Clinical Value of Machine Learning in the Automated Detection of Focal Cortical Dysplasia Using Quantitative Multimodal Surface-Based Features

Development and prospective clinical validation of a convolutional neural network for automated detection and segmentation of focal cortical dysplasias

PURPOSE

Focal cortical dysplasias (FCDs) are a leading cause of drug-resistant epilepsy. Early detection and resection of FCDs have favorable prognostic implications for postoperative seizure freedom. Despite advancements in imaging methods, FCD detection remains challenging. House et al. (2021) introduced a convolutional neural network (CNN) for automated FCD detection and segmentation, achieving a sensitivity of 77.8%. However, its clinical applicability was limited due to a low specificity of 5.5%. The objective of this study was to improve the CNN's performance through data-driven training and algorithm optimization, followed by a prospective validation on daily-routine MRIs.

MATERIAL AND METHODS

A dataset of 300 3 T MRIs from daily clinical practice, including 3D T1 and FLAIR sequences, was prospectively compiled. The MRIs were visually evaluated by two neuroradiologists and underwent morphometric assessment by two epileptologists. The dataset included 30 FCD cases (11 female, mean age: 28.1 ± 10.1 years) and a control group of 150 normal cases (97 female, mean age: 32.8 ± 14.9 years), along with 120 non-FCD pathological cases (64 female, mean age: 38.4 ± 18.4 years). The dataset was divided into three subsets, each analyzed by the CNN. Subsequently, the CNN underwent a two-phase-training process, incorporating subset MRIs and expert-labeled FCD maps. This training employed both classical and continual learning techniques. The CNN's performance was validated by comparing the baseline model with the trained models at two training levels.

RESULTS

In prospective validation, the best model trained using continual learning achieved a sensitivity of 90.0%, specificity of 70.0%, and accuracy of 72.0%, with an average of 0.41 false positive clusters detected per MRI. For FCD segmentation, an average Dice coefficient of 0.56 was attained. The model's performance improved in each training phase while maintaining a high level of sensitivity. Continual learning outperformed classical learning in this regard.

CONCLUSIONS

Our study presents a promising CNN for FCD detection and segmentation, exhibiting both high sensitivity and specificity. Furthermore, the model demonstrates continuous improvement with the inclusion of more clinical MRI data. We consider our CNN a valuable tool for automated, examiner-independent FCD detection in daily clinical practice, potentially addressing the underutilization of epilepsy surgery in drug-resistant focal epilepsy and thereby improving patient outcomes.

The clinical, imaging, pathological and genetic landscape of bottom-of-sulcus dysplasia

Bottom-of-sulcus dysplasia (BOSD) is increasingly recognized as a cause of drug-resistant, surgically-remediable, focal epilepsy, often in seemingly MRI-negative patients. We describe the clinical manifestations, morphological features, localization patterns and genetics of BOSD, with the aims of improving management and understanding pathogenesis. We studied 85 patients with BOSD diagnosed between 2005-2022. Presenting seizure and EEG characteristics, clinical course, genetic findings and treatment response were obtained from medical records. MRI (3 T) and 18F-FDG-PET scans were reviewed systematically for BOSD morphology and metabolism. Histopathological analysis and tissue genetic testing were performed in 64 operated patients. BOSD locations were transposed to common imaging space to study anatomical location, functional network localization and relationship to normal MTOR gene expression. All patients presented with stereotyped focal seizures with rapidly escalating frequency, prompting hospitalization in 48%. Despite 42% patients having seizure remissions, usually with sodium channel blocking medications, most eventually became drug-resistant and underwent surgery (86% seizure-free). Prior developmental delay was uncommon but intellectual, language and executive dysfunction were present in 24%, 48% and 29% when assessed preoperatively, low intellect being associated with greater epilepsy duration. BOSDs were missed on initial MRI in 68%, being ultimately recognized following repeat MRI, 18F-FDG-PET or image postprocessing. MRI features were grey-white junction blurring (100%), cortical thickening (91%), transmantle band (62%), increased cortical T1 signal (46%) and increased subcortical FLAIR signal (26%). BOSD hypometabolism was present on 18F-FDG-PET in 99%. Additional areas of cortical malformation or grey matter heterotopia were present in eight patients. BOSDs predominated in frontal and pericentral cortex and related functional networks, mostly sparing temporal and occipital cortex, and limbic and visual networks. Genetic testing yielded pathogenic mTOR pathway variants in 63% patients, including somatic MTOR variants in 47% operated patients and germline DEPDC5 or NPRL3 variants in 73% patients with familial focal epilepsy. BOSDs tended to occur in regions where the healthy brain normally shows lower MTOR expression, suggesting these regions may be more vulnerable to upregulation of MTOR activity. Consistent with the existing literature, these results highlight (i) clinical features raising suspicion of BOSD; (ii) the role of somatic and germline mTOR pathway variants in patients with sporadic and familial focal epilepsy associated with BOSD; and (iii) the role of 18F-FDG-PET alongside high-field MRI in detecting subtle BOSD. The anatomical and functional distribution of BOSDs likely explain their seizure, EEG and cognitive manifestations and may relate to relative MTOR expression.

Automated detection of focal cortical dysplasia based on magnetic resonance imaging and positron emission tomography

PURPOSE

Focal cortical dysplasia (FCD) is a common etiology of drug-resistant focal epilepsy. Visual identification of FCD is usually time-consuming and depends on personal experience. Herein, we propose an automated type II FCD detection approach utilizing multi-modal data and 3D convolutional neural network (CNN).

METHODS

MRI and positron emission tomography (PET) data of 82 patients with FCD were collected, including 55 (67.1%) histopathologically, and 27 (32.9%) radiologically diagnosed patients. Three types of morphometric feature maps and three types of tissue maps were extracted from the T1-weighted images. These maps, T1, and PET images formed the inputs for CNN. Five-fold cross-validations were carried out on the training set containing 62 patients, and the model behaving best was chosen to detect FCD on the test set of 20 patients. Furthermore, ablation experiments were performed to estimate the value of PET data and CNN.

