here we analyse a cross-sectional multicentre structural MRI dataset of 696 people with epilepsy and 118 control subjects. We use an innovative machine-learning algorithm, Subtype and Stage Inference, to develop a novel data-driven disease taxonomy, whereby epilepsy subtypes correspond to distinct patterns of spatiotemporal progression of brain atrophy.In a discovery cohort of 814 individuals, we identify two subtypes common to focal and idiopathic generalized epilepsies, characterized by progression of grey matter atrophy driven by the cortex or the basal ganglia. A third subtype, only detected in focal epilepsies, was characterized by hippocampal atrophy. We corroborate external validity via an independent cohort of 254 people and confirm that the basal ganglia subtype is associated with the most severe epilepsy.Our findings suggest fundamental processes underlying the progression of epilepsy-related brain atrophy. We deliver a novel MRI- and AI-guided epilepsy taxonomy, which could be used for individualized prognostics and targeted therapeutics.

Advances in Image Processing for Epileptogenic Zone Detection with MRI

Focal epilepsy is a common and severe neurologic disorder. Neuroimaging aims to identify the epileptogenic zone (EZ), preferably as a macroscopic structural lesion. For approximately a third of patients with chronic drug-resistant focal epilepsy, the EZ cannot be precisely identified using standard 3.0-T MRI. This may be due to either the EZ being undetectable at imaging or the seizure activity being caused by a physiologic abnormality rather than a structural lesion. Computational image processing has recently been shown to aid radiologic assessments and increase the success rate of uncovering suspicious regions by enhancing their visual conspicuity. While structural image analysis is at the forefront of EZ detection, physiologic image analysis has also been shown to provide valuable information about EZ location. This narrative review summarizes and explains the current state-of-the-art computational approaches for image analysis and presents their potential for EZ detection. Current limitations of the methods and possible future directions to augment EZ detection are discussed.

Recent advances in epilepsy surgery

PURPOSE OF REVIEW

Technological innovations in the preoperative evaluation, surgical techniques and outcome prediction in epilepsy surgery have grown exponentially over the last decade. This review highlights and emphasizes relevant updates in techniques and diagnostic tools, discussing their context within standard practice at comprehensive epilepsy centres.

RECENT FINDINGS

High-resolution structural imaging has set an unprecedented opportunity to detect previously unrecognized subtle abnormalities. Machine learning and computer science are impacting the methodologies to analyse presurgical and surgical outcome data, building more accurate prediction models to tailor treatment strategies. Robotic-assisted placement of depth electrodes has increased the safety and ability to sample epileptogenic nodes within deep structures, improving our understanding of the seizure networks in drug-resistant epilepsy. The current available minimally invasive techniques are reasonable surgical alternatives to ablate or disrupt epileptogenic regions, although their sustained efficacy is still an active area of research.

SUMMARY

Epilepsy surgery is still underutilized worldwide. Every patient who continues with seizures despite adequate trials of two well selected and tolerated antiseizure medications should be evaluated for surgical candidacy. Collaboration between academic epilepsy centres is of paramount importance to answer long-standing questions in epilepsy surgery regarding the understanding of spatio-temporal dynamics in epileptogenic networks and its impact on surgical outcomes.

The Temporal Lobe as a Symptomatogenic Zone in Medial Parietal Lobe Epilepsy

Some surgical failures after temporal lobe epilepsy surgery may be due to the presence of an extratemporal epileptogenic zone. Of particular interest is the medial parietal lobe due to its robust connectivity with mesial temporal structures. Seizures in that area may be clinically silent before propagating to the symptomatogenic temporal lobe. In this paper, we present an overview of the anatomical connectivity, semiology, radiology, electroencephalography, neuropsychology, and outcomes in medial parietal lobe epilepsy. We also present two illustrative cases of seizures originating from the precuneus and the posterior cingulate cortex. We conclude that the medial parietal lobe should be strongly considered for sampling by intracranial electrodes in individuals with nonlesional temporal lobe epilepsy, especially if scrutinizing the presurgical data produces discordant findings.

Non-parametric combination of multimodal MRI for lesion detection in focal epilepsy

•

Multivariate voxel-based analysis useful for lesion detection in focal epilepsy.

•

Non-parametric combination algorithm used to combine data from various MR sequences.

•

Successful lesion detection demonstrated in MRI-positive and MRI-negative patients.

•

Multimodal analysis detected abnormalities from diverse epileptogenic pathologies.

•

Sensitivity of multivariate analysis notably higher than univariate analyses.

