Medial temporal lobe heterotopia as a cause of increased hippocampal and amygdaloid MRI volumes

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.

Increased expression of caspase 2 in experimental and human temporal lobe epilepsy

Temporal lobe epilepsy (TLE) is often caused by a neurodegenerative brain insult that triggers epileptogenesis, and eventually results in spontaneous seizures, i.e., epilepsy. Understanding the mechanisms of cell death is a key for designing new drug therapies for preventing the neurodegeneration associated with TLE. Here, we investigated the expression of caspase 2, a protein involved in programmed cell death, during the course of epilepsy. We investigated caspase 2 expression in hippocampal samples derived from patients operated on for drug refractory TLE. To understand the evolution of altered-caspase 2 expression during the epileptic process, we also examined caspase 2 expression and activity in the rat hippocampus after status epilepticus-induced acute damage, during epileptogenesis, and after the onset of epilepsy. Caspase 2 expression was enhanced in the hippocampal neurons in chronic TLE patients. In rats, status epilepticus-induced caspase 2 labeling paralleled the progression of neurodegeneration. Proteolytic activation and cleavage of caspase 2 was also detected in the rat brain undergoing epileptogenesis. Our data suggest that caspase 2-mediated programmed cell death participates in the seizure-induced degenerative process in experimental and human TLE.

Amygdala volumetry in "imaging-negative" temporal lobe epilepsy

OBJECTIVE

Although amygdala abnormalities are sometimes suspected in "imaging-negative" patients with video EEG confirmed unilateral focal epilepsy suggestive of temporal lobe epilepsy (TLE), amygdala asymmetry is difficult to assess visually. This study examined a group of "imaging-negative" TLE patients, estimating amygdala volumes, to determine whether cryptic amygdala lesions might be detected.

METHODS

Review of video EEG monitoring data yielded 11 patients with EEG lateralised TLE and normal structural imaging. Amygdala volumes were estimated in this group, in 77 patients with pathologically verified hippocampal sclerosis (HS), and in 77 controls.

RESULTS

Seven of 11 "imaging-negative" cases had both significant amygdala asymmetry and amygdala enlargement, concordant with seizure lateralisation. Although significant amygdala asymmetry occurred in 35 of 77 HS patients, it was never attributable to an abnormally large ipsilateral amygdala. Compared with patients with HS, patients with amygdala enlargement were less likely to have suffered secondarily generalised seizures (p<0.05), and had an older age of seizure onset (p<0.01).

CONCLUSION

Abnormal amygdala enlargement is reported in seven cases of "imaging-negative" TLE. Such abnormalities are not observed in patients with HS. It is postulated that amygdala enlargement may be attributable to a developmental abnormality or low grade tumour. It is suggested that amygdala volumetry is indicated in the investigation and diagnosis of "imaging-negative" TLE.