RESULTS

On the validation set, FCD was detected in 90.3% of the cases, with an average of 1.7 possible lesions per patient. The sensitivity on the test set was 90.0%, with 1.85 possible lesions per patient. Without the PET data, the sensitivity decreased to 80.0%, and the average lesion number increased to 2.05 on the test set. If an artificial neural network replaced the CNN, the sensitivity decreased to 85.0%, and the average lesion number increased to 4.65.

SIGNIFICANCE

Automated detection of FCD with high sensitivity and few false-positive findings is feasible based on multi-modal data. PET data and CNN could improve the performance of automated detection.

Deep learning-based automated lesion segmentation on pediatric focal cortical dysplasia II preoperative MRI: a reliable approach

OBJECTIVES

Focal cortical dysplasia (FCD) represents one of the most common causes of refractory epilepsy in children. Deep learning demonstrates great power in tissue discrimination by analyzing MRI data. A prediction model was built and verified using 3D full-resolution nnU-Net for automatic lesion detection and segmentation of children with FCD II.

METHODS

High-resolution brain MRI structure data from 65 patients, confirmed with FCD II by pathology, were retrospectively studied. Experienced neuroradiologists segmented and labeled the lesions as the ground truth. Also, we used 3D full-resolution nnU-Net to segment lesions automatically, generating detection maps. The algorithm was trained using fivefold cross-validation, with data partitioned into training (N = 200) and testing (N = 15). To evaluate performance, detection maps were compared to expert manual labels. The Dice-Sørensen coefficient (DSC) and sensitivity were used to assess the algorithm performance.

RESULTS

The 3D nnU-Net showed a good performance for FCD lesion detection at the voxel level, with a sensitivity of 0.73. The best segmentation model achieved a mean DSC score of 0.57 on the testing dataset.

CONCLUSION

This pilot study confirmed that 3D full-resolution nnU-Net can automatically segment FCD lesions with reliable outcomes. This provides a novel approach to FCD lesion detection.

CRITICAL RELEVANCE STATEMENT

Our fully automatic models could process the 3D T1-MPRAGE data and segment FCD II lesions with reliable outcomes.

KEY POINTS

• Simplified image processing promotes the DL model implemented in clinical practice. • The histopathological confirmed lesion masks enhance the clinical credibility of the AI model. • The voxel-level evaluation metrics benefit lesion detection and clinical decisions.

Cortical atrophy in early-stage patients with anti-NMDA receptor encephalitis: a machine-learning MRI study with various feature extraction

Conventional brain magnetic resonance imaging (MRI) of anti-N-methyl-D-aspartate-receptor encephalitis (NMDARE) is non-specific, thus showing little differential diagnostic value, especially for MRI-negative patients. To characterize patterns of structural alterations and facilitate the diagnosis of MRI-negative NMDARE patients, we build two support vector machine models (NMDARE versus healthy controls [HC] model and NMDARE versus viral encephalitis [VE] model) based on radiomics features extracted from brain MRI. A total of 109 MRI-negative NMDARE patients in the acute phase, 108 HCs and 84 acute MRI-negative VE cases were included for training. Another 29 NMDARE patients, 28 HCs and 26 VE cases were included for validation. Eighty features discriminated NMDARE patients from HCs, with area under the receiver operating characteristic curve (AUC) of 0.963 in validation set. NMDARE patients presented with significantly lower thickness, area, and volume and higher mean curvature than HCs. Potential atrophy predominately presented in the frontal lobe (cumulative weight = 4.3725, contribution rate of 29.86%), and temporal lobe (cumulative weight = 2.573, contribution rate of 17.57%). The NMDARE versus VE model achieved certain diagnostic power, with AUC of 0.879 in validation set. Our research shows potential atrophy across the entire cerebral cortex in acute NMDARE patients, and MRI machine learning model has a potential to facilitate the diagnosis MRI-negative NMDARE.

Surface-Based Neuroimaging Pattern of Multiple System Atrophy

RATIONALE AND OBJECTIVES

Overlapping parkinsonism, cerebellar ataxia, and pyramidal signs render challenges in the clinical diagnosis of multiple system atrophy (MSA). The neuroimaging pattern is valuable to understand its pathophysiology and improve its diagnostic effect.

MATERIALS AND METHODS

We retrospectively obtained magnetic resonance imaging and susceptibility-weighted imaging in patients with MSA (including parkinsonian type [MSA-P] and cerebellar type [MSA-C]), Parkinson's disease, and normal controls. We quantified neuroimaging features to identify the optimal threshold for diagnosis. Furthermore, we explore neuroimaging patterns of MSA by mapping the subcortical morphological alterations and constructing a diagnostic model.

RESULTS

Compared to controls, normalized putaminal volume significantly decreased in patients with MSA-P (P < .001) and normalized pontine volume significantly decreased in patients with MSA-C (P < .001). The Youden index of the threshold-based clinical prediction model was 0.871-0.928 in patients with MSA. The neuroimaging pattern in patients with MSA was jointly located in the lateral putamen, and the neuroimaging pattern prediction model achieved a classification accuracy of 83.9%-100%.

CONCLUSION

The quantitative neuroimaging features and surface-based morphologic anomalies represent the markers of MSA and open new avenues for personalized clinical diagnosis.

Enhanced focal cortical dysplasia detection in pediatric frontal lobe epilepsy with asymmetric radiomic and morphological features

Objective

Focal cortical dysplasia (FCD) is the most common pathological cause for pediatric epilepsy, with frontal lobe epilepsy (FLE) being the most prevalent in the pediatric population. We attempted to utilize radiomic and morphological methods on MRI and PET to detect FCD in children with FLE.

