Diagnostic Performance of MRI Volumetry in Epilepsy Patients With Hippocampal Sclerosis Supported Through a Random Forest Automatic Classification Algorithm

Segmentation of hippocampal subregion via 5-Tesla magnetic resonance imaging: a comparative study

Background

Small voxel sizes in two-dimensional (2D) hippocampal magnetic resonance imaging (MRI) facilitate subfield segmentation. However, thinner slices remain challenging despite the advancements in in-plane resolution due to the tradeoffs in image quality. A higher signal-to-noise ratio (SNR) at 5 Tesla (T) compared to 3 T could enable thinner slices; however, the effects of changes in tissue contrast and inhomogeneous B1 variation may be weaken the SNR and contrast-to-noise ratio (CNR). However, the effectiveness of current automatic segmentation tools in handling 5-T images, particularly for thinner slices, remains to be fully elucidated. This study thus aimed to assess hippocampus image quality improvement from 3 to 5 T and to evaluate the autosegmentation of the hippocampal subregion on 5 T with different slice thicknesses.

Methods

This prospective study included 28 healthy participants (14 females, with a mean age of 31.8±8.96 years). T2-weighted series with 2-, 1-, and 0.7-mm slice thicknesses were acquired on 3 T and 5T. CNR, SNR, and visual scores of the molecular layer were compared. Hippocampal subregions were segmented with 5-T data using FreeSurfer and HippUnfold toolboxes. Dice coefficients of stratum radiatum, lacunosum, and moleculare (SRLM) were compared between FreeSurfer and HippUnfold. Fleiss' Kappa, the Kruskal-Wallis test, and the pairwise Wilcoxon signed-rank test were used for statistical analyses, with a P value <0.05 being considered statistically significant.

Results

SNR, CNR, and visual scores were higher on 5 T than on 3 T at the same slice thickness. The SNR of the 1-mm T2 images on 5 T was comparable to that of 2-mm images on 3 T, and the CNR of the 0.7-mm images on 5 T was comparable to that of the 2-mm images on 3 T. With a reduction in slice thickness, the Dice coefficients of SRLM between HippUnfold and manual segmentation increased, while those between FreeSurfer and manual segmentation decreased.

Conclusions

The 5-T magnetic resonance system offers higher SNR and CNR values for hippocampal imaging as compared to 3-T imaging. A slice thickness of 0.7-1 mm is recommended for 2D T2-weighted imaging. However, in the selection of automatic segmentation tools for hippocampal subregions, caution is advised if the slice thickness is less than 2 mm.

The effect of hippocampal subfield volumes on cognitive decline and incident dementia in a memory clinic cohort

BackgroundThe hippocampus plays a central role in cognition and hippocampal atrophy is a key hallmark of Alzheimer's disease. Evidence has suggested associations between hippocampal subfield volumes and specific cognitive domains and dementia risk. However, to our knowledge, no study has examined the role of hippocampal subfield volumes in cognitive decline across different domains over time.ObjectiveWe investigated associations between hippocampal subfield volumes and changes in cognitive domains together with incident dementia in a memory clinic cohort.MethodsAssociations between hippocampal subfield volumes and cognitive decline over three years (n = 443) were analyzed using generalized estimating equations, and associations with incident dementia (n = 283) using multiple logistic regression.ResultsAt baseline, all hippocampal subfield volumes were associated with diagnosis of dementia, while the CA4-dentate gyrus, molecular layer, subicular complex, and fimbria volumes were associated with diagnosis of CIND. Over three years, all subfields except the hippocampal fissure were associated with memory. Decreased molecular layer (OR:2.26, 95%CI:1.50;3.50) size was associated with increased risk of dementia.ConclusionsOur findings suggest that hippocampal atrophy of the cornu ammonis, CA4-dentate gyrus, and molecular layer may first manifest with cognitive impairment in memory before other subfields of the hippocampus, and that molecular layer volume may be an early biomarker of dementia. Further research demonstrating the biological role of hippocampal subfields in specific cognitive domains is required.

Morphometric evaluation of the human hippocampus and hippocampal subfield volume characteristics by VolBrain/HIPS

