Epilepsy & Seizures
Hippocampal and parahippocampal seizures
Jul. 31, 2022
MedLink®, LLC
3525 Del Mar Heights Rd, Ste 304
San Diego, CA 92130-2122
Toll Free (U.S. + Canada): 800-452-2400
US Number: +1-619-640-4660
Support: service@medlink.com
Editor: editor@medlink.com
ISSN: 2831-9125
Toll Free (U.S. + Canada): 800-452-2400
US Number: +1-619-640-4660
Support: service@medlink.com
Editor: editor@medlink.com
ISSN: 2831-9125
Worddefinition
At vero eos et accusamus et iusto odio dignissimos ducimus qui blanditiis praesentium voluptatum deleniti atque corrupti quos dolores et quas.
In this article, the author provides a review of hypothalamic hamartoma, with an emphasis on treatment options. Hypothalamic hamartoma is a rare malformation in the ventral hypothalamus, resulting in treatment-resistant (drug-resistant) epilepsy, including gelastic seizures. However, multiple surgical approaches are now available. Treatment should be individualized to the patient’s clinical course and the surgical anatomy of the hypothalamic hamartoma. MR-guided laser interstitial thermal therapy including near real-time MR thermography is the most recently developed surgical treatment option for this disease. Reports describe relatively favorable outcomes for safety and efficacy. Minimally invasive stereotactic techniques (with either laser interstitial thermal therapy or radiofrequency thermoablation) are gaining favor as treatment of first choice for most hypothalamic hamartoma patients.
• Hypothalamic hamartoma should be considered in any patient with gelastic seizures and in any child with early onset of precocious puberty. | |
• Seizures associated with hypothalamic hamartoma are rarely controlled with antiepilepsy drugs. | |
• Cognitive impairment and psychiatric symptoms are common comorbid features with hypothalamic hamartoma and epilepsy. | |
• Surgical treatment of the hypothalamic hamartoma can control seizures and stabilize (or even improve) cognitive and psychiatric symptoms. | |
• The best surgical approach is chosen after considering each patient’s clinical course and surgical anatomy. |
The first description of hypothalamic hamartoma as a cause of precocious puberty (and probably gelastic seizures, although the symptoms were not recognized as such) was published in 1934 (52). In 1958, List and colleagues recognized the association between hypothalamic hamartoma and gelastic seizures (54). In 1988, Berkovic and colleagues described 4 children with hypothalamic hamartoma, treatment-resistant epilepsy, and progressive neurobehavioral deficits, providing the first definitive description of the catastrophic epilepsy syndrome that we recognize today (04).
Hypothalamic hamartoma can cause 2 distinct clinical syndromes (36; 29). Central precocious puberty is associated with hypothalamic hamartoma lesions that attach anteriorly to the ventral hypothalamus, near the tuber cinereum and pituitary stalk. These have been classified as “parahypothalamic” or “pedunculated” based on the local anatomy.
The second syndrome consists of the neurologic symptoms, usually beginning with gelastic (laughing) seizures, but often progressing to additional, more disabling seizure types, along with cognitive impairment and behavioral symptoms. These hypothalamic hamartoma lesions attach posteriorly in the ventral hypothalamus, in the region of the mammillary bodies, and have been referred to as “intrahypothalamic” or “sessile”. Approximately 40% of hypothalamic hamartoma patients with epilepsy also have central precocious puberty, due to larger lesions that have both an anterior and posterior plane of attachment to the hypothalamus.
Hypothalamic hamartoma can cause central precocious puberty, with an isosexual pattern of secondary sexual characteristics (that is, abnormally early progression through the normal sequence of thelarche, adrenarche, and menarche seen with normal puberty). Hypothalamic hamartoma is a particularly common cause of central precocious puberty in young children and should always be excluded by MRI of the brain in boys or girls with central precocious puberty who are younger than 6 years of age (35). Most children with central precocious puberty alone do not experience epilepsy or the other neurobehavioral symptoms to be discussed. However, approximately 40% of children with hypothalamic hamartoma and epilepsy experience central precocious puberty. In our experience, gelastic seizures are usually the first manifestation in this group. However, clinical diversity from patient to patient is 1 of the hallmark features of the hypothalamic hamartoma syndrome, and exceptions to the most common clinical patterns should be anticipated.
Gelastic seizures are the prototypical seizure type associated with hypothalamic hamartoma (30). For those patients with epilepsy, gelastic seizures are usually the first seizure type and occur very early in life. The correct diagnosis is often delayed, but the majority of hypothalamic hamartoma patients begin having gelastic seizures before 1 year of age, and many of these begin during the first month of life. Gelastic seizures are brief (duration is usually less than 20 seconds) and frequent (usually with multiple gelastic seizures per day). They may mimic true laughter but more often are peculiar and mirthless even to the casual observer, and may incorporate behavioral elements of grimacing or crying (dacrystic seizures) (39). They may or may not be associated with altered consciousness, which is often difficult to determine in infants. They can be very subtle, or even subjective, as “pressure to laugh” as an ictal manifestation is reported in adults (82). Gelastic seizures are rarely controlled with antiepilepsy drug therapy (84; 24).
Unfortunately, additional types of seizures develop in 80% of patients. These additional seizures can be disabling. Complex partial seizures are most common, and can mimic seizures that arise from either temporal or frontal lobe regions. Generalized seizure types can include tonic-clonic, tonic, atonic, or even absence. Severely affected patients can develop epileptic encephalopathies with all the features of Lennox-Gastaut syndrome, including drop attacks (05; 25). Approximately 5% of children with hypothalamic hamartoma develop infantile spasms (42).
Significant cognitive deficits occur in roughly 80% of patients and may be progressive in up to 50% (71). When mild, difficulties with processing speed and short-term memory are most common. However, 50% of hypothalamic hamartoma patients with treatment-resistant epilepsy are intellectually disabled (full-scale intelligence quotient [IQ] or developmental quotient [DQ] less than 70) (69; 90). Some patients experience developmental regression and loss of previously learned abilities, usually occurring at the time when seizures worsen.
Psychiatric problems are also common for the cohort of patients with hypothalamic hamartoma and epilepsy (89; 16). Children with hypothalamic hamartoma and epilepsy may have oppositional-defiant disorder (83%), attention deficit/hyperactivity disorder (75%), conduct disorder (33%), and mood disorder (17%) (93). Rage attacks, consisting of aggressive and sometimes destructive behavior arising from minor frustration, are a particular problem for patients with hypothalamic hamartoma and epilepsy and can sometimes be the most disabling aspect of the disease (60).