Methods

Thirty-seven children with FLE and 20 controls were included in the primary cohort, and a five-fold cross-validation was performed. In addition, we validated the performance in an independent site of 12 FLE children. A two-stage experiments including frontal lobe and subregions were employed to detect the lesion area of FCD, incorporating the asymmetric feature between the left and right hemispheres. Specifically, for the radiomics approach, we used gray matter (GM), white matter (WM), GM and WM, and the gray-white matter boundary regions of interest to extract features. Then, we employed a Multi-Layer Perceptron classifier to achieve FCD lesion localization based on both radiomic and morphological methods.

Results

The Multi-Layer Perceptron model based on the asymmetric feature exhibited excellent performance both in the frontal lobe and subregions. In the primary cohort and independent site, the radiomics analysis with GM and WM asymmetric features had the highest sensitivity (89.2 and 91.7%) and AUC (98.9 and 99.3%) in frontal lobe. While in the subregions, the GM asymmetric features had the highest sensitivity (85.6 and 79.7%). Furthermore, relying on the highest sensitivity of GM and WM asymmetric features in frontal lobe, when integrated with the subregions results, our approach exhibited overlaps with GM asymmetric features (55.4 and 52.4%), as well as morphological asymmetric features (54.4 and 53.8%), both in the primary cohort and at the independent site.

Significance

This study demonstrates that a two-stage design based on the asymmetry of radiomic and morphological features can improve FCD detection. Specifically, incorporating regions of interest for GM, WM, GM, and WM, and the gray-white matter boundary significantly enhances the localization capabilities for lesion detection within the radiomics approach.

Deep learning in neuroimaging of epilepsy

In recent years, artificial intelligence, particularly deep learning (DL), has demonstrated utility in diverse areas of medicine. DL uses neural networks to automatically learn features from the raw data while this is not possible with conventional machine learning. It is helpful for the assessment of patients with epilepsy and whilst most published studies have been aimed at the automatic detection and prediction of seizures from electroencephalographic records, there is a growing number of investigations that use neuroimaging modalities (structural and functional magnetic resonance imaging, diffusion-weighted imaging and positron emission tomography) as input data. We review the application of DL to neuroimaging (sMRI, fMRI, DWI and PET) of focal epilepsy, specifically presurgical evaluation of drug-refractory epilepsy. First, a brief theoretical overview of artificial neural networks and deep learning is presented. Next, we review applications of deep learning to neuroimaging of epilepsy: diagnosis and lateralization, automated detection of lesion, presurgical evaluation and prediction of postsurgical outcome. Finally, the limitations, challenges and possible future directions in the application of these methods in the study of epilepsies are discussed. This approach could become an essential tool in clinical practice, particularly in the evaluation of images considered negative by visual inspection, in individualized treatments, and in the approach to epilepsy as a network disorder. However, greater multicenter collaboration is required to achieve the collection of sufficient data with the required quality together with the open access availability of the developed codes and tools.

Automatic Detection of Focal Cortical Dysplasia Using MRI: A Systematic Review

Focal cortical dysplasia (FCD) is a congenital brain malformation that is closely associated with epilepsy. Early and accurate diagnosis is essential for effectively treating and managing FCD. Magnetic resonance imaging (MRI)-one of the most commonly used non-invasive neuroimaging methods for evaluating the structure of the brain-is often implemented along with automatic methods to diagnose FCD. In this review, we define three categories for FCD identification based on MRI: visual, semi-automatic, and fully automatic methods. By conducting a systematic review following the PRISMA statement, we identified 65 relevant papers that have contributed to our understanding of automatic FCD identification techniques. The results of this review present a comprehensive overview of the current state-of-the-art in the field of automatic FCD identification and highlight the progress made and challenges ahead in developing reliable, efficient methods for automatic FCD diagnosis using MRI images. Future developments in this area will most likely lead to the integration of these automatic identification tools into medical image-viewing software, providing neurologists and radiologists with enhanced diagnostic capabilities. Moreover, new MRI sequences and higher-field-strength scanners will offer improved resolution and anatomical detail for precise FCD characterization. This review summarizes the current state of automatic FCD identification, thereby contributing to a deeper understanding and the advancement of FCD diagnosis and management.

Detecting Cortical Thickness Changes in Epileptogenic Lesions Using Machine Learning

Epilepsy is a neurological disorder characterized by abnormal brain activity. Epileptic patients suffer from unpredictable seizures, which may cause a loss of awareness. Seizures are considered drug resistant if treatment does not affect success. This leads practitioners to calculate the cortical thickness to measure the distance between the brain’s white and grey matter surfaces at various locations to perform a surgical intervention. In this study, we introduce using machine learning as an approach to classify extracted measurements from T1-weighted magnetic resonance imaging. Data were collected from the epilepsy unit at King Abdulaziz University Hospital. We applied two trials to classify the extracted measurements from T1-weighted MRI for drug-resistant epilepsy and healthy control subjects. The preprocessing sequence on T1-weighted MRI images was performed using C++ through BrainSuite’s pipeline. The first trial was performed on seven different combinations of four commonly selected measurements. The best performance was achieved in Exp6 and Exp7, with 80.00% accuracy, 83.00% recall score, and 83.88% precision. It is noticeable that grey matter volume and white matter volume measurements are more significant than the cortical thickness measurement. The second trial applied four different machine learning classifiers after applying 10-fold cross-validation and principal component analysis on all extracted measurements as in the first trial based on the mentioned previous works. The K-nearest neighbours model outperformed the other machine learning classifiers with 97.11% accuracy, 75.00% recall score, and 75.00% precision.