One third of patients with medically refractory focal epilepsy have normal-appearing MRI scans. This poses a problem as identification of the epileptogenic region is required for surgical treatment. This study performs a multimodal voxel-based analysis (VBA) to identify brain abnormalities in MRI-negative focal epilepsy. Data was collected from 69 focal epilepsy patients (42 with discrete lesions on MRI scans, 27 with no visible findings on scans), and 62 healthy controls. MR images comprised T1-weighted, fluid-attenuated inversion recovery (FLAIR), fractional anisotropy (FA) and mean diffusivity (MD) from diffusion tensor imaging, and neurite density index (NDI) from neurite orientation dispersion and density imaging. These multimodal images were coregistered to T1-weighted scans, normalized to a standard space, and smoothed with 8 mm FWHM. Initial analysis performed voxel-wise one-tailed t-tests separately on grey matter concentration (GMC), FLAIR, FA, MD, and NDI, comparing patients with epilepsy to controls. A multimodal non-parametric combination (NPC) analysis was also performed simultaneously on FLAIR, FA, MD, and NDI. Resulting p-maps were family-wise error rate corrected, threshold-free cluster enhanced, and thresholded at p < 0.05. Sensitivity was established through visual comparison of results to manually drawn lesion masks or seizure onset zone (SOZ) from stereoelectroencephalography. A leave-one-out cross-validation with the same analysis protocols was performed on controls to determine specificity. NDI was the best performing individual modality, detecting focal abnormalities in 38% of patients with normal MRI and conclusive SOZ. GMC demonstrated the lowest sensitivity at 19%. NPC provided superior performance to univariate analyses with 50% sensitivity. Specificity in controls ranged between 96 and 100% for all analyses. This study demonstrated the utility of a multimodal VBA utilizing NPC for detecting epileptogenic lesions in MRI-negative focal epilepsy. Future work will apply this approach to datasets from other centres and will experiment with different combinations of MR sequences.

The benefit of the diffusion kurtosis imaging in presurgical evaluation in patients with focal MR-negative epilepsy

The effectivity of diffusion-weighted MRI methods in detecting the epileptogenic zone (EZ) was tested. Patients with refractory epilepsy (N=25) who subsequently underwent resective surgery were recruited. First, the extent of white matter (WM) asymmetry from mean kurtosis (MK) was calculated in order to detect the lobe with the strongest impairment. Second, a newly developed metric was used, reflecting a selection of brain areas with concurrently increased mean Diffusivity, reduced fractional Anisotropy, and reduced mean Kurtosis (iDrArK). A two-step EZ detection was performed as (1) lobe-specific detection, (2) iDrArK voxel-wise detection (with a possible lobe-specific restriction if the result of the first step was significant in a given subject). The method results were compared with the surgery resection zones. From the whole cohort (N=25), the numbers of patients with significant results were: 10 patients in lobe detection and 9 patients in EZ detection. From these subsets of patients with significant results, the impaired lobe was successfully detected with 100% accuracy; the EZ was successfully detected with 89% accuracy. The detection of the EZ using iDrArK was substantially more successful when compared with solo diffusional parameters (or their pairwise combinations). For a subgroup with significant results from step one (N=10), iDrArK without lobe restriction achieved 37.5% accuracy; lobe-restricted iDrArK achieved 100% accuracy. The study shows the plausibility of MK for detecting widespread WM changes and the benefit of combining different diffusional voxel-wise parameters.

Clinical Application of Machine Learning Models for Brain Imaging in Epilepsy: A Review

Epilepsy is a common neurological disorder characterized by recurrent and disabling seizures. An increasing number of clinical and experimental applications of machine learning (ML) methods for epilepsy and other neurological and psychiatric disorders are available. ML methods have the potential to provide a reliable and optimal performance for clinical diagnoses, prediction, and personalized medicine by using mathematical algorithms and computational approaches. There are now several applications of ML for epilepsy, including neuroimaging analyses. For precise and reliable clinical applications in epilepsy and neuroimaging, the diverse ML methodologies should be examined and validated. We review the clinical applications of ML models for brain imaging in epilepsy obtained from a PubMed database search in February 2021. We first present an overview of typical neuroimaging modalities and ML models used in the epilepsy studies and then focus on the existing applications of ML models for brain imaging in epilepsy based on the following clinical aspects: (i) distinguishing individuals with epilepsy from healthy controls, (ii) lateralization of the temporal lobe epilepsy focus, (iii) the identification of epileptogenic foci, (iv) the prediction of clinical outcomes, and (v) brain-age prediction. We address the practical problems and challenges described in the literature and suggest some future research directions.