In the literature, research often utilizes digital segmentation methods like VolBrain, aiding in anatomical understanding. However, racial and population-specific differences remain unclear but may be important for clinical interpretation. We seek to enhance clinical perspectives by providing normative data and insights into hippocampal morphology, potentially advancing diagnostic and prognostic methodologies in neurological disorders. We think that it is necessary to discuss the reported asymmetry data by VolBrain. There might be inaccuracies or inconsistencies in how asymmetry is measured and reported. MRIs of the 138 healthy individuals (66 females, 72 males) were included. The VolBrain-HIPS pipeline was preferred for automatic segmentation. Alternative methods were recruited, including the total hippocampal volume for the adjustment of the volumetric data. Asymmetry index was accepted as the indicator of the magnitude of the asymmetry and negative asymmetry index data was positivized. Gender and side comparisons of the hippocampal data were presented not only for absolute measurement but also for the adjusted volumetric data. Comparisons for the magnitude of the asymmetry were carried out as well as the regression modeling based on age and gender. Hippocampus-adjusted volume information should also be considered when analyzing VolBrain data and making clinical or anatomical decisions. The asymmetry data produced by VolBrain should be considered as a magnitude scale of the asymmetry in individuals with right or left dominant hippocampus or hippocampus subfield. Alternative significant regression models were introduced to observe how the volumetric numerical composition of the hippocampus subfields changes with age.

Exploring Cortical and Hippocampal Changes in Temporal Lobe Epilepsy Using Automated MRI Segmentation Techniques

Background

To investigate the applicability of MR-based automated segmentation techniques in evaluating cortical and hippocampal changes in adults with temporal lobe epilepsy (TLE), specifically emphasizing the affected hemisphere.

Methods

A retrospective analysis involved 48 cases diagnosed with TLE based on clinical and EEG criteria. The cohort comprised 30 patients with hippocampal sclerosis (HS) and 18 with nonlesional temporal lobe epilepsy (TLE-NL) on MR. 30 healthy volunteers constituted the control group. FreeSurfer software facilitated the segmentation of cortical regions and hippocampal subfields, generating numerical values for cortical thickness and hippocampal subfield volumes on the left hemisphere. Independent sample Wilcoxon rank-sum tests enabled pairwise comparisons of cortical thickness and hippocampal subfield volumes between the control, TLE-NL, and HS groups.

Results

Significant differences emerged in hippocampal total volume and volumes of the head, body, and tail regions between the control and HS groups and the TLE-NL and HS groups. Cortical thickness of 6 regions exhibited statistical differences between the control and TLE-NL groups, while 15 regions showed distinctions between the control and HS groups. 2 regions displayed variations in cortical thickness between the TLE-NL and HS groups.

Conclusion

MRI-based automated segmentation techniques provide valuable insights into cortical and hippocampal structural variations in distinct TLE subtypes. This methodology effectively delineates changes in cortical regions and hippocampal subfields, augmenting clinical comprehension of TLE progression.

Machine learning localization to identify the epileptogenic side in mesial temporal lobe epilepsy

BACKGROUND

Mesial temporal sclerosis (MTS) is the most common pathology associated with drug-resistant mesial temporal lobe epilepsy (mTLE) in adults. Most atrophic hippocampi can be identified using MRI based on standard epilepsy protocols; however, difficulties can arise in cases where sclerotic changes in the hippocampus are subtle or non-epilepsy-specific protocols have been implemented. In such cases, quantitative methods, such as T1-weighted axial series MRIs, are valuable additional tools to complement epilepsy-specific protocols. In the current study, we applied machine learning (ML) techniques to the analysis of brain regions of interest (ROIs), including the hippocampus, thalamus, and cortical areas, to enhance the accuracy of lesion lateralization in MRI.

METHODS

This study included 104 patients diagnosed with mTLE, including 55 with lesions on the right side and 49 with lesions on the left side. FreeSurfer software was used to extract features from high-resolution T1-weighted axial brain MRI scans for use in computing lateralization indices (LI) for various brain regions. After using feature selection to pinpoint critical ROIs, the corresponding LI values were used as parameters in training the ML model.

RESULTS

The proposed ML model demonstrated exceptional performance in the lateralization of mTLE, achieving test accuracy of 92.38 % with an AUROC of 0.97.

CONCLUSION

This study demonstrated the efficacy of ML in detecting instances of MTS from thin-slice T1 images. The proposed method provides valuable insights for surgical planning and treatment. Nonetheless, additional research will be required to enhance the robustness of the model and rigorously validate its effectiveness and applicability in clinical settings.

Glucose Metabolism of Hippocampal Subfields in Medial Temporal Lobe Epilepsy

PURPOSE

Reduced glucose metabolism in the hippocampus is commonly observed in cases of medial temporal lobe epilepsy (MTLE) with hippocampal sclerosis (HS). Glucose metabolism among the various hippocampal subfields has not been thoroughly investigated.

PATIENTS AND METHODS

This study examined 29 patients (18 females; 15-58 years) diagnosed with HS who underwent surgery for drug-resistant epilepsy. FreeSurfer 7.1.1 was used in the processing of MRI data and 18 F-FDG PET scans to derive volumetric data and the FDG SUVr in the whole hippocampus and hippocampal subfields, including the CA1, CA2-4, granule cell and molecular layer of the dentate gyrus (GC-ML-DG), and subiculum. Asymmetries in the volume and SUVr between the 2 sides from the subfields of the hippocampus were defined in terms of an asymmetry index. Comparisons of the asymmetry index among these regions were performed. The correlations between asymmetry index values and postoperative outcomes and presurgical neuropsychological test results were also evaluated.