Interpretable surface-based detection of focal cortical dysplasias: a Multi-centre Epilepsy Lesion Detection study

One outstanding challenge for machine learning in diagnostic biomedical imaging is algorithm interpretability. A key application is the identification of subtle epileptogenic focal cortical dysplasias (FCDs) from structural MRI. FCDs are difficult to visualize on structural MRI but are often amenable to surgical resection. We aimed to develop an open-source, interpretable, surface-based machine-learning algorithm to automatically identify FCDs on heterogeneous structural MRI data from epilepsy surgery centres worldwide. The Multi-centre Epilepsy Lesion Detection (MELD) Project collated and harmonized a retrospective MRI cohort of 1015 participants, 618 patients with focal FCD-related epilepsy and 397 controls, from 22 epilepsy centres worldwide. We created a neural network for FCD detection based on 33 surface-based features. The network was trained and cross-validated on 50% of the total cohort and tested on the remaining 50% as well as on 2 independent test sites. Multidimensional feature analysis and integrated gradient saliencies were used to interrogate network performance. Our pipeline outputs individual patient reports, which identify the location of predicted lesions, alongside their imaging features and relative saliency to the classifier. On a restricted 'gold-standard' subcohort of seizure-free patients with FCD type IIB who had T1 and fluid-attenuated inversion recovery MRI data, the MELD FCD surface-based algorithm had a sensitivity of 85%. Across the entire withheld test cohort the sensitivity was 59% and specificity was 54%. After including a border zone around lesions, to account for uncertainty around the borders of manually delineated lesion masks, the sensitivity was 67%. This multicentre, multinational study with open access protocols and code has developed a robust and interpretable machine-learning algorithm for automated detection of focal cortical dysplasias, giving physicians greater confidence in the identification of subtle MRI lesions in individuals with epilepsy.

Automated Detection and Surgical Planning for Focal Cortical Dysplasia with Multicenter Validation

BACKGROUND

In patients with surgically amenable focal cortical dysplasia (FCD), subtle neuroimaging representation and the risk of open surgery lead to gaps in surgical treatment and delays in surgery.

OBJECTIVE

To construct an integrated platform that can accurately detect FCD and automatically establish trajectory planning for magnetic resonance-guided laser interstitial thermal therapy.

METHODS

This multicenter study included retrospective patients to train the automated detection model, prospective patients for model evaluation, and an additional cohort for construction of the automated trajectory planning algorithm. For automated detection, we evaluated the performance and generalization of the conventional neural network in different multicenter cohorts. For automated trajectory planning, feasibility/noninferiority and safety score were calculated to evaluate the clinical value.

RESULTS

Of the 260 patients screened for eligibility, 202 were finally included. Eighty-eight patients were selected for conventional neural network training, 88 for generalizability testing, and 26 for the establishment of an automated trajectory planning algorithm. The model trained using preprocessed and multimodal neuroimaging displayed the best performance in diagnosing FCD (figure of merit = 0.827 and accuracy range = 75.0%-91.7% across centers). None of the clinical variables had a significant effect on prediction performance. Moreover, the automated trajectory was feasible and noninferior to the manual trajectory (χ2 = 3.540, P = .060) and significantly safer (overall: test statistic = 30.423, P < .001).

CONCLUSION

The integrated platform validated based on multicenter, prospective cohorts exhibited advantages of easy implementation, high performance, and generalizability, thereby indicating its potential in the diagnosis and minimally invasive treatment of FCD.

Metabolism and Intracranial Epileptogenicity in Temporal Lobe Long-Term Epilepsy-Associated Tumor

Brain tumors are common in epilepsy surgery and frequently occur in the temporal lobe, but the optimal surgical strategies to remove the tumor and epileptogenic zone remain controversial. We aim at illustrating the positron emission tomography (PET) metabolism and the stereoelectroencephalography (SEEG) epileptogenicity of temporal lobe long-term epilepsy-associated tumors (LEAT). In this study, 70 patients and 25 healthy controls were included. Our analysis leveraged group-level analysis to reveal the whole-brain metabolic pattern of temporal lobe LEATs. The SEEG-based epileptogenicity mapping was performed to verify the PET findings in the epileptic network. Compared to controls, patients with a temporal lobe LEAT showed a more widespread epileptic network based on 18FDG-PET in patients with a mesial temporal lobe LEAT than in those with a lateral temporal lobe LEAT. The significant brain clusters mainly involved the paracentral lobule (ANOVA F = 9.731, p < 0.001), caudate nucleus (ANOVA F = 20.749, p < 0.001), putamen (Kruskal−Wallis H = 19.258, p < 0.001), and thalamus (ANOVA F = 4.754, p = 0.011). Subgroup analysis and SEEG-based epileptogenicity mapping are similar to the metabolic pattern. Our findings demonstrate the metabolic and electrophysiological organization of the temporal lobe LEAT epileptic network, which may assist in a patient-specific surgical strategy.

DeepEZ: A Graph Convolutional Network for Automated Epileptogenic Zone Localization from Resting-State fMRI Connectivity

OBJECTIVE

Epileptogenic zone (EZ) localization is a crucial step during diagnostic work up and therapeutic planning in medication refractory epilepsy. In this paper, we present the first deep learning approach to localize the EZ based on resting-state fMRI (rs-fMRI) data.

METHODS

Our network, called DeepEZ, uses a cascade of graph convolutions that emphasize signal propagation along expected anatomical pathways. We also integrate domain-specific information, such as an asymmetry term on the predicted EZ and a learned subject-specific bias to mitigate environmental confounds.

RESULTS

We validate DeepEZ on rs-fMRI collected from 14 patients with focal epilepsy at the University of Wisconsin Madison. Using cross validation, we demonstrate that DeepEZ achieves consistently high EZ localization performance (Accuracy: 0.88 ± 0.03; AUC: 0.73 ± 0.03) that far outstripped any of the baseline methods. This performance is notable given the variability in EZ locations and scanner type across the cohort.

CONCLUSION

Our results highlight the promise of using DeepEZ as an accurate and noninvasive therapeutic planning tool for medication refractory epilepsy.

