Clinical manifestations

Presentation and course

Intracranial lipomas are usually discovered as incidental CT or MRI findings in asymptomatic patients, or at postmortem examination. They occur throughout the neuroaxis but are most commonly found in the midline or just lateral to midline structures (15). As they are rare, a comprehensive understanding of their incidence based on neuroanatomic location is limited. Approximately half are thought to arise in the corpus callosum, a fourth in the quadrigeminal or superior cerebellar cisterns, and the remainder in the suprasellar or interpeduncular region, cerebellopontine angle, and in the Sylvian fissure (63). A retrospective study revealed a similar distribution of locations in lipomas, with the pericallosal region being the most common site (30). Clinical manifestations, when present, depend on the location. Epilepsy is the most common presenting symptom in supratentorial lipomas (32). Seizures are especially associated with lipomas involving the Sylvian fissure (32; 38; 68; 24; 45). Corpus callosal lipomas are also frequently associated seizures (07). Other presenting symptoms of corpus callosal lipomas include headaches, behavioral disorders, developmental/mental retardation, memory loss, and craniofacial abnormalities (07). Occasionally, corpus callosal lipomas have an intraventricular extension and can result in hydrocephalus. In rare cases, they present with extracranial extension (20; 18). In addition, corpus callosal lipomas are frequently associated with other midline anomalies, such as callosal agenesis or frontal bone defects (06; 51). Congenital malformations are less commonly associated with lipomas situated in the posterior fossa. Lipomas of the cerebellopontine angle are rare and can be mistaken for vestibular schwannomas (03; 34; 65; 62). Lagman and colleagues reviewed 117 cases and found hearing loss (53%), tinnitus (30%), vertigo or dizziness (25%), and facial symptoms (25%) to be the most common presenting symptoms (34). One third of cerebellopontine or internal auditory canal lipomas are identified as incidental findings, and surveillance with sequential imaging, rather than surgical intervention, is often recommended to follow the growth rate of the lipoma (03). Sudden death due to cardiocirculatory failure secondary to uncal herniation has been reported in lipomas in the cerebellopontine angle (60). Lipomas that are in the cisternal space have been reported to be associated with cranial nerve abnormalities and visual changes (29; 58). A small fraction of intracranial lipomas involves the hypothalamic region, reportedly 2.4% in one series (01). The majority of hypothalamic lipomas are asymptomatic and are a result of incidental findings (01). A case report documented an association between a hypothalamic lipoma and growth hormone deficiencies (53).

Intraspinal lipomas are often associated with spinal dysraphic states (eg, spina bifida). As a result, attention may be drawn to the lumbar spine at birth because of associated cutaneous changes (hair tuft, acral pit) in otherwise asymptomatic patients. Intraspinal lipomas are typically classified based on their location in the spinal cord (21). Most spinal lipomas are located in the lumbosacral region and can be divided into two categories: those involving the conus medullaris and those involving the filum terminale. A third category of subpial lipomas are much rarer, likely have a different embryologic origin, and can be located throughout the spinal cord. Lipomas involving the filum terminale and conus medullaris can be asymptomatic incidental findings, or they may present with symptoms due to spinal cord tethering. These include urinary symptoms such as hyperactive bladder, frequent urinary tract infections, and incomplete voiding. Secondary orthopedic complications due to asymmetric muscle weakness is also common and results in scoliosis, club foot, and asymmetric leg length. Finally, lumbar pain is particularly prominent in older children and adults and presents in a patchy, nondermatomal pattern. Stretching that increases traction on the spinal cord and physical exertion typically exacerbates the pain.

Lipomas are usually slowly progressive over several years, although there are case reports of rapid growth over a short period of time. Most symptomatic lipomas occur in children and young adults during the first 4 decades of life, but they have been recorded in utero and in people as old as 93 years (32; 12).

Clinical vignette

A 35-year-old man presented with a 6-week history of unilateral intermittent throbbing headaches occasionally associated with vomiting. There was no history of migraine, but he had recently changed jobs and had not been sleeping well. He found that acetaminophen only partially relieved his headache. On neurologic examination, there were no abnormal findings.

CT brain scan revealed low density in the region of the posterior corpus callosum, suggestive of a lipoma. An MRI brain scan confirmed the presence of a corpus callosum lipoma.

The patient was prescribed propranolol as a migraine prophylaxis and advised to take analgesia for acute headache. The headaches settled within 1 week and the propranolol was withdrawn after 3 months.

