Voxel-Based Morphometry-from Hype to Hope. A Study on Hippocampal Atrophy in Mesial Temporal Lobe Epilepsy

Relationship between visuoperceptual functions and parietal structural abnormalities in temporal lobe epilepsy

Progressive gray matter volume reductions beyond the epileptogenic area has been described in temporal lobe epilepsy. There is less evidence regarding correlations between gray and white matter volume changepres and multi-domain cognitive performance in this setting. We aimed to investigate correlations between volume changes in parietal structures and visuospatial performance in temporal lobe epilepsy patients. we performed a cross-sectional study comparing global and regional brain volume data from 34 temporal lobe epilepsy patients and 30 healthy controls. 3D T1-weighted sequences were obtained on a 3.0 T magnet, and data were analyzed using age and sex-adjusted linear regression models. Global and regional brain volumes and cortical thickness in patients were correlated with standardized visual memory, visuoperceptual, visuospatial, and visuoconstructive parameters obtained in a per-protocol neuropsychological assessment. temporal lobe epilepsy patients had smaller volume fractions of the deep gray matter structures, putamen and nucleus accumbens, and larger cerebrospinal fluid volume fraction than controls. Correlations were found between: 1) visual memory and precuneus and inferior parietal cortical thickness; 2) visuoperceptual performance and precuneus and supramarginal white matter volumes; 3) visuospatial skills and precuneus, postcentral, and inferior and superior parietal white matter volumes; 4) visuoconstructive performance and inferior parietal white matter volume. Brain volume loss is widespread in temporal lobe epilepsy. Volumetric reductions in parietal lobe structures were associated with visuoperceptual cognitive performance.

[MR morphometry in epileptology: progress and perspectives]

Morphometric MRI analysis improves neuroimaging of structural changes in epilepsy.

OBJECTIVE

To investigate diagnostic potential of MR brain morphometry in neurosurgical epileptology.

MATERIAL AND METHODS

An interdisciplinary working group reviewed the studies devoted to MR morphometry in epileptology as a part of state assignment No. 056-00119-22-00. Study subject was trials of MR-morphometry in epilepsy. Searching for literature data was conducted in international and national databases between 2017 and 2022 using certain keywords. Final analysis included 36 publications.

RESULTS

Currently, MR brain morphometry allows measurement of cortical volume and thickness, surface area and depth of furrows, as well as analysis of cortical tortuosity and fractal changes. In neurosurgical epileptology, MR-morphometry has the greatest diagnostic value in MR-negative epilepsy. This method simplifies preoperative diagnosis and reduces costs.

CONCLUSION

Morphometry in neurosurgical epileptology is an additional method for verifying the epileptogenic zone. Automated programs simplify application of this method.

Interpretable deep learning-based hippocampal sclerosis classification

OBJECTIVE

To evaluate the performance of a deep learning model for hippocampal sclerosis classification on the clinical dataset and suggest plausible visual interpretation for the model prediction.

METHODS

T2-weighted oblique coronal images of the brain MRI epilepsy protocol performed on patients were used. The training set included 320 participants with 160 no, 100 left and 60 right hippocampal sclerosis, and cross-validation was implemented. The test set consisted of 302 participants with 252 no, 25 left and 25 right hippocampal sclerosis. As the test set was imbalanced, we took an average of the accuracy achieved within each group to measure a balanced accuracy for multiclass and binary classifications. The dataset was composed to include not only healthy participants but also participants with abnormalities besides hippocampal sclerosis in the control group. We visualized the reasons for the model prediction using the layer-wise relevance propagation method.

RESULTS

When evaluated on the validation of the training set, we achieved multiclass and binary classification accuracy of 87.5% and 88.8% from the voting ensemble of six models. Evaluated on the test sets, we achieved multiclass and binary classification accuracy of 91.5% and 89.76%. The distinctly sparse visual interpretations were provided for each individual participant and group to suggest the contribution of each input voxel to the prediction on the MRI.

SIGNIFICANCE

The current interpretable deep learning-based model is promising for adapting effectively to clinical settings by utilizing commonly used data, such as MRI, with realistic abnormalities faced by neurologists to support the diagnosis of hippocampal sclerosis with plausible visual interpretation.