SIGNIFICANCE

While prior work in EZ localization focused on identifying localized aberrant signatures, there is growing evidence that epileptic seizures affect inter-regional connectivity in the brain. DeepEZ allows clinicians to harness this information from noninvasive imaging that can easily be integrated into the existing clinical workflow.

Decoding Intracranial EEG With Machine Learning: A Systematic Review

Advances in intracranial electroencephalography (iEEG) and neurophysiology have enabled the study of previously inaccessible brain regions with high fidelity temporal and spatial resolution. Studies of iEEG have revealed a rich neural code subserving healthy brain function and which fails in disease states. Machine learning (ML), a form of artificial intelligence, is a modern tool that may be able to better decode complex neural signals and enhance interpretation of these data. To date, a number of publications have applied ML to iEEG, but clinician awareness of these techniques and their relevance to neurosurgery, has been limited. The present work presents a review of existing applications of ML techniques in iEEG data, discusses the relative merits and limitations of the various approaches, and examines potential avenues for clinical translation in neurosurgery. One-hundred-seven articles examining artificial intelligence applications to iEEG were identified from 3 databases. Clinical applications of ML from these articles were categorized into 4 domains: i) seizure analysis, ii) motor tasks, iii) cognitive assessment, and iv) sleep staging. The review revealed that supervised algorithms were most commonly used across studies and often leveraged publicly available timeseries datasets. We conclude with recommendations for future work and potential clinical applications.

Metabolic phenotyping of hand automatisms in mesial temporal lobe epilepsy

Purpose

Hand automatisms (HA) are common clinical manifestations in mesial temporal lobe epilepsy. However, the location of the symptomatogenic zone (EZ) in HA as well as the networks involved, are still unclear. To have a better understanding of HA underlying mechanisms, we analyzed images from interictal [¹⁸F] fluorodeoxyglucose-positron emission tomography (FDG-PET) in patients with mesial temporal lobe epilepsy (mTLE).

Methods

We retrospectively recruited 79 mTLE patients and 18 healthy people that substituted the control group for the analysis. All patients underwent anterior temporal lobectomy and were seizure-free. Based on the semiology of the HA occurrence, the patients were divided into three subgroups: patients with unilateral HA (Uni-HA), with bilateral HA (Bil-HA) and without HA (None-HA). We performed the intergroup comparison analysis of the interictal FDG-PET images and compared the functional connectivity within metabolic communities.

Results

Our analysis showed that the metabolic patterns varied among the different groups. The Uni-HA subgroup had significant differences in the extratemporal lobe brain areas, mostly in the ipsilateral supplementary motor area (SMA) and middle cingulate cortex (MCC) when compared to the healthy control group. The Bil-HA subgroup demonstrated that the bilateral SMA and MCC areas were differentially affected, whereas in the None-HA subgroup the differences were evident in limited brain areas. The metabolic network involving HA showed a constrained network embedding the SMA and MCC brain regions. Furthermore, the increased metabolic synchronization between SMA and MCC was significantly correlated with HA.

Conclusion

The metabolic pattern of HA was most conspicuous in SMA and MCC brain regions. Increased metabolic synchronization within SMA and MCC was considered as the major EZ of HA.

Metabolic pattern analysis allowed allocation of the symptomatogenic zone (EZ) and brain network of hand automatisms (HA) in mesial temporal lobe epilepsy (mTLE).

The involved network of bilateral HA was larger than the unilateral one, probably due to the occurrence of contralateral dystonic posturing.

Increased metabolic synchronization within supplementary motor area (SMA) and middle cingulate cortex (MCC) regions were engaged in the representation and modulation of HA, suggesting these regions as the EZ for HA.

Whole-brain metabolic pattern analysis in patients with anti-LGI1 encephalitis

BACKGROUND AND PURPOSE

Faciobrachial dystonic seizures (FBDS) and hyponatraemia are the distinct clinical features of autoimmune encephalitis (AE) caused by antibodies against leucine-rich glioma-inactivated 1 (LGI1). The pathophysiological pattern and neural mechanisms underlying these symptoms remain largely unexplored.

METHODS

We included 30 patients with anti-LGI1 AE and 30 controls from a retrospective observational cohort. Whole-brain metabolic pattern analysis was performed to assess the pathological network of anti-LGI1 AE, as well as the symptomatic networks of FBDS. Logistic regression was applied to explore independent predictors of FBDS. Finally, we applied multiple regression model to investigate the hyponatraemia-associated brain network and its effect on serum sodium levels.

RESULTS

The pathological network of anti-LGI1 AE involved a hypermetabolism in cerebellum, subcortical structures, and Rolandic area, as well as a hypometabolism in the medial prefrontal cortex. The symptomatic network of FBDS shown a hypometabolism in cerebellum and Rolandic area (P_FDR < 0.05). Hypometabolism in the cerebellum was an independent predictor of FBDS (P < 0.001). Hyponatraemia-associated network highlighted a negative effect on caudate nucleus, frontal and temporal white matter. Serum sodium level had the negative trend with metabolism of hypothalamus (Pearson's R = -0.180, P = 0.342) but the mediation was not detected (path c' = -7.238, 95% CI = -30.947 to 16.472).

CONCLUSIONS

Our results provide insights into the whole-brain metabolic patterns of patients with anti-LGI1 AE, including the symptomatic network FBDS and the hyponatraemia-associated brain network, which is conducive to understanding the neural mechanisms and evaluating disease progress of anti-LGI1 AE.