Lipomas are frequently an incidental finding. Posterior corpus callosal lipomas are best managed conservatively.

Biological basis

Etiology and pathogenesis

Most lipomas are congenital, and their etiology is unknown. In 1929, Verga proposed that intracranial lipomas derive from the embryologic meninx primitiva that gives rise to the meninges. It has also been suggested that lipomas arise from a mesenchymal derivative of the neural crest. The current, most likely explanation is that lipomas result from the abnormal persistence of the meninx primitiva and its subsequent maldifferentiation (63). This would account for why there is such a high association with other dysraphic states and why vessels can run through a lipoma. Macroscopically, lipomas appear as yellow, greasy, fatty tissue that can be well vascularized and is interspersed with fibrous tissue. The lipoma may be adherent to nerves or underlying brain or spinal cord (12). Microscopically, lipomas consist of adipocytes that appear histologically normal. A thick capsule of fibroconnective tissue may separate the lipoma from the brain, and several blood vessels and thin delicate fibrous septa may course through the lipoma. Lipogenesis in lipomas is no different to that in normal fatty tissue (26).

Many classification schemes exist for lipomas of the spinal canal. One clinically important way to divide these is by their location in the spinal cord into lipomas of the conus medullaris and of the filum terminale (21). Conus medullaris lipomas are more common and are associated with dysraphic states and overlaying cutaneous stigmata. These are thought to arise as a result of premature separation of the neural tube from the ectoderm. The resulting opening allows for entry of mesenchymal cells, which ultimately give rise to lipomatous tissue (44; 21). Filum terminale lipomas are less commonly associated with cutaneous stigmata and may present with tethered cord syndrome. Their pathogenesis is not well understood.

Epidemiology

As intracranial and intraspinal lipomas are rare, their true incidence and prevalence are unknown. Two autopsy series from 1944 and 1974 report lipomas in 4 of 5000 (0.08%) and 9 of 1956 (0.46%) brain autopsies, though autopsy selection criteria were not clear. Among 3200 patients with brain tumors studied by CT, Kazner and colleagues counted only 11 lipoma cases (0.34%) (32). A survey of 319 cases of primary intracranial tumors in children counted no cases of lipoma (52). This study also reported two lipomas in 21 spine tumor cases. Cerebellopontine angle lipomas account for up to 1% of cerebellopontine angle tumors (41). Intraspinal lipomas are also often discovered in asymptomatic patients but are found more much more frequently than their intracranial counterparts (36). Lipomas of the filum terminale are found in 4% to 6% of cases in postmortem series. Among patients undergoing spinal MRI for unrelated reasons, lipomas are incidentally discovered in about 4% of cases (49). Steroids are associated with development of extra-CNS lipomas. It is unclear if they are associated with increased risk of CNS lipomas, although at least one case has suggested it (25). If steroids do contribute to the risk, it is likely very low.

Prevention

There are no known methods to prevent the development of lipomas.

Differential diagnosis

Confusing conditions

In patients with spinal dysraphism or other congenital malformations, differential diagnosis includes other congenital tumors or cysts (eg, dermoid, epidermoid, teratoma). Imaging of the site of clinical involvement using CT or MRI usually leads to the correct diagnosis preoperatively, although lipomas may be difficult to distinguish from dermoid or epidermoid cysts (05). CT scan usually shows a low-density lesion (generally -50 HU to -120 HU) with no contrast enhancement. Occasionally, lipomas have a fine rim of calcium around them, as can epidermoids. Epidermoids and dermoids can have a similar CT density to lipomas. The characteristic sites for dermoids and teratomas are the pineal, subfrontal, and sphenoid bone regions. The nonmidline sites would be uncommon for lipomas. On T1-weighted MRI sequences and proton density images, lipomas, epidermoids, dermoids, and teratomas can exhibit a high signal (bright or white). Lipomas, epidermoids, and dermoids produce diminution of signal as T2-weighting increases and are low signal area on standard T2-weighted sequences. Three-dimensional T1-GRE images demonstrate more clearly lipomas than standard sequences (42). Dermoids have a nonhomogeneous signal pattern, whereas lipomas of at least 25 mm in diameter are homogeneous. In pericallosal lipomas, vascular flow voids within the lipoma usually represent the anterior cerebral arteries. Small intracranial lipomas close to a cerebral artery are hyperintense on time of flight MR images and could be mistaken for partially thrombosed aneurysms and associated flow-related artifact. Lipomas are associated with chemical shift artifacts in MRIs that can mimic hemorrhages (31).

