Anatomical features decide the atypical seizure manifestation of parahypothalamic hamartomas

Abstract

Background: The intrahypothalamic phenotype of hypothalamic hamartomas (HH) is associated with epilepsy, and the parahypothalamic phenotype usually leads to central precocious puberty but not neurological comorbidities or seizures. No study has confirmed the pathological role of parahypothalamic hamartomas in epileptogenesis, and the underlying mechanism is yet to be elucidated.

Objective: We aimed to investigate whether parahypothalamic hamartomas are intrinsically epileptogenic and elucidate the underlying pathway of epileptogenesis.

Methods: We reviewed 92 patients with HH-related epilepsy, categorized them by the classification system of Delalande and Fohlen, and further classified Type I (corresponding to parahypothalamic HH) into the following three groups based on the relationship between the lesion and mammillary bodies (MB): entirely invaded (Group 1), partially connected (Group 2), and not connected at all (Group 3). We examined different anatomical features with their relationship to clinical manifestations. Stereoelectroencephalography (SEEG) was implanted in both HH and extra-HH cortices in different groups to identify the epileptogenic zone. Corticocortical evoked potentials (CCEPs) were also used to determine the pathological correlation among different regions to determine the related epileptogenic network.

Results: A total of 13 patients presented with parahypothalamic HH and 10 (76.9%) presented with non-GS only, with late-onset age and normal cognitive development, which is different from classical clinical features. SEEG showed that HH is intrinsically epileptogenic in MB-involved parahypothalamic groups. No statistical difference was found in onset age (p = 0.213), and lesions horizontally oriented from the tuber cinereum without connection to MB were not involved in seizure genesis. CCEP indicated a pathological connection among HH, middle cingulate cortex, and insular cortex.

Conclusion: The parahypothalamic HH can also cause epilepsy and is different from classic HH-related seizures, by non-GS only with the late-onset age and normal cognitive development. MB is proven to be related to non-GS by the mamillo-cingulate-cortex pathway.

Keywords: hypothalamic hamartoma; network; seizure; semiology; stereo-electroencephalography.

Figure 1

The diagram shows the hypothalamus-related structures from the coronal scan (A) and sagittal scan (B) of MRI. The lateral hypothalamus from the vertical plane is colored in blue, and the structures on the horizontal plane are colored in yellow. The mammillary bodies are marked by white asterisks. The classification system in this study is illustrated in (C) and (D) using coronal and sagittal scans. HHs are colored in red. (C) The modified classification of Delalande Type I (lesions have a horizontal base of attachment below the normal position of the subventricular floor) based on the relationship between the lesion and MB: entire invasion (Group 1), partial connection (Group 2), and no connection (Group 3). (D) Other three types of Delalande classification. Type II lesions have a vertical plane of attachment to the wall of the third ventricle, completely above the subventricular floor. Type III lesions have a plane of attachment that extends both above and below the subventricular floor. Consequently, these lesions have both vertical and horizontal planes of attachment. Type IV lesions were termed “giant” with bilateral insertion. MRI, magnetic resonance imaging; HH, hypothalamic hamartoma; AC, anterior commissure; M, midbrain; OC, optic chiasma; PS, pituitary stalk; TC, tuber cinereum; TV, the third ventricle.

Figure 2

Co-registered images from MRI and PET-CT of each patient by coronal (left), axial (upper right), and sagittal (lower right) scans. Ipsilateral MB and contralateral MB are marked by white and black asterisks, respectively. (A) Five patients from Group 1 showed that the hypothalamic connection was established only on one side of the subventricular floor by a sessile attachment, with invaded ipsilateral MB and preserved contralateral MB. (B) Five patients from Group 2 showed that HH adheres to both TC and MB outside the third ventricle. In this way, the lesion was bilaterally connected to the MB by its caudal part, and on the coronal plane, narrow attachments were observed on both sides. (C) Three patients from Group 3 showed HH horizontally oriented from TC; therefore, the lesion adhered to the subventricular floor only by a pedunculated attachment, being suspended and sparing the MB. (D) The relationship between MB and other Delalande types of lesions. MRI, magnetic resonance imaging; PET-CT, positron emission tomography-computed tomography; MB, mammillary bodies; HH, hypothalamic hamartoma; TC, tuber cinereum.

Figure 3

(A) Patient 8 was a 12-year-old boy with normal cognitive and physiological development. GS started at the age of 9 years and disappeared after HH-resection 3 years ago, and FBTCS relapsed. Semiology includes eye blinking and a contralateral facial tonic seizure and finally evolving into a bilateral asymmetric tonic seizure. The ictal discharge first appeared in HH and then propagated to the INS, insular operculum, and MCC. (B) Patient 13 was a 15-year-old boy with a seizure duration of 5 years. The only seizure type of this patient was eye blinking with awareness, accompanied by flush and vomit sometimes. SEEG recording showed HH onset and then propagated to the INS, insular operculum, and MCC. (C) Patient 10 was a 22-year-old male with normal cognitive and physiological development. Seizure presented at the age of 21 years with typical unconscious oroalimentary and hand automatism. SEEG recording showed an onset from the left HIP and temporal pole and then propagated to the ipsilateral ACC, contralateral HIP, and HH. Even though the HH attachment of HH has rhythm changed, the interictal discharge of HH remained nearly the same. GS, gelastic seizures; FBTCS, focal to bilateral tonic-clonic seizures; SEEG, stereo-electroencephalography; HH, hypothalamic hamartoma; ACC, anterior cingulate cortex; MCC, middle cingulate cortex; AIC, anterior insular cortex; PIC, posterior insular cortex; FO, frontal operculum; PO, parietal operculum; HIP, hippocampus; TP, temporal pole; L, left; R, right.

Figure 4

Showing CCEP results of three patients and their combining results. Five regions were selected to be shown in different ranks by stimulated and recorded conditions. The x-axis stands for the stimulated regions, and the y-axis stands for the recorded regions. Three degrees of the color bar were used to show the different combining results, with a darker bar corresponded to a higher positive rate. Pathological relationships were observed between HH-HIP, MCC-INS, and HH-MCC. CCEP, cortico-cortical evoked potentials; HH, hypothalamic hamartoma; HIP, hippocampus; ACC, anterior cingulate cortex; MCC, middle cingulate cortex; INS, insular cortex.