Using a recurrent neural network with S2 characteristics, efficient identification of localised cortical dysplasia

Focal cortical dysplasia (FCD) is an inborn anomaly in brain growth and morphological deformation in lesions of the brain which induces focal seizures. Neurosurgical therapies were performed for the detection of FCD. Furthermore, it can be overcome through the presurgical evaluation of epilepsy. The surgical result is attained basically through the output of the presurgical output. In preprocessing the process of increasing true positives with the decrease in false negatives occurs which results in an effective outcome. MRI (Magnetic Resonance Imaging) outputs are efficient to predict the FCD lesions through T1- MPRAGE and T2- FLAIR efficient output can be obtained. In our proposed work we extract the S2 features through the testing of T1, T2 images. Using RNN-LSTM (Recurrent neural network-Long short-term memory) test images were trained and the FCD lesions were segmented. The output of our work is compared with the proposed work yields better results compared to the existing system such as artificial neural network (ANN), support vector machine (SVM), and convolution neural network (CNN). This approach obtained an accuracy rate of 0.195% (ANN), 0.20% (SVM), 0.14% (CNN), specificity rate of 0.23% (ANN), 0.15% (SVM), 0.13% (CNN) and sensitivity rate of 0.22% (ANN), 0.14% (SVM), 0.08% (CNN) respectively in comparison with RNN-LSTM.

Neuroimaging gradient alterations and epileptogenic prediction in focal cortical dysplasia Ⅲa

INTRODUCTION

Focal cortical dysplasia Type Ⅲa (FCD Ⅲa) is a highly prevalent temporal lobe epilepsy but the seizure outcomes are not satisfactory after epilepsy surgery. Hence, quantitative neuroimaging, epileptogenic alterations, as well as their values in guiding surgery are worth exploring.

METHODS

We examined 69 patients with pathologically verified FCD Ⅲa using multimodal neuroimaging and stereoelectroencephalography (SEEG). Among them, 18 received postoperative imaging which showed the extent of surgical resection and 9 underwent SEEG implantation. We also explored neuroimaging gradient alterations along with the distance to the temporal pole. Subsequently, the machine learning regression model was employed to predict whole-brain epileptogenicity. Lastly, the correlation between neuroimaging or epileptogenicity and surgical cavities was assessed.

RESULTS

FCD Ⅲa displayed neuroimaging gradient alterations on the temporal neocortex, morphology-signal intensity decoupling, low similarity of intra-morphological features and high similarity of intra-signal intensity features. The support vector regression model was successfully applied at the whole-brain level to calculate the continuous epileptogenic value at each vertex (mean-squared error = 13.8 ± 9.8).

CONCLUSION

Our study investigated the neuroimaging gradient alterations and epileptogenicity of FCD Ⅲa, along with their potential values in guiding suitable resection range and in predicting postoperative seizure outcomes. The conclusions from this study may facilitate an accurate presurgical examination of FCD Ⅲa. However, further investigation including a larger cohort is necessary to confirm the results.

Whole-brain morphological alterations associated with trigeminal neuralgia

Background

Novel neuroimaging strategies have the potential to offer new insights into the mechanistic basis for trigeminal neuralgia (TN). The present study aims to conduct whole-brain morphometry analyses of TN patients and to assess the value of group-level neocortical and subcortical structural patterns as tools for diagnostic biomarker exploration.

Methods

Cortical thickness, surface area, and myelin levels in the neocortex were measured via magnetic resonance imaging (MRI). The radial distance and the Jacobian determinant of the subcortex in 43 TN patients and 43 matched controls were compared. Pattern learning algorithms were employed to establish the utility of group-level MRI findings as tools for predicting TN. An additional 40 control patients with hemifacial spasms were then evaluated to assess algorithm sensitivity and specificity.

Results

TN patients exhibited reductions in cortical indices in the anterior cingulate cortex (ACC), the midcingulate cortex (MCC), and the posterior cingulate cortex (PCC) relative to controls. They further presented with widespread subcortical volume reduction that was most evident in the putamen, the thalamus, the accumbens, the pallidum, and the hippocampus. Whole brain-level morphological alterations successfully enable automated TN diagnosis with high specificity (TN: 95.35 %; disease controls: 46.51 %).

Conclusions

TN is associated with a distinctive whole-brain structural neuroimaging pattern, underscoring the value of machine learning as an approach to differentiating between morphological phenotypes, ultimately revealing the full spectrum of this disease and highlighting relevant diagnostic biomarkers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s10194-021-01308-5.

Sulcus-centered resection for focal cortical dysplasia type II: surgical techniques and outcomes

Focal cortical dysplasia type II (FCD II) is a common histopathological substrate of epilepsy surgery. Here, the authors propose a sulcus-centered resection strategy for this malformation, provide technical details, and assess the efficacy and safety of this technique. The main purpose of the sulcus-centered resection is to remove the folded gray matter surrounding a dysplastic sulcus, particularly that at the bottom of the sulcus. The authors also retrospectively reviewed the records of 88 consecutive patients with FCD II treated with resective surgery between January 2015 and December 2018. The demographics, clinical characteristics, electrophysiological recordings, neuroimaging studies, histopathological findings, surgical outcomes, and complications were collected. After the exclusion of diffusely distributed and gyrus-based lesions, 71 patients (30 females, 41 males) who had undergone sulcus-centered resection were included in this study. The mean (± standard deviation) age of the cohort was 17.78 ± 10.54 years (38 pediatric patients, 33 adults). Thirty-five lesions (49%) were demonstrated on MRI; 42 patients (59%) underwent stereo-EEG monitoring before resective surgery; and 37 (52%) and 34 (48%) lesions were histopathologically proven to be FCD IIa and IIb, respectively. At a mean follow-up of 3.34 ± 1.17 years, 64 patients (90%) remained seizure free, and 7 (10%) had permanent neurological deficits including motor weakness, sensory deficits, and visual field deficits. The study findings showed that in carefully selected FCD II cases, sulcus-centered resection is an effective and safe surgical strategy.