Diagnostic workup

Intracranial lipomas are typically diagnosed based on characteristic CT or MRI imaging. They appear as well-demarcated extra-axial adipose masses. On CT, they are hypodense and nonenhancing. Calcifications are often present in the fibrous capsule encasing the mass. Lipomas are hyperintense on T1- and T2-weighted MRI images. On fat-saturated T1-weighted images, these lesions are hypointense (13). Lipomas are hypointense on T2-weighted sequences. Associated arterial abnormalities may be seen on vessel imaging. Additionally, midline brain abnormalities are commonly seen in conjunction with intracranial lipomas (28).

Suspicion for intraspinal lipoma may be raised when plain x-rays show spinal canal widening and thinning of the inner aspects of the pedicles. As for intracranial lipomas, their diagnosis is made by the identification of adipose tissue by CT hypodensity and MRI T1- and T2-hypointensity. Other associated dysraphic abnormalities may also be identified, such as spina bifida, vertebral anomalies, and widened interpedicular distance. Although not routinely used, myelography may help assess the relationship between the lipoma and the surrounding subarachnoid spaces. Intracranial cisternal lipomas that are associated with cortical dysplasia have been shown to be better characterized by the Dixon technique, which is a fat-water separation method on MRI (27). Fat is seen in the filum terminale in 3% to 5% of MRIs of the lumbosacral spine (02). In an MR study of 50 patients with occult spinal dysraphism, the conus lay below L3 in 86%, and about half of patients had a lipoma that was commonly associated with an open central canal (56). In cases where there is diagnostic doubt, further investigation depends on the site and suspected differential diagnosis (eg, possibly angiography) if vessels course through the lipoma. Definitive diagnosis is made at operation and by histology.

Management

Most intracranial lipomas are asymptomatic and should be left alone. Given their rarity, there are no clinical trials for their management. Surgical removal of corpus callosal, choroid plexus, and quadrigeminal lipomas should generally be avoided. Surgical removal of Sylvian fissure lipomas has only been reported a handful of times and is considered very risky given the highly vascularized environment (10). Seizures are best managed medically. Hydrocephalus can be treated by ventriculoperitoneal shunting (32). Lipomas in the cerebellopontine angle are usually not associated with other congenital malformations. Decompression should only be considered in cases of an enlarging tumor and compressive symptoms. In a review of 98 cases of cerebellopontine angle lipomas, Tankere and colleagues found that original symptoms improve in 50% of patients postoperatively (61). However, new postoperative deficits were present in 72%, and preservation of hearing was achieved in one quarter of patients.

Spinal lipomas are surgically managed when they are symptomatic. In cases of symptomatic intraspinal lipomas in which the symptoms are progressive, earlier operation has been shown to produce some symptom improvements in patients (57). In cases of symptomatic intraspinal lipomas where there are progressive symptoms or enlarging lipoma or syrinx demonstrated on MRI, there is a general feeling that earlier operation is easier and will produce some improvement in up to 60% of patients. In cases where the lipoma is associated with spinal dysraphism, the natural history is one of progressive neurologic dysfunction, and early operation may be preventative (08). The goal of surgery is to reverse spinal cord tethering.

There is considerable debate in the field as to whether asymptomatic spinal lipomas should undergo prophylactic surgery. Filum terminale lipomas are frequently diagnosed in asymptomatic patients by lumbosacral MRI. When symptoms are present, they tend to manifest the tethered cord syndrome. In a retrospective analysis of the natural history of filum terminale lipomas, Coors and colleagues followed 249 patients over a mean of 3.5 years and found that only one person developed new urinary symptoms that could be attributable to the lipoma. Thus, these authors and others recommend against routine prophylactic surgery for asymptomatic filum terminale lipomas given the apparently benign natural history of the vast majority of asymptomatic cases (17). On the other hand, low surgical complication rates and favorable surgical outcomes lead others to propose early prophylactic surgery (35; 64; 37). In a single center analysis of 174 children who underwent operative management of filum terminale lipomas, the authors note that the mean age at surgery for symptomatic children was 8.3 years, suggesting that a longer follow-up period is needed to determine what percentage of asymptomatic filum terminale lipomas remain asymptomatic if left untreated (64).

Management of asymptomatic lipomas of the conus medullaris is also a topic of much debate. Some authors favor early prophylactic surgery to release the conus and reduce the risk of progressive neurologic deterioration as the child grows (35; 14). The rationale for early intervention is largely based on case series and retrospective analyses in which patients who underwent surgery while asymptomatic had lower rates of clinical deterioration as compared to patients who were symptomatic at the time of surgery (35; 23). In contrast, Wykes and colleagues noted that the majority of patients (60%) with conus medullaris lipomas managed nonoperatively remained asymptomatic at 10-year follow-up, and they instead advocated for close surveillance and surgery only when symptomatic (67). As the arguments for and against prophylactic surgery are based on retrospective data, selection and publication biases likely confound data interpretation. Resolution of this issue will likely require a randomized study of “early” intervention versus “delayed” intervention in asymptomatic spinal lipomas.

Outcomes

Lipomas can remain asymptomatic or be nonprogressive for years. Surgical resection of intracranial lipomas is usually avoided due to high surgical risk. Given their rarity, little data exist on surgical outcomes for resection of corpus callosum and Sylvian fissure lipomas. Attempts to remove cerebellopontine lipomas commonly result in severe deficits. Therefore, surgery should only be considered in patients with definite lipoma enlargement with intractable symptoms from mass effect.

The mortality from operation on lumbosacral lipomas or lipomeningoceles is less than 3%, and between 3% and 32% of patients have deterioration in neurologic or urologic function postoperatively (08; 50; 55; 64). Factors that might predispose to a poorer outcome may be older age at intervention, extension of the lipoma in the spinal canal, and the type of lipoma (transitional-type have more severe deficits) (33). Formal preoperative urological assessment is a strong determinant of eventual urological outcome. Patients with poor bladder studies are less likely to have a good outcome and tend to be older than 3 years of age at intervention. In a retrospective study of untethering of the cord in the presence of a lipomyelomeningocele, the procedure was commonly associated with transient bladder symptoms and did not prevent deterioration, with the majority of patients progressing over a median of 3 years (11). Large lipomas and sacral position may also be poor prognostic indicators (16). Children with asymptomatic lipomas of the conus who selected surgery and were followed for a mean of 10 years showed permanent surgical morbidity in 3% and tethered cord syndrome in 10% compared with 29% in those electing a conservative approach over a similar period (59). Younger age at surgery and cord to sac ratio of less than 50% seemed to increase the risk of tethered cord syndrome. In children younger than 1 year of age with symptomatic lipomas or lipomeningoceles, 29% to 39% show improvements in neurologic or urologic function after an operation to untether the cord (08; 55). Urodynamic testing can be predictive of neurologic deterioration (66). Retethering following operation is common in this group. In older children with symptomatic spinal dysraphism, 18% of whom had lipomas, the outlook following operation was less positive; 5% improved and 26% worsened (16). The current trend is for early operation; however, care must be taken not to over-interpret the results from these studies of presumably different patients. One case report demonstrated that a lumbosacral lipoma spontaneously shrank over a 4-year period, and this coincided with the boy losing a significant amount of weight (19). A publication covering 20 years of follow-up of more than 300 patients with complex spinal cord lipomas reported a neurologic complication rate of about 4% with total and 5% with partial resection (46; 48). Progression-free survival was 88.1% at 20 years for total resection and 34.6% at 10 years for partial resection. Predictors of long-term outcome included age, sex, lipoma type, preoperative symptoms, previous surgery, and postoperative cord-to-sac ratio. The best outcome was in asymptomatic children less than 2 years of age without previous surgery. Where complete resection was not possible, there was a high incidence of scarring of the neural placode and a poorer prognosis compared with no surgery (46; 47). However, there are no controlled trials of early versus later operation. Because complete resection is commonly not possible, recurrences do occur, and pain, arachnoiditis, and neurologic deterioration can complicate recovery of intraspinal lipomas. Approximately 10% of patients with spinal lipomas suffer from symptomatic recurrent tethering of the cord (09). Adult patients with symptomatic lumbosacral lipomas can still benefit from total or near-total resection, with symptom improvements mainly observed in pain and neuropathic ulcers (04).

Special considerations

Pregnancy

There is no evidence that pregnancy has any influence on the growth potential of lipomas; however, it is interesting that in one small series of intradural spinal lipomas not associated with spinal dysraphism, three of four patients had symptomatic deterioration during