Deep learning-based diagnosis of temporal lobe epilepsy associated with hippocampal sclerosis: An MRI study

PURPOSE

The currently available indicators-sensitivity and specificity of expert radiological evaluation of MRIs-to identify mesial temporal lobe epilepsy (MTLE) associated with hippocampal sclerosis (HS) are deficient, as they cannot be easily assessed. We developed and investigated the use of a novel convolutional neural network trained on preoperative MRIs to aid diagnosis of these conditions.

SUBJECTS AND METHODS

We enrolled 141 individuals: 85 with clinically diagnosed mesial temporal lobe epilepsy (MTLE) and hippocampal sclerosis International League Against Epilepsy (HS ILAE) type 1 who had undergone anterior temporal lobe hippocampectomy were assigned to the MTLE-HS group, and 56 epilepsy clinic outpatients diagnosed as nonepileptic were assigned to the normal group. We fine-tuned a modified CNN (mCNN) to classify the fully connected layers of ImageNet-pretrained VGG16 network models into the MTLE-HS and control groups. MTLE-HS was diagnosed using MRI both by the fine-tuned mCNN and epilepsy specialists. Their performances were compared.

RESULTS

The fine-tuned mCNN achieved excellent diagnostic performance, including 91.1% [85%, 96%] mean sensitivity and 83.5% [75%, 91%] mean specificity. The area under the resulting receiver operating characteristic curve was 0.94 [0.90, 0.98] (DeLong's method). Expert interpretation of the same image data achieved a mean sensitivity of 73.1% [65%, 82%] and specificity of 66.3% [50%, 82%]. These confidence intervals were located entirely under the receiver operating characteristic curve of the fine-tuned mCNN.

CONCLUSIONS

Deep learning-based diagnosis of MTLE-HS from preoperative MR images using our fine-tuned mCNN achieved a performance superior to the visual interpretation by epilepsy specialists. Our model could serve as a useful preoperative diagnostic tool for ascertaining hippocampal atrophy in patients with MTLE.

Histological Correlates of Neuroanatomical Changes in a Rat Model of Levodopa-Induced Dyskinesia Based on Voxel-Based Morphometry

Long-term therapy with levodopa (L-DOPA) in patients with Parkinson’s disease (PD) often triggers motor complications termed as L-DOPA-induced dyskinesia (LID). However, few studies have explored the pathogenesis of LID from the perspective of neuroanatomy. This study aimed to investigate macroscopic structural changes in a rat model of LID and the underlying histological mechanisms. First, we established the hemiparkinsonism rat model through stereotaxic injection of 6-hydroxydopamine (6-OHDA) into the right medial forebrain bundle, followed by administration of saline (PD) or L-DOPA to induce LID. Magnetic resonance imaging (MRI) and behavioral evaluations were performed at different time points. Histological analysis was conducted to assess the correlations between MRI signal changes and cellular contributors. Voxel-based morphometry (VBM) analysis revealed progressive bilateral volume reduction in the cortical and subcortical areas in PD rats compared with the sham rats. These changes were partially reversed by chronic L-DOPA administration; moreover, there was a significant volume increase mainly in the dorsolateral striatum, substantia nigra, and piriform cortex of the lesioned side compared with that of PD rats. At the striatal cellular level, glial fibrillary acidic protein-positive (GFAP+) astrocytes were significantly increased in the lesioned dorsolateral striatum of PD rats compared with the intact side and the sham group. Prolonged L-DOPA treatment further increased GFAP levels. Neither 6-OHDA damage nor L-DOPA treatment influenced the striatal expression of vascular endothelial growth factor (VEGF). Additionally, there was a considerable increase in synapse-associated proteins (SYP, PSD95, and SAP97) in the lesioned striatum of LID rats relative to the PD rats. Golgi-Cox staining analysis of the dendritic spine morphology revealed an increased density of dendritic spines after chronic L-DOPA treatment. Taken together, our findings suggest that striatal volume changes in LID rats involve astrocyte activation, enrichment of synaptic ultrastructure and signaling proteins in the ipsilateral striatum. Meanwhile, the data highlight the enormous potential of structural MRI, especially VBM analysis, in determining the morphological phenotype of rodent models of LID.