Automatic analysis of integrated magnetic resonance and positron emission tomography images improves the accuracy of detection of focal cortical dysplasia type IIb lesions

We aimed to develop an efficient and objective pre-evaluation method to identify the precise location of a focal cortical dysplasia lesion before surgical resection to reduce medication use and decrease the post-operative frequency of seizure attacks. We developed a novel machine learning-based approach using cortical surface-based features by integrating MRI and metabolic PET to identify focal cortical dysplasia lesions. Significant surface-based features of 22 patients with histopathologically proven FCD IIb lesions were extracted from PET and MRI images using FreeSurfer. We modified significant parameters, trained and tested the XGBoost model using these surface-based features, and made predictions. We detected lesions in all 20 patients using the XGBoost model, with an accuracy of 91%. We used one-way chi-squared test to test the null hypothesis that the population proportion was 50% (p = 0.0001), indicating that our classification of the algorithm was statistically significant. The sensitivity, specificity, and false-positive rates were 93%, 91%, and 9%, respectively. We developed an objective, quantitative XGBoost classifier that combined MRI and PET imaging features to locate focal cortical dysplasia. This automated method yielded better outcomes than conventional visual analysis and single modality quantitative analysis for surgical pre-evaluation, especially in subtle or visually unidentifiable FCD lesions. This time-efficient method would also help doctors identify otherwise overlooked details.

External validation of automated focal cortical dysplasia detection using morphometric analysis

OBJECTIVE

Focal cortical dysplasias (FCDs) are a common cause of drug-resistant focal epilepsy but frequently remain undetected by conventional magnetic resonance imaging (MRI) assessment. The visual detection can be facilitated by morphometric analysis of T1-weighted images, for example, using the Morphometric Analysis Program (v2018; MAP18), which was introduced in 2005, independently validated for its clinical benefits, and successfully integrated in standard presurgical workflows of numerous epilepsy centers worldwide. Here we aimed to develop an artificial neural network (ANN) classifier for robust automated detection of FCDs based on these morphometric maps and probe its generalization performance in a large, independent data set.

METHODS

In this retrospective study, we created a feed-forward ANN for FCD detection based on the morphometric output maps of MAP18. The ANN was trained and cross-validated on 113 patients (62 female, mean age ± SD =29.5 ± 13.6 years) with manually segmented FCDs and 362 healthy controls (161 female, mean age ± SD =30.2 ± 9.6 years) acquired on 13 different scanners. In addition, we validated the performance of the trained ANN on an independent, unseen data set of 60 FCD patients (28 female, mean age ± SD =30 ± 15.26 years) and 70 healthy controls (42 females, mean age ± SD = 40.0 ± 12.54 years).

RESULTS

In the cross-validation, the ANN achieved a sensitivity of 87.4% at a specificity of 85.4% on the training data set. On the independent validation data set, our method still reached a sensitivity of 81.0% at a comparably high specificity of 84.3%.

SIGNIFICANCE

Our method shows a robust automated detection of FCDs and performance generalizability, largely independent of scanning site or MR-sequence parameters. Taken together with the minimal input requirements of a standard T1 image, our approach constitutes a clinically viable and useful tool in the presurgical diagnostic routine for drug-resistant focal epilepsy.

Automated detection and segmentation of focal cortical dysplasias (FCDs) with artificial intelligence: Presentation of a novel convolutional neural network and its prospective clinical validation

PURPOSE

Focal cortical dysplasias (FCDs) represent one of the most frequent causes of pharmaco-resistant focal epilepsies. Despite improved clinical imaging methods over the past years, FCD detection remains challenging, as FCDs vary in location, size, and shape and commonly blend into surrounding tissues without clear definable boundaries. We developed a novel convolutional neural network for FCD detection and segmentation and validated it prospectively on daily-routine MRIs.

MATERIAL AND METHODS

The neural network was trained on 201 T1 and FLAIR 3 T MRI volume sequences of 158 patients with mainly FCDs, regardless of type, and 7 focal PMG. Non-FCD/PMG MRIs, drawn from 100 normal MRIs and 50 MRIs with non-FCD/PMG pathologies, were added to the training. We applied the algorithm prospectively on 100 consecutive MRIs of patients with focal epilepsy from daily clinical practice. The results were compared with corresponding neuroradiological reports and morphometric MRI analyses evaluated by an experienced epileptologist.

RESULTS

Best training results reached a sensitivity (recall) of 70.1 % and a precision of 54.3 % for detecting FCDs. Applied on the daily-routine MRIs, 7 out of 9 FCDs were detected and segmented correctly with a sensitivity of 77.8 % and a specificity of 5.5 %. The results of conventional visual analyses were 33.3 % and 94.5 %, respectively (3/9 FCDs detected); the results of morphometric analyses with overall epileptologic evaluation were both 100 % (9/9 FCDs detected) and thus served as reference.

CONCLUSION

We developed a 3D convolutional neural network with autoencoder regularization for FCD detection and segmentation. Our algorithm employs the largest FCD training dataset to date with various types of FCDs and some focal PMG. It provided a higher sensitivity in detecting FCDs than conventional visual analyses. Despite its low specificity, the number of false positively predicted lesions per MRI was lower than with morphometric analysis. We consider our algorithm already useful for FCD pre-screening in everyday clinical practice.

Automatic Detection of Focal Cortical Dysplasia Type II in MRI: Is the Application of Surface-Based Morphometry and Machine Learning Promising?

Background and Objectives

Focal cortical dysplasia (FCD) is a type of malformations of cortical development and one of the leading causes of drug-resistant epilepsy. Postoperative results improve the diagnosis of lesions on structural MRIs. Advances in quantitative algorithms have increased the identification of FCD lesions. However, due to significant differences in size, shape, and location of the lesion in different patients and a big deal of time for the objective diagnosis of lesion as well as the dependence of individual interpretation, sensitive approaches are required to address the challenge of lesion diagnosis. In this research, a FCD computer-aided diagnostic system to improve existing methods is presented.