RESULT

The CA1, CA2-4, subiculum, GC-ML-DG, and whole hippocampus presented reductions in volume and hypometabolism ipsilateral to MTLE. Asymmetries in volume and SUVr were significantly less pronounced in the CA1 and subiculum than in the CA2-4 or GC-ML-DG. Postoperative seizure outcomes were not correlated with the asymmetry index for volume or SUVr in any hippocampal subfield. In cases of left MTLE, scores of immediate logical memory and delayed logical memory were positively correlated with the asymmetry index for SUVr in the following subfields: CA1 ( R = 0.829, P = 0.021; R = 0.770, P = 0.043), CA2-4 ( R = 0.825, P = 0.022; R = 0.894, P = 0.007), subiculum ( R = 0.882, P = 0.009; R = 0.853, P = 0.015), GC-ML-DG ( R = 0.850, P = 0.015; R = 0.796, P = 0.032), and whole hippocampus ( R = 0.841, P = 0.018; R = 0.822, P = 0.023). In cases of right MTLE, the scores for delayed face memory were positively correlated with the asymmetry index for SUVr in the subiculum ( R = 0.935, P = 0.006).

CONCLUSIONS

In cases of HS, changes in glucose metabolism levels varied among the hippocampal subfields. Asymmetries in glucose metabolism among the CA-1, CA2-4, subiculum, and GC-ML-DG subregions were correlated with scores for verbal memory among patients with left MTLE. Asymmetric glucose metabolism in the subiculum was also correlated with visual memory scores among patients with right MTLE.

Clinical Evaluation of a Quantitative Imaging Biomarker Supporting Radiological Assessment of Hippocampal Sclerosis

OBJECTIVE

To evaluate the influence of quantitative reports (QReports) on the radiological assessment of hippocampal sclerosis (HS) from MRI of patients with epilepsy in a setting mimicking clinical reality.

METHODS

The study included 40 patients with epilepsy, among them 20 with structural abnormalities in the mesial temporal lobe (13 with HS). Six raters blinded to the diagnosis assessed the 3T MRI in two rounds, first using MRI only and later with both MRI and the QReport. Results were evaluated using inter-rater agreement (Fleiss' kappa [Formula: see text]) and comparison with a consensus of two radiological experts derived from clinical and imaging data, including 7T MRI.

RESULTS

For the primary outcome, diagnosis of HS, the mean accuracy of the raters improved from 77.5% with MRI only to 86.3% with the additional QReport (effect size [Formula: see text]). Inter-rater agreement increased from [Formula: see text] to [Formula: see text]. Five of the six raters reached higher accuracies, and all reported higher confidence when using the QReports.

CONCLUSION

In this pre-use clinical evaluation study, we demonstrated clinical feasibility and usefulness as well as the potential impact of a previously suggested imaging biomarker for radiological assessment of HS.

Deep learning and radiomics based automatic diagnosis of hippocampal sclerosis

Accurate and rapid segmentation of the hippocampus can help doctors perform intractable temporal lobe epilepsy (TLE) preoperative evaluations to identify good surgical candidates. This study aims to establish a radiomics system for the automatic diagnosis of hippocampal sclerosis with the help of machine learning. A total of 240 cases were analysed to develop a diagnostic model. First, an automatic hippocampal segmentation process was established that exploits a priori knowledge of the relatively fixed location of the hippocampus in brain partitions, as well as a deep-learning segmentation network based on an Attention U-net. Then, we extracted 527 radiomics features from each side of the segmented hippocampus. The iterative sparse representation based on feature selection and a support vector machine classifier were finally used to establish the diagnostic model of hippocampal sclerosis. The diagnostic model consists of two consecutive steps: distinguish hippocampal sclerosis (HS) from normal control (NC) and detect whether the HS is located on the left or right side. When the automatic diagnosis model identified HS and NC, the sensitivity and specificity reached 0.941 and 0.917 in the 10-fold cross-validation set and 0.920 and 0.909 in the independent testing set. When the diagnostic model detected HS lateralization, the sensitivity and specificity reached 0.923 and 0.920 in cross-validation and 0.909 and 0.929 in independent testing. Our results show that the developed radiomics model can help detect TLE patients with hippocampal sclerosis and has the potential to simplify preoperative evaluations and select surgical candidates.