Clinical features that are positive predictors for the presence of psychiatric symptoms (such as rage behaviors) include male gender, younger age at time of first seizure, presence of intellectual disability, and seizure severity (46; 16).
For those patients with hypothalamic hamartoma and central precocious puberty only, the prognosis is favorable. Almost all of these patients respond to medical therapy with gonadotropin-releasing hormone (GnRH) agonists (29). Once the normal age range for puberty is reached, medical therapy is discontinued, and the normal developmental program of puberty occurs. These patients typically do not have neurologic problems later in life.
For those patients with gelastic seizures, the prognosis is highly variable. Patients with onset of gelastic seizures during adolescence or adulthood (approximately 10% of the population of patients with hypothalamic hamartoma and epilepsy) can have a relatively benign prognosis and may not go on to experience other seizure types or cognitive issues (58; 28).
However, the more common scenario with hypothalamic hamartoma and epilepsy is that gelastic seizures begin early, prior to 5 years of age and often before 1 year of age. These patients are at high risk (80% likelihood) of developing additional seizure types, and 50% have a clinical course consistent with an epileptic encephalopathy, with objective evidence of deterioration in cognition or behavioral health (04; 41; 81). Early intervention with surgical therapy may favorably influence long-term cognitive and behavioral outcome.
The patient was a 5-year-old, right-handed girl with a history of treatment-resistant epilepsy associated with hypothalamic hamartoma. Her mother’s pregnancy was normal, and she experienced an uncomplicated vaginal delivery. Peculiar, highly-stereotyped behaviors were noted during the first week of life, later diagnosed as gelastic seizures. These consisted of unprovoked giggling, accompanied by a frightened look with eyes wide open, and clenching of the fists. These rarely lasted longer than 15 seconds but occurred up to 40 times per day. Her early development was normal for age.
At 3 years of age, she developed complex partial seizures, usually beginning with the gelastic features but leading to activity arrest, behavioral unresponsiveness, and staring. She was lethargic after these events and would sometimes sleep. She was experiencing up to 5 of these daily. Brain MRI demonstrated a hypothalamic hamartoma almost filling the third ventricle and attached to the left side immediately above the mammillary body. The brain MRI was otherwise normal. An EEG demonstrated rare interictal spikes over the left midtemporal region. Three of her usual gelastic seizures were captured, which were associated with diffuse and nonlocalizing rhythmical features on the simultaneous EEG.
Seizure frequency was unaffected by trials of antiepilepsy drugs including levetiracetam, oxcarbazepine, and topiramate. She was described as deteriorating with her socialization and behavior, with severe mood swings and tantrums, and failing to make progress with learning and language skills. There were no recognized endocrine problems.
She was evaluated at our center for surgical treatment at 4.5 years of age, with up to 10 brief complex partial seizures per day. Her family was concerned about her social skills and failure to make developmental progress. Neuropsychological testing revealed Full Scale IQ 82, Verbal IQ 82, and Performance IQ 84. Processing speed was notable as a relative strength. She had significant deficits with language-based skills relative to visual motor skills. Review of her films at multidisciplinary conference showed the lesion to be Delalande Classification Type II (20). Based on the surgical anatomy (hypothalamic hamartoma lesion completely above the floor of the third ventricle and insufficient room within the third ventricle to maneuver a surgical endoscope), a transcallosal interforniceal approach was recommended for optimal resection.
Her postoperative course was uneventful, and she was discharged on the fourth postoperative day. She did not experience diabetes insipidus or obvious short-term memory problems. Postoperative MRI showed no residual hypothalamic hamartoma. One year following surgery, she was completely free of seizures while taking 1 antiepilepsy drug. Previously thin for her age, she had an excessive appetite, and the family struggled to maintain her weight. Her follow-up EEG was normal. Behavior and socialization were regarded as normal for age. Postoperative neuropsychological testing showed stable results, with improved performance for verbal expression and vocabulary.
Most patients with hypothalamic hamartoma (90%-95%) have sporadic lesions unassociated with an identifiable syndrome or other congenital malformations. However, hypothalamic hamartoma can occur with dysmorphology syndromes, most commonly Pallister-Hall syndrome (OMIM #146510) and oral-facial-digital syndrome type VI (OMIM #277170), among other rare associations. Pallister-Hall syndrome is known to be due to germline (total body) loss-of-function mutations within specific portions of the GLI3 gene (expressing a transcription factor in the Sonic hedgehog intracellular signaling pathway) whereas oral-facial-digital syndrome type VI is related to germline mutations of the ciliopathy-associated gene CPLANE1 (09).
Genetics. The recognition that these rare germline mutations could result in hypothalamic hamartoma led to a search for possible somatic (tumor only) mutations as a cause for sporadic hypothalamic hamartoma. Studies utilize the strategy of comparing DNA extracted from surgically-resected hypothalamic hamartoma tissue with DNA derived from leukocytes from the same patient with the application of genotyping techniques such as whole-exome sequencing, targeted sequencing, and single nucleotide polymorphism microarrays.
As a result, approximately 40% of sporadic (nonsyndrome-related) hypothalamic hamartoma are associated with somatic mutations in multiple genes associated with the Sonic hedgehog pathway (including GLI3, SMO, PRKACA, and others) and ciliary structure and function (DYNC2H1 and others) (17; 32; 26). These pathways converge with regard to Sonic hedgehog intracellular signaling participation in ciliary genesis (63). Hypothalamic hamartoma may be a ciliopathy.
Pathology. As a hamartoma, hypothalamic hamartoma lesions contain normal-appearing (ie, not developmentally immature or neoplastic) cells. However, hypothalamic hamartoma tissue is distinct from adjacent normal hypothalamus, which is characterized by nuclei with large neurons. Conversely, hypothalamic hamartoma contain predominately (90%) small neurons (10 to 16 µm diameter cell bodies) that occur in clusters, with relatively sparse intermixed larger neurons. The abundance of neurons and the prominence of neuron clusters vary from case to case (15). Tracts of myelinated fibers are seen in the subependymal region (that is, at the periphery of the lesions), but the specific details of how hypothalamic hamartoma connect to normal brain networks are unknown.