MRI-Based Machine Learning Prediction Framework to Lateralize Hippocampal Sclerosis in Patients With Temporal Lobe Epilepsy

BACKGROUND AND OBJECTIVES

MRI fails to reveal hippocampal pathology in 30% to 50% of temporal lobe epilepsy (TLE) surgical candidates. To address this clinical challenge, we developed an automated MRI-based classifier that lateralizes the side of covert hippocampal pathology in TLE.

METHODS

We trained a surface-based linear discriminant classifier that uses T1-weighted (morphology) and T2-weighted and fluid-attenuated inversion recovery (FLAIR)/T1 (intensity) features. The classifier was trained on 60 patients with TLE (mean age 35.6 years, 58% female) with histologically verified hippocampal sclerosis (HS). Images were deemed to be MRI negative in 42% of cases on the basis of neuroradiologic reading (40% based on hippocampal volumetry). The predictive model automatically labeled patients as having left or right TLE. Lateralization accuracy was compared to electroclinical data, including side of surgery. Accuracy of the classifier was further assessed in 2 independent TLE cohorts with similar demographics and electroclinical characteristics (n = 57, 58% MRI negative).

RESULTS

The overall lateralization accuracy was 93% (95% confidence interval 92%-94%), regardless of HS visibility. In MRI-negative TLE, the combination of T2 and FLAIR/T1 intensities provided the highest accuracy in both the training (84%, area under the curve [AUC] 0.95 ± 0.02) and validation (cohort 1 90%, AUC 0.99; cohort 2 76%, AUC 0.94) cohorts.

DISCUSSION

This prediction model for TLE lateralization operates on readily available conventional MRI contrasts and offers gain in accuracy over visual radiologic assessment. The combined contribution of decreased T1- and increased T2-weighted intensities makes the synthetic FLAIR/T1 contrast particularly effective in MRI-negative HS, setting the basis for broad clinical translation.

CLASSIFICATION OF EVIDENCE

This study provides Class II evidence that in people with TLE and MRI-negative HS, an automated MRI-based classifier accurately determines the side of pathology.

Automated volumetry of hippocampal subfields in temporal lobe epilepsy

INTRODUCTION

Hippocampal sclerosis is the most frequent pathological substrate in drug resistant temporal lobe epilepsy (TLE). Recently 4 types of hippocampal sclerosis (HS) have been defined in a task force by the International League Against Epilepsy (ILAE), based on patterns of cell loss in specific hippocampal subfields. Type 1 HS is most frequent and has the most favorable outcome after epilepsy surgery. We hypothesized that volume loss in specific hippocampal subfields determined by automated volumetry of high resolution MRI would correspond to cell loss in histological reports.

MATERIAL AND METHODS

In a group of well characterized patients with drug resistant TLE (N = 26 patients, 14 with right-sided focus, 12 with left-sided focus) volumes of the right and left hippocampus and the hippocampal subfields CA1, CA2 + 3, CA4 and dentate gyrus (DG) were estimated automatically using FreeSurfer version 6.0 from high-resolution cerebral MRI and compared to a large group of healthy controls (N = 121). HS subtype classification was attempted based on histological reports.

RESULTS

Volumes of the whole hippocampus and all investigated hippocampal subfields (CA1, CA2 + 3, CA4 and DG) were significantly lower on the ipsilateral compared the contralateral side (p < 0.001) and compared to the healthy controls (p < 0.001). Conversely, whole hippocampal and hippocampal subfield volumes were not significantly different from healthy control values on the contralateral side. In 12 of 20 patients the pattern of hippocampal volume loss in specific subfields was in accordance with HS types from histology. The highest overlap between automated MRI and histology was achieved for type 1 HS (in 10 of 12 cases).

CONCLUSION

The automated volumetry of hippocampal subfields, based on high resolution MRI, may have the potential to predict the pattern of cell loss in hippocampal sclerosis before operation.