Methods

Magnetic resonance imaging (MRI) data were collected from 58 participants (30 with histologically confirmed FCD type II and 28 without a record of any neurological prognosis). Morphological and intensity-based features were calculated for each cortical surface and inserted into an artificial neural network. Statistical examinations evaluated classifier efficiency.

Results

Neural network evaluation metrics—sensitivity, specificity, and accuracy—were 96.7, 100, and 98.6%, respectively. Furthermore, the accuracy of the classifier for the detection of the lobe and hemisphere of the brain, where the FCD lesion is located, was 84.2 and 77.3%, respectively.

Conclusion

Analyzing surface-based features by automated machine learning can give a quantitative and objective diagnosis of FCD lesions in presurgical assessment and improve postsurgical outcomes.

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

Objective

To compare the location of suspect lesions detected by computational analysis of multimodal magnetic resonance imaging data with areas of seizure onset, early propagation, and interictal epileptiform discharges (IEDs) identified with stereoelectroencephalography (SEEG) in a cohort of patients with medically refractory focal epilepsy and radiologically normal magnetic resonance imaging (MRI) scans.

Methods

We developed a method of lesion detection using computational analysis of multimodal MRI data in a cohort of 62 control subjects, and 42 patients with focal epilepsy and MRI‐visible lesions. We then applied it to detect covert lesions in 27 focal epilepsy patients with radiologically normal MRI scans, comparing our findings with the areas of seizure onset, early propagation, and IEDs identified at SEEG.

Results

Seizure‐onset zones (SoZs) were identified at SEEG in 18 of the 27 patients (67%) with radiologically normal MRI scans. In 11 of these 18 cases (61%), concordant abnormalities were detected by our method. In the remaining seven cases, either early seizure propagation or IEDs were observed within the abnormalities detected, or there were additional areas of imaging abnormalities found by our method that were not sampled at SEEG. In one of the nine patients (11%) in whom SEEG was inconclusive, an abnormality, which may have been involved in seizures, was identified by our method and was not sampled at SEEG.

Significance

Computational analysis of multimodal MRI data revealed covert abnormalities in the majority of patients with refractory focal epilepsy and radiologically normal MRI that co‐located with SEEG defined zones of seizure onset. The method could help identify areas that should be targeted with SEEG when considering epilepsy surgery.

Deep Learning Model for the Automated Detection and Histopathological Prediction of Meningioma

The volumetric assessment and accurate grading of meningiomas before surgery are highly relevant for therapy planning and prognosis prediction. This study was to design a deep learning algorithm and evaluate the performance in detecting meningioma lesions and grade classification. In total, 5088 patients with histopathologically confirmed meningioma were retrospectively included. The pyramid scene parsing network (PSPNet) was trained to automatically detect and delineate the meningiomas. The results were compared to manual segmentations by evaluating the mean intersection over union (mIoU). The performance of grade classification was evaluated by accuracy. For the automated detection and segmentation of meningiomas, the mean pixel accuracy, tumor accuracy, background accuracy and mIoU were 99.68%, 81.36%, 99.88% and 81.36% for all patients; 99.52%, 84.86%, 99.93% and 84.86% for grade I meningiomas; 99.57%, 80.11%, 99.92% and 80.12% for grade II meningiomas; and 99.75%, 78.40%, 99.99% and 78.40% for grade III meningiomas, respectively. For grade classification, the accuracy values of the training and test datasets were 99.93% and 81.52% for all patients; 99.98% and 98.51% for grade I meningiomas; 99.91% and 66.67% for grade II meningiomas; and 99.88% and 73.91% for grade III meningiomas, respectively. The automated detection, segmentation and grade classification of meningiomas based on deep learning were accurate and reliable and may improve the monitoring and treatment of this frequently occurring tumor entity. Furthermore, the method could function as a useful tool for preassessment and preselection for radiologists, offering auxiliary information for clinical decision making in presurgical evaluation.

Planning stereoelectroencephalography using automated lesion detection: Retrospective feasibility study

Objective

This retrospective, cross‐sectional study evaluated the feasibility and potential benefits of incorporating deep‐learning on structural magnetic resonance imaging (MRI) into planning stereoelectroencephalography (sEEG) implantation in pediatric patients with diagnostically complex drug‐resistant epilepsy. This study aimed to assess the degree of colocalization between automated lesion detection and the seizure onset zone (SOZ) as assessed by sEEG.

Methods

A neural network classifier was applied to cortical features from MRI data from three cohorts. (1) The network was trained and cross‐validated using 34 patients with visible focal cortical dysplasias (FCDs). (2) Specificity was assessed in 20 pediatric healthy controls. (3) Feasibility of incorporation into sEEG implantation plans was evaluated in 34 sEEG patients. Coordinates of sEEG contacts were coregistered with classifier‐predicted lesions. sEEG contacts in seizure onset and irritative tissue were identified by clinical neurophysiologists. A distance of <10 mm between SOZ contacts and classifier‐predicted lesions was considered colocalization.

Results

In patients with radiologically defined lesions, classifier sensitivity was 74% (25/34 lesions detected). No clusters were detected in the controls (specificity = 100%). Of the total 34 sEEG patients, 21 patients had a focal cortical SOZ, of whom eight were histopathologically confirmed as having an FCD. The algorithm correctly detected seven of eight of these FCDs (86%). In patients with histopathologically heterogeneous focal cortical lesions, there was colocalization between classifier output and SOZ contacts in 62%. In three patients, the electroclinical profile was indicative of focal epilepsy, but no SOZ was localized on sEEG. In these patients, the classifier identified additional abnormalities that had not been implanted.

Significance

There was a high degree of colocalization between automated lesion detection and sEEG. We have created a framework for incorporation of deep‐learning–based MRI lesion detection into sEEG implantation planning. Our findings support the prospective evaluation of automated MRI analysis to plan optimal electrode trajectories.