Human motor sequence learning drives transient changes in network topology and hippocampal connectivity early during memory consolidation

In the last decade, the exclusive role of the hippocampus in human declarative learning has been challenged. Recently, we have shown that gains in performance observed in motor sequence learning (MSL) during the quiet rest periods interleaved with practice are associated with increased hippocampal activity, suggesting a role of this structure in motor memory reactivation. Yet, skill also develops offline as memory stabilizes after training and overnight. To examine whether the hippocampus contributes to motor sequence memory consolidation, here we used a network neuroscience strategy to track its functional connectivity offline 30 min and 24 h post learning using resting-state functional magnetic resonance imaging. Using a graph-analytical approach we found that MSL transiently increased network modularity, reflected in an increment in local information processing at 30 min that returned to baseline at 24 h. Within the same time window, MSL decreased the connectivity of a hippocampal-sensorimotor network, and increased the connectivity of a striatal-premotor network in an antagonistic manner. Finally, a supervised classification identified a low-dimensional pattern of hippocampal connectivity that discriminated between control and MSL data with high accuracy. The fact that changes in hippocampal connectivity were detected shortly after training supports a relevant role of the hippocampus in early stages of motor memory consolidation.

Incomplete hippocampal inversion in patients with mutations in genes involved in sonic hedgehog signaling

Sonic hedgehog (Shh) signaling pathways are known to play an important role in the morphological development of the hippocampus in vivo, but their actual roles in humans have not been clarified. Hypothalamic hamartoma (HH) is known to be associated with germline or somatic gene mutations of Shh signaling. We hypothesized that patients with HH and mutations of Shh-related genes also show hippocampal maldevelopment and an abnormal hippocampal infolding angle (HIA). We analyzed 45 patients (age: 1-37 years) with HH who underwent stereotactic radiofrequency thermocoagulation and found Shh-related gene mutations in 20 patients. In addition, 44 pediatric patients without HH (age: 2-25 years) who underwent magnetic resonance imaging (MRI) examinations under the same conditions during the same period were included in this study as a control group. HIA evaluated on MRI was compared between patients with gene mutations and the control group. The median HIA at the cerebral peduncle slice in patients with the gene mutation was 74.36° on the left and 76.11° on the right, and these values were significantly smaller than the corresponding values in the control group (80.46° and 80.56°, respectively, p < 0.01). Thus, mutations of Shh-related genes were correlated to incomplete hippocampal inversion. The HIA, particularly at the cerebral peduncle slice, is a potential indicator of abnormalities of the Shh-signaling pathway.

Mapping hippocampal glutamate in mesial temporal lobe epilepsy with glutamate weighted CEST (GluCEST) imaging

Temporal lobe epilepsy (TLE) is one of the most common subtypes of focal epilepsy, with mesial temporal sclerosis (MTS) being a common radiological and histopathological finding. Accurate identification of MTS during presurgical evaluation confers an increased chance of good surgical outcome. Here we propose the use of glutamate-weighted chemical exchange saturation transfer (GluCEST) magnetic resonance imaging (MRI) at 7 Tesla for mapping hippocampal glutamate distribution in epilepsy, allowing to differentiate lesional from non-lesional mesial TLE. We demonstrate that a directional asymmetry index, which quantifies the relative difference between GluCEST contrast in hippocampi ipsilateral and contralateral to the seizure onset zone, can differentiate between sclerotic and non-sclerotic hippocampi, even in instances where traditional presurgical MRI assessments did not provide evidence of sclerosis. Overall, our results suggest that hippocampal glutamate mapping through GluCEST imaging is a valuable addition to the presurgical epilepsy evaluation toolbox.

Epilepsy in Pediatric Patients—Evaluation of Brain Structures’ Volume Using VolBrain Software

Epilepsy is one of the most frequent serious brain disorders. Approximately 30,000 of the 150,000 children and adolescents who experience unprovoked seizures are diagnosed with epilepsy each year. Magnetic resonance imaging is the method of choice in diagnosing and monitoring patients with this condition. However, one very effective tool using MR images is volBrain software, which automatically generates information about the volume of brain structures. A total of 57 consecutive patients (study group) suffering from epilepsy and 34 healthy patients (control group) who underwent MR examination qualified for the study. Images were then evaluated by volBrain. Results showed atrophy of the brain and particular structures—GM, cerebrum, cerebellum, brainstem, putamen, thalamus, hippocampus and nucleus accumbens volume. Moreover, the statistically significant difference in the volume between the study and the control group was found for brain, lateral ventricle and putamen. A volumetric analysis of the CNS in children with epilepsy confirms a decrease in the volume of brain tissue. A volumetric assessment of brain structures based on MR data has the potential to be a useful diagnostic tool in children with epilepsy and can be implemented in clinical work; however, further studies are necessary to enhance the effectiveness of this software.