Cellular pathophysiology. Hypothalamic hamartoma are associated with central precocious puberty. The exact molecular mechanisms by which this is mediated are unknown. Chan and colleagues found that tissue expression of various mediators of normal puberty (including gonadotropin-releasing hormone [GnRH]) did not differ between those hypothalamic hamartoma patients with and without a history of central precocious puberty. However, the anatomy of the attachment of these lesions to the hypothalamus did differ significantly. Patients with prior history of central precocious puberty had hypothalamic hamartoma that attached anteriorly in the region of the tuber cinereum and pituitary stalk, whereas patients with epilepsy but without prior history of central precocious puberty had lesions that attached posteriorly in the hypothalamus and lacked attachment to the region of the tuber cinereum (13).
Hypothalamic hamartoma are also intrinsically epileptogenic, which has been established by seizure recordings in which electrodes have been surgically implanted into the hypothalamic hamartoma lesion itself (38; 06; 49). Experimentation with surgically-resected hypothalamic hamartoma tissue has led to a preliminary cellular model for seizure generation (97; 44). The small hypothalamic hamartoma neurons (comprising approximately 90% of total hypothalamic hamartoma neurons) appear to be gamma-amino-butyric-acid (GABA) expressing with an interneuron-like phenotype and possess intrinsic pacemaker-like firing behavior (98; 15; 44).
Conversely, large hypothalamic hamartoma neurons (approximately 10% of total hypothalamic hamartoma neurons) have a phenotype consistent with excitatory, projection type neurons. These large neurons also have the functionally immature property of depolarizing and firing in response to GABA agonists (48; 47; 95; 77). The immature response of the large hypothalamic hamartoma neurons may lead to the unfavorable balance of excitatory and inhibitory activity that is a requisite feature of epileptic tissue. Work by Wu and colleagues has also suggested that gap junctions are present between hypothalamic hamartoma neurons and may functionally contribute to increased synchrony of cellular network firing (96).
Hypothalamic hamartomas are uncommon. Hypothalamic hamartoma associated with epilepsy is reported to have a prevalence of 1 in 200,000 children and adolescents in Sweden (10) and 1 in 250,000 children in Israel (79). Epidemiological studies addressing the prevalence of hypothalamic hamartoma associated with central precocious puberty are not available. Approximately 40% of patients with hypothalamic hamartoma and treatment-resistant epilepsy have a history of central precocious puberty (23).
There are no recognized ethnic or geographical differences with respect to the prevalence of hypothalamic hamartoma. A modest (1.5 to 1) male-to-female predominance is reported in some studies (59; 36). The number of completely asymptomatic cases is likely small, although a handful of these patients have been encountered in our experience as a referral center (less than 2% of total cases). Hypothalamic hamartoma as an incidental finding on MR imaging has not been quantified in any population-based study.
There are no recognized preventative measures for this congenital malformation.
Precocious puberty can be isosexual (early initiation of the complete developmental program of puberty) or isolated to the changes that accompany abnormal female sex steroid production (thelarche = breast development and menarche = female reproductive changes leading to menses) or male sex steroid production (adrenarche = axillary and pubic hair in both genders, and changes involving the scrotum and penis in males). A full discussion of the differential diagnosis of early puberty is beyond the scope of this article. Consultation with a pediatric endocrinologist is recommended. Hypothalamic hamartoma is 1 of the most common causes of central (isosexual) precocious puberty in children younger than 6 years of age. Brain MRI to exclude hypothalamic hamartoma as a cause of precocious puberty in this age group is indicated.
For patients with gelastic seizures, one must always exclude hypothalamic hamartoma as a cause with high-resolution MRI. However, gelastic seizures can arise from other brain regions, most commonly from temporal and frontal lobe structures (14; 75). Aside from hypothalamic hamartoma, other pathologies occurring in or adjacent to the hypothalamus (such as craniopharyngioma, juvenile pilocytic astrocytoma, optic nerve glioma, and colloid cyst of the third ventricle) are rarely associated with gelastic seizures.
Magnetic resonance imaging. MRI is the most appropriate structural imaging technology for diagnosing (or excluding) hypothalamic hamartoma (23; 30). Coronal, axial, and sagittal T1- and T2-weighted and fluid-attenuated inversion recovery (FLAIR) images are recommended. For diagnosis and surgical planning, we find coronal T2 fast spin echo (FSE) with thin cuts and minimal slice gaps to be the single most useful sequence. Sedation is usually required for young or uncooperative patients to minimize movement and to obtain high-quality images.
We recommend at least 1 follow-up MRI to exclude a progressive mass lesion. However, there is good evidence that hypothalamic hamartoma do not grow or expand relative to the normal growth trajectory of the brain (86; 23). Serial imaging is usually not required. Gadolinium administration (contrast enhancement) is recommend for the initial (or first follow-up) imaging study to exclude a contrast-enhancing lesion. Hypothalamic hamartoma lesions do not enhance. Cystic components occur in approximately 1% to 2% of hypothalamic hamartoma (68; 56).
Hypothalamic hamartoma have MRI signal characteristics similar to that of normal grey matter, particularly if the lesions are small. Larger lesions commonly show high signal features with T2-weighted and FLAIR imaging relative to normal grey matter, which correlates with decreased neuronal density as shown by magnetic resonance spectroscopy (03) and stereology on pathological specimens (43). Hypothalamic hamartoma lesions vary tremendously in size. In our series of over 200 patients undergoing surgical treatment, the mean hypothalamic hamartoma lesion volume is 1.74 cm3 (median 0.57 cm3; range 0.04 to 20.04 cm3 [using the technique for calculating the volume of an ellipsoid, derived from conventional measurements of the 3 major axes]).
There are 2 clinicopathological (and clinicoradiological) subtypes based on the location of the hypothalamic hamartoma lesion relative to the floor of the third ventricle (36; 29).
The first phenotype consists of central precocious puberty, with hypothalamic hamartoma lesions that are characterized as parahypothalamic (referring to the position of the lesion located below the floor of the third ventricle and usually having a horizontal base of attachment to the ventral surface of the hypothalamus) or pedunculated (referring to the presence of a stalk or infundibulum that results in the attachment). The second phenotype consists of epilepsy (usually including gelastic seizures) and the cognitive and psychiatric symptoms that are comorbid features of epilepsy. These hypothalamic hamartoma lesions are characterized as intrahypothalamic (referring to the presence of the hypothalamic hamartoma within the third ventricle and having at least some plane of attachment to the vertical wall [or walls] of the third ventricle) or sessile (referring to a broad base of attachment rather than a peduncle).
However, over the last 10 years studies have added diagnostic emphasis on the location of hypothalamic hamartoma attachment along the anterior to posterior axis of the ventral hypothalamus. For patients with central precocious puberty, there appears to be universal attachment of the hypothalamic hamartoma lesion to the region of the tuber cinereum and pituitary stalk in the anterior hypothalamus (regardless of the presence or absence of a peduncle) (13). For patients with epilepsy, there appears to be universal attachment to the region of the mammillary bodies in the posterior hypothalamus (66). Patients with attachment to both the anterior and posterior regions (correlating with large hypothalamic hamartoma lesions) have both central precocious puberty and epilepsy (approximately 40% of hypothalamic hamartoma patients with treatment-resistant epilepsy also have a history of central precocious puberty). In summary, the clinical phenotype of a patient with hypothalamic hamartoma can be predicted by the location of the lesion on sagittal MRI sequences.
Most hypothalamic hamartoma lesions are not accompanied by abnormalities located elsewhere in the brain. However, approximately 5% of hypothalamic hamartoma patients do have other brain findings with a diverse range of features, usually malformations. These include periventricular nodular heterotopias, malformations of cortical development, or even midline developmental defects such as holoprosencephaly, optic pathway dysplasia, or dysgenesis of the corpus callosum (23; 62).
Although the location of attachment of the hypothalamic hamartoma is predictive of the clinical phenotype, the details of synaptic contact and network connectivity between the hypothalamic hamartoma and normal brain are unknown. There is circumstantial evidence from functional imaging techniques that hypothalamic hamartoma associated with epilepsy connect with the limbic circuit (including mammillary bodies, mammillothalamic tracts, anterior nucleus of the thalamus, cingulate gyrus, and mesial temporal structures) (08; 87). This may explain why complex partial seizures associated with hypothalamic hamartoma often mimic seizures of temporal lobe origin (11). However, it also appears likely that gelastic seizures with hypothalamic hamartoma are associated with functional network connectivity to the dorsal pons and associated brainstem and cerebellar structures, possibly via the mammillotegmental tracts (08; 87).
Computed tomography imaging. CT imaging detects large hypothalamic hamartoma lesions but is not adequate to fully characterize soft-tissue mass lesions in this region of the brain. Additionally, many small hypothalamic hamartoma lesions are missed entirely with CT. Normal CT imaging does not exclude a diagnosis of hypothalamic hamartoma.
Electroencephalogram. Although EEG is a conventional biomarker for patients with epilepsy, there are limitations when interpreting EEG results on patients with gelastic seizures and hypothalamic hamartoma (30).
Perhaps most importantly, EEG with standard electrode placement has significantly reduced sensitivity for gelastic seizures. That is, for patients that have exclusively gelastic seizures (more likely in younger hypothalamic hamartoma patients), the interictal recordings are often normal (84). Ictal recordings, capturing gelastic seizure events, often show no change on the simultaneous EEG recording. Troester and colleagues have reported EEG findings on a cohort of 133 hypothalamic hamartoma patients undergoing presurgical evaluation. Based on scalp EEG and standard visual analysis, 56% of all patients experiencing gelastic seizures (and 75% of all gelastic seizure events) did not show a discernible change on the ictal EEG (85).
The interictal and ictal EEG is more likely to show abnormalities in older patients with additional seizure types (84). Here, the problem is a relative lack of specificity, with a diversity of focal and generalized epileptiform changes (84; 59; 85). Interictal focal spikes can be seen from virtually any brain region but are most common over the temporal regions, correlating with the concept that partial seizures arising in the hypothalamic hamartoma probably spread through limbic pathways and can mimic temporal lobe epilepsy. Generalized EEG abnormalities are also commonly observed, including generalized spike-wave discharges, consistent with epileptic encephalopathies such as Lennox-Gastaut syndrome (04; 59; 25). Generalized EEG abnormalities often correlate with generalized seizures types, including tonic, atonic, tonic-clonic, and absence. Patients with hypothalamic hamartoma and infantile spasms may have hypsarrhythmia (42).
Surgical implantation of recording depth wires into the brain, including placement into the hypothalamic hamartoma, has shown that gelastic seizures (and some of the other seizure types) originate in the hypothalamic hamartoma itself, demonstrating that the hypothalamic hamartoma is intrinsically epileptogenic (38; 49; 64; 37). Other seizures, more commonly those with generalized EEG and clinical features, arise from neocortical brain regions without onset in the hypothalamic hamartoma (25; 37). These neocortical seizure foci likely arise over time through a process known as secondary epileptogenesis (41). However, seizures arising from these distant neocortical foci may disappear over weeks or months following surgical resection of the hypothalamic hamartoma, a process known as the “running down phenomenon” (25; 41; Ng et al 2006). As the hypothalamic hamartoma is the appropriate surgical target for most patients, intracranial recordings with surgical implantation of depth electrodes is usually not recommended.
Single photon emission computed tomography. Ictal studies have shown hyperperfusion in the hypothalamic hamartoma and thalamus (34; 49). This can serve as a noninvasive confirmatory test but need not be considered standard in clinical practice.
Positron emission tomography. PET imaging following administration of 18F-flouro-deoxyglucose (FDG) shows generally nonspecific features when obtained during the interictal state. Ictal studies with FDG administration and PET imaging have demonstrated increased metabolism in the hypothalamic hamartoma (65; 78). As with SPECT, FDG-PET may be regarded as a noninvasive ancillary test but need not be considered standard in clinical practice.
Magnetoencephalography. Magnetic dipole mapping within the intracranial space can identify spike origination within hypothalamic hamartoma (51; 50). MEG is a noninvasive ancillary test but need not be considered standard in clinical practice.
Resting state functional MRI. Boerwinkle and colleagues have utilized resting state functional MRI (rs-fMRI) to help guide surgical targeting in 36 children with hypothalamic hamartoma and epilepsy undergoing treatment with stereotactic laser interstitial thermal therapy. This cohort was compared with a nonrandomized control group of 15 patients in which rs-fMRI was not available (07). Posttreatment imaging showed no difference in the treatment (ablation) volume in the 2 groups. However, the cohort with rs-fMRI showed a significant improvement in Engel class I outcome (92% Engel class I for the rs-fMRI group vs. 47% for the non-rs-fMRI cohort; p=0.001). Resting state functional MRI is a technically challenging imaging modality (with regard to signal to noise issues and defining the receiver operating characteristic curve for binary decisions) but may offer value for surgical planning for patients with hypothalamic hamartoma and epilepsy.
Central precocious puberty. Patients with hypothalamic hamartoma and central precocious puberty are treated with gonadotropin-releasing hormone (GnRH) agonists, usually with once-monthly intramuscular injection of leuprolide or related compounds (19; 29). Intramuscular depot formulations of leuprolide give rise to consistently high GnRH agonist levels and mask the pulsatile release of GnRH that is required to trigger puberty. For those rare central precocious puberty patients who fail to respond to medical management (or are hypersensitive to GnRH agonist compounds), surgical resection of the hypothalamic hamartoma is effective for arresting early puberty (53; 56). Once the normal age for puberty is reached, medical therapy is discontinued, and the normal developmental program of puberty occurs.
Gelastic seizures and epilepsy. There are no published trials (controlled or otherwise) of antiepilepsy drug therapy for hypothalamic hamartoma. Probably less than 5% of patients with hypothalamic hamartoma and epilepsy are optimally controlled on antiepilepsy drugs alone, although reports from referral centers likely include ascertainment bias with regard to describing antiepilepsy drug treatment resistance (41). Antiepilepsy drugs seem to be particularly ineffective against gelastic seizures, whereas they probably do decrease the frequency (without providing complete control) of other seizure types. There is no evidence for the superiority of 1 antiepilepsy drug over another in this regard. We recommend a trial of at least 2 antiepilepsy drugs before making a determination of treatment resistance and considering surgical therapy. Hypothalamic hamartoma patients with epilepsy usually have at least 1 seizure per day (often many more), so therapeutic trials of antiepilepsy drugs do not take a long time.
Cognitive deficits. All patients with hypothalamic hamartoma and epilepsy should undergo neuropsychological assessment. A diverse spectrum of deficits is possible from normal to severe intellectual disability (69; 90). There are no cognitive interventions that are specific to hypothalamic hamartoma, but awareness of deficits and intervention with special needs education or functional rehabilitation is recommended.
Psychiatric symptoms. Psychiatric symptoms (most commonly consisting of poor frustration tolerance and rage attacks but may include mood disorder, depression, anxiety, attention deficit disorder, and obsessive-compulsive disorder) may be the most disabling aspect of hypothalamic hamartoma for some patients (93; 46; 16). Pharmacotherapy for these target symptoms, utilizing standard medications, should be considered when appropriate. However, there are no studies that provide evidence for efficacy for psychiatric problems in this population.
General comments. Surgical interventions for hypothalamic hamartoma are usually undertaken to treat refractory epilepsy. From a historical perspective, this is an evolving and still unsettled field, which has moved rapidly from the 1990s (in which surgical intervention was usually not possible or was ill-advised) to 2019 (in which we have multiple surgical treatment choices). New innovations have occurred regularly over the past decade, and today some of the newest and most-promising innovations have limited peer-reviewed evidence regarding safety and efficacy (45). Consequently, this is a dynamic field for epilepsy treatment.
It is important to recognize that neocortical resection is usually not the appropriate treatment choice for patients with hypothalamic hamartoma and epilepsy. Even for those patients with evidence of secondary epileptogenesis, in which seizures are arising from neocortical regions, these seizures may disappear with removal or treatment of the hypothalamic hamartoma lesion (the running-down phenomenon) (25; 41; Ng et al 2006; 81). Neocortical resection has a poor track record for improving seizures in patients with hypothalamic hamartoma and epilepsy (11). Surgical treatment in adults with hypothalamic hamartoma and long-standing epilepsy is associated with a lower success rate and a higher likelihood of complications (21).
The natural history of hypothalamic hamartoma associated with epilepsy is dynamic, rather than static. Patients develop additional seizures types along with cognitive and psychiatric problems over time. Some patients will functionally deteriorate with loss of previously acquired skills and developmental milestones. Secondary epileptogenesis (leading to independent seizure onset at distant sites) also occurs over time. Accordingly, for those with disabling seizures or significant cognitive or behavioral impairments, a proactive approach with earlier surgical intervention should be considered.
Success of surgery for controlling seizures correlates with younger age at time of surgery and a shorter lifetime duration of epilepsy at time of surgery (Ng et al 2006; 76). The likelihood of improved cognition subsequent to hypothalamic hamartoma surgery also correlates with younger age at the time of surgery (94). An independent predictor of surgical success for controlling seizures is the extent of hypothalamic hamartoma resection: completely resected lesions are more likely to result in complete seizure control (Ng et al 2006).
Selecting the single best surgical treatment for hypothalamic hamartoma patients is based on an appreciation of the natural history (stable vs. deteriorating), which, in turn, influences the degree of urgency for immediate (resection, disconnection, or thermal destruction) versus delayed (gamma knife radiosurgery) treatment. Equally important is the surgical anatomy, which is unique to each patient.
Classification and surgical anatomy. Several classification systems have been proposed for hypothalamic hamartoma but our group prefers the system proposed by Delalande and Fohlen as it most directly translates into the realm of surgical planning (20).
Type I. Hypothalamic hamartoma lesion in which the horizontal plane of attachment is completely below the floor of the third ventricle. These would correspond to the designation “parahypothalamic” or “pedunculated.” Many of the lesions that cause central precocious puberty would be Type I. However, Type I lesions with a broad base of attachment that includes the region of the mammillary bodies can also cause epilepsy.
Type II. Hypothalamic hamartoma lesion in which there is a vertical plane of attachment to the walls of the third ventricle, completely above the floor of the third ventricle. These would be the classic “intrahypothalamic” lesions, and would also be considered “sessile” because a stalk or peduncle is not present. These are highly associated with epilepsy, and infrequently associated with central precocious puberty.
Type III. Hypothalamic hamartoma lesion in which the plane of attachment is both above and below the floor of the third ventricle, and thereby possessing a plane of attachment that is both vertical (in the third ventricle) and horizontal (attached to the inferior surface of the hypothalamus). These are larger lesions than Type II and often include attachment that extends anteriorly (to the region of the tuber cinereum) and posteriorly (to the region of the mammillary bodies). Consequently, these hypothalamic hamartoma lesions often include both central precocious puberty and epilepsy.
Type IV. These were characterized as “giant” hypothalamic hamartoma lesions by Delalande (20), without offering clear criteria for the boundary between Types III and IV. Our research group currently utilizes a volume of 4 cm3 as the boundary between III and IV (as determined by utilizing the volume of an ellipsoid from the 3 major axes). However, the surgical planning considerations for Types III and IV are similar.
The surgical approaches are described in the order of their historical development.
Pterional (orbitozygomatic) approach. This was the usual way of resecting hypothalamic hamartoma lesions prior to 1999 (64). When this approach was used for all hypothalamic hamartoma patients, efficacy was low and the complication rate was high (64). However, for Type I patients (with a horizontal plane of attachment below the floor of the third ventricle), this is the optimal surgical approach for open resection. Abla and colleagues report a cohort of 10 patients in whom the choice of pterional resection was individualized to the surgical anatomy (01). With at least 1 year of follow-up, 40% are completely seizure-free (66% of those with 100% hypothalamic hamartoma resection). An additional 40% are at least 50% improved with seizure frequency. See Table 1 for additional details.
Transcallosal anterior interforniceal approach. This surgical approach was pioneered by Walter Dandy in 1923 for lesions within the third ventricle. However, Jeffrey Rosenfeld of Melbourne, Australia was the first to utilize this surgical approach for hypothalamic hamartoma (74). This approach “revolutionized” hypothalamic hamartoma surgery, providing the first major step forward with respect to the surgical treatment of this disease.
Two large uncontrolled outcome studies report similar results. Harvey and colleagues, with a cohort of 29 patients, reported complete seizure control in 52% (31), whereas Ng and colleagues, with 26 patients, reported complete seizure control in 54% (Ng et al 2006). Deficits with short-term memory emerged as a possible complication, with residual postoperative impairment of short-term memory in 8% of subjects (same result in both studies). More recently, van Tonder and colleagues reported a cohort of 10 patients undergoing transcallosal resection assisted by the use of intraoperative MR imaging, with complete control of seizures in 70% and favorable profile for complications (88). See Table 1 for additional details.
Transcallosal resection is a treatment option for those patients with hypothalamic hamartoma lesions with vertical attachment to the walls of the third ventricle (Types II to IV), particularly for those with large lesions that fill the third ventricle (as with the illustrative case noted previously) or for lesions with bilateral attachment.
Transventricular endoscopic approach. Two large uncontrolled trials report similar findings: Procaccini and colleagues report a cohort of 33 patients undergoing endoscopic hypothalamic hamartoma resection, with complete seizure-control in 49% (70), whereas Ng and colleagues have published a cohort of 37 patients, with complete seizure control in 49% (62). Small ischemic infarcts of the thalamus (usually asymptomatic and detected on diffusion-weighted postoperative MRI) emerged as a complication in 30%, possibly due to trauma to small perforating arteries with manipulation of the endoscope (62).
Endoscopic resection is a treatment option for those with smaller intraventricular hypothalamic hamartoma lesions (Type II), particularly with those that have a clearly unilateral base of attachment. See Table 1 for additional details.
Wethe and colleagues at Barrow have reported postoperative neuropsychological testing results for patients with hypothalamic hamartoma and treatment-resistant epilepsy (94). Of the cohort of 32 patients, 63% underwent endoscopic resection. For the entire cohort, there was a statistically significant improvement in full-scale intelligence quotient (IQ), with a pre-operative mean of 83.0 and postoperative mean 91.3 (p< 0.001). For the entire group, there was no significant difference between preoperative and postoperative scores relating to learning and memory, although individual patients did demonstrate diverse (some favorable and others unfavorable) changes between pre-and postoperative testing. Improvement in cognitive functioning was most likely to occur in patients who were younger at the time of surgery (and had a shorter lifetime duration of epilepsy) and in those with lower scores with preoperative testing.
Combined or staged approach for hypothalamic hamartoma surgery. For patients with hypothalamic hamartoma Types III and IV, which have attachment both above and below the normal position of the floor of the third ventricle (and, therefore, have both vertical and horizontal planes of attachment), a staged approach may be required. This would consist of treatment from above (with transcallosal approach, endoscopic approach, or stereotactic thermoablation) and subsequently from below by the pterional approach (27).
Gamma knife radiosurgery. Treatment with Gamma knife radiosurgery delivers a subnecrotizing dose of radiation to the responsible epileptic lesion. Although the exact cellular mechanisms are unknown, it likely does involve neuronal death rather than purely modulatory effects (43). The side effect profile is excellent with little or no risk of adverse events (12).
Regis and colleagues updated their longstanding prospective series of patients with hypothalamic hamartoma and epilepsy undergoing radiosurgical treatment (72). Patients were treated with 17 Gray to the 50% isodose line, which usually corresponded to the margin of the hypothalamic hamartoma as visualized on MR imaging. With follow-up of at least 3 years after treatment, they reported a cohort of 48 patients (of whom 28 [58%] required at least 2 courses of radiosurgery). At last follow-up, 37% had Engel class I outcome (69% for combined Engel class I and II). No residual neurologic deficits (and specifically, no memory deficits) were observed. Transient increases in seizure frequency were noted in 17% of patients after radiosurgery (lasting a median of 30 days [range 9-90 days]) (72). Abla and colleagues report a cohort of 10 patients who underwent Gamma knife surgery, with 40% seizure-free at follow-up (02).
The relative weakness for Gamma knife radiosurgery as a treatment choice is the delay in efficacy, which commonly occurs 6 to 18 months after treatment. Patients may have transient worsening of seizure frequency during a window of time several weeks to several months after Gamma knife surgery, as described in the series of Regis and colleagues (72). Accordingly, we favor the use of Gamma knife for those patients who are stable, that is, able to tolerate the wait time for efficacy. This is more likely to be the case in patients who are older and have gelastic seizures as their only seizure type. We also favor Gamma knife radiosurgery for treating patients with residual hypothalamic hamartoma after undergoing thermoablation or surgical resection. See Table 1 for additional details.
Stereotactic radiofrequency thermoablation. This technique utilizes a stereotactic radiofrequency probe to physically heat the hypothalamic hamartoma lesion, thereby resulting in neuronal injury and death in the target. This technique has the advantage of being minimally invasive relative to open resection but nevertheless still carries a low risk of hemorrhage that accompanies any stereotactic procedure. It has the advantage of immediate effectiveness for those patients who are successfully treated.
Kameyama and colleagues provided an updated report on their cohort of 100 patients undergoing stereotactic radiofrequency thermoablation of hypothalamic hamartoma for treatment-resistant epilepsy (40). With at least 1 year of follow-up (median duration of follow-up 3 years), they reported 71% of patients are completely seizure-free. Psychiatric symptoms were reported to be improved (if not completely resolved) in all subjects. For those patients with both pre- and post-operative neuropsychological testing (69% of the cohort), there was a statistically significant group-wise improvement in full-scale IQ (mean increase in postoperative full-scale IQ 6.1 points; p< 0.001).
Short-term adverse events included Horner syndrome (60%), hyperphagia (28%), and short-term memory loss 8.6%. However, these proved to be transient in most instances. Residual (long-term) complications included delayed puberty (9%), other pituitary dysfunction (2%), and excessive weight gain (7%). See Table 1 for additional details.
Wei and colleagues reported a cohort of 9 patients undergoing radiofrequency thermoablation for hypothalamic hamartoma and epilepsy, with Engel I outcome in 5 patients (56%) (92). Tandon and colleagues added the technical nuance of probe placement for radiofrequency thermoablation with the use of a robotic system, reporting Engel I seizure outcome for 4 of 5 (80%) patients (83).
Stereotactic laser interstitial thermal therapy (LITT) with real-time MR thermography. This surgical technique is the latest innovation for treatment of hypothalamic hamartoma with epilepsy, consisting of stereotactic thermoablation (utilizing a laser-mediated heat source) with the added safety measure of near real-time magnetic resonance thermography. The thermography signal can enable the system to shut off when predetermined temperature parameters are reached at selected anatomical structures during the treatment process in an effort to limit heat-related injury to normal structures (33; 80).
Several substantial treatment series utilizing laser interstitial thermal therapy for hypothalamic hamartoma and epilepsy are now reported. Curry and colleagues from Texas Children’s Hospital have updated their series of 71 patients treated between 2011 and 2018 with an age range at the time of treatment between 5 months and 20 years. Sixteen patients (23%) required at least 2 laser interstitial thermal therapy treatments. At the time of last follow-up, 93% of patients were completely free of gelastic seizures. Unfortunately, efficacy for other seizure types is not available (18).
Xu and colleagues at the Barrow Neurological Institute have published a laser interstitial thermal therapy treatment series of 18 patients with hypothalamic hamartoma and epilepsy (age range 3.3 to 69 years; median age 11 years). Three patients (17%) required repeat laser interstitial thermal therapy treatment. Of the patients with gelastic seizures at the time of surgery, 12 of 15 (80%) were seizure-free. For the patients with other types of seizures, 5 of 9 (56%) were seizure-free. Residual deficits were observed in 22% of patients, including 1 with clinically significant hemiparesis (99).
Du and colleagues at the Zucker School of Medicine have reported a cohort of 8 patients undergoing laser interstitial thermal therapy for hypothalamic hamartoma, age range 3 to 40 years. One patient was treated for psychiatric symptoms of rage attacks whereas the other 7 had intractable epilepsy. Of the 7 patients with epilepsy, 6 (86%) were seizure-free with Engel I outcomes. The patient with psychiatric symptoms was not improved. There was 1 epidural hemorrhage at the time of surgery requiring craniotomy for evacuation without residual neurologic deficit (22).
Age at surgery (years) | HH lesion size | Seizure efficacy | Cognitive and psychiatric outcome | Most common adverse event |
Transcallosal interforniceal approach (31): N=29; Mean follow-up=30 months | ||||
4-23 mean: 10.0 | Diameter range from 0.7 to 4.2 cm | Seizure-free 52% >90% reduction 24% 50-90% reduction 10% No improvement 14% | NA | Transient STM 48% Residual STM 14% Small thalamic infarcts 7% |
Transcallosal interforniceal approach (Ng et al 2006): N=26; Mean follow-up=20 months | ||||
2.1 - 24.2 mean: 10.0 | Volume mean 3.9 cm3 | Seizure-free 54% >90% reduction 35% 50-90% reduction 4% No Improvement 8% | Cognitive improvement 65% Behavioral improvement 88% (subjective parent assessment) | Transient STM 58% Residual STM 8% Weight gain 19% DI 15% |
Transcallosal interforniceal approach (88): N=10; Mean follow-up=37 months | ||||
2.1 – 17.6 mean: 7.3 | Volume NA | Seizure-free 70% > 90% reduction 30% | Cognitive improvement NA Behavioral improvement NA | Transient STM 0% Residual STM 0% Weight gain 10% Transient DI 20% |
Transventricular endoscopic approach (70): N=33; Mean follow-up=19 months | ||||
0.75 - 34 mean: 10.5 | NA | Seizure-free 49% Engel class II and II improvement 49% No improvement 3% | Cognitive improvement 65% Behavioral improvement 75% (subjective report) | STM NA Weight gain 15% Panhypopituitarism 6% Transient DI 3% |
Transventricular endoscopic approach (62): N=37 ; Median follow-up=21 months | ||||
0.6 - 55 median: 11.8 | Volume mean 1.0 cm3 | Seizure-free 49% >90% reduction 22% 50-90% reduction 22% No improvement 8% | Cognitive improvement NA Behavioral improvement NA | Transient STM 14% Residual STM 8% Small thalamic infarcts 30% Weight gain NA |
Pterional approach (64): N=13; Mean follow-up=32 months | ||||
2.5-33 mean: 8.4 | NA | Seizure-free 15% >90% reduction 38% 50-90% reduction 23% No improvement 23% | Cognitive improvement 100% Behavioral improvement 100% (subjective parent assessment) | STM NA Weight gain 8% Small thalamic infarcts 30% CN III paresis 30% |
Pterional approach (01): N=10; Mean follow-up=37 months | ||||
0.7 - 42.7 mean: 18.3 | Volume mean=2.9 cm3 | Seizure-free 40% >90% reduction 10% 50-90% reduction 30% No Improvement 20% | Cognitive improvement 40% Behavioral improvement 20% (subjective report) | Transient STM 40% Residual STM 30% Weight Gain 40% DI 20% |
Gamma Knife radiosurgery (72): N=48; Minimum follow-up=36 months | ||||
3 - 50 median: 16.5 | Volume median 0.4 cm3 | Engel I 40% Engel II 29% Engel III 17% Engel IV 15% | Neuropsychology Testing N = 39 “No significant declines” | Transient worsening of seizures 17% Transient poikilothermia 6% No permanent deficits |
Gamma Knife radiosurgery (01): N=10; Follow-up 18 to 81 months (mean 43) | ||||
5.7 – 29.3 mean: 15.1 | Volume Mean=0.7 cm3 | Seizure-free 40% >90% reduction 0% 50-90% reduction 30% No improvement 10% | Cognitive improvement 30% Behavioral improvement 50% | Transient STM 0% Residual STM 0% Weight gain 20% Transient poikilothermia 10% |
Stereotactic radiofrequency thermoablation (40): N=100; Follow-up=12 to 204 months (median 36 months) | ||||
1-50 median=10 | Maximal diameter 1.5 cm | Seizure-free 71% | Cognitive improvement: Mean full-scale IQ increased 6.1 points; p< 0.001 Behavioral improvement 100% | Transient: Horner syndrome 60% Hyperphagia 28% STM deficit 8.6% Residual: Delayed puberty 9% Other pituitary deficit 2% Weight gain 7% |
Laser interstitial thermal therapy (LITT) (18): N=71; Follow-up=1 to 96 months | ||||
0.4 – 20 mean = NA | Maximal diameter 0.8 – 6 cm | Reported for gelastic seizures only Seizure-free 93% | NA | STM deficit: 0.2% DI 0.2% Symptomatic hyponatremia 0.4% |
Laser interstitial thermal therapy (LITT) (99): N=18; Mean follow-up=17.4 months | ||||
3.3 – 68.9 median: 11 | NA | Seizure-free (gelastic seizures) 80% Seizure-free (other seizure types) 56% | NA | STM 22% Weight gain 22% Hypothyroidism 11% Hemiparesis 6% |
Treatments on the horizon. A technological innovation deserving study as a possible treatment modality for hypothalamic hamartoma associated with treatment-resistant epilepsy is focused ultrasound (57). This is a noninvasive technique delivering multiple transcranial trajectories of ultrasound energy to converge at a predetermined target volume, resulting in thermal injury and hopefully ablation of the target. Near real-time MR thermography is utilized to monitor the delivery of thermal injury and enhance safety. A human research safety board-approved trial has been given approval to recruit patients with deep epileptic lesions, including hypothalamic hamartoma, for possible treatment, but publication of peer-reviewed results is not available at this time (as of June 2019).
Concluding comments. The single best treatment choice for each patient cannot be completely evidence-based at this time. Clinical judgment is required. Head-to-head trials between the different surgical treatment options for patients with hypothalamic hamartoma and epilepsy are unlikely given the rarity of the condition and lack of funding for expensive trials of this nature. In addition, new technologies are still emerging.
Hypothalamic hamartoma patients are diverse with respect to their symptoms and their natural history. Hypothalamic hamartoma lesions vary tremendously with respect to their size, position, and attachment. Accordingly, treatment decisions should be individualized to the circumstances of each case and the expertise of each center (67; 55; 73; 91; 45). Ideally, referral centers for hypothalamic hamartoma should be able to offer most or all treatment modalities and can therefore make the best decision for each patient without bias in favor of only 1 technique.
Treatment algorithm for hypothalamic hamartoma. This algorithm is based on the surgical anatomy of the hypothalamic hamartoma as classified into types I to IV according to the system proposed by Delalande and Fohlen (20).
This approach is utilized at the Barrow Neurological Institute as a guide to treatment decision-making. The algorithm was developed on the basis of expert opinion and the available literature as discussed in this section. Controlled, randomized treatment trials do not exist. Decision-making should always be individualized to the clinical circumstances of each patient and the experience of the local institution. Ideally, patients should be evaluated and treated at hypothalamic hamartoma referral centers with established multidisciplinary programs. Hypothalamic hamartoma type is based on the Delalande classification.
Women with hypothalamic hamartoma are capable of pregnancy and successful childbirth. We are not aware of any obstetrical issues for this patient cohort. For those women with hypothalamic hamartoma trying to become pregnant, consultation with a reproductive endocrinologist may be useful due to possible endocrine dysfunction within the reproductive axis.
Hypothalamic hamartoma patients with no prior history of hypothalamic hamartoma surgery are not likely to be at risk from anesthesia. However, hypothalamic hamartoma patients who have undergone prior surgical resection in the hypothalamus may have deficits relating to sodium and electrolyte balance (diabetes insipidus) or adrenal or thyroid insufficiency. Preoperative evaluation for these possible deficits is indicated.
John F Kerrigan MD
Dr. Kerrigan of Barrow Neurological Institute at Phoenix Children's Hospital and University of Arizona College of Medicine has no relevant financial relationships to disclose.
See ProfileJerome Engel Jr MD PhD
Dr. Engel of the David Geffen School of Medicine at the University of California, Los Angeles, received honorariums from Cerebel for advisory committee membership.
See ProfileNearly 3,000 illustrations, including video clips of neurologic disorders.
Every article is reviewed by our esteemed Editorial Board for accuracy and currency.
Full spectrum of neurology in 1,200 comprehensive articles.
MedLink®, LLC
3525 Del Mar Heights Rd, Ste 304
San Diego, CA 92130-2122
Toll Free (U.S. + Canada): 800-452-2400
US Number: +1-619-640-4660
Support: service@medlink.com
Editor: editor@medlink.com
ISSN: 2831-9125
Epilepsy & Seizures
Jul. 31, 2022
Epilepsy & Seizures
Jul. 31, 2022
Epilepsy & Seizures
Jul. 31, 2022
Epilepsy & Seizures
Jul. 30, 2022
Epilepsy & Seizures
Jul. 19, 2022
Epilepsy & Seizures
Jul. 14, 2022
Epilepsy & Seizures
Jul. 06, 2022
Epilepsy & Seizures
Jul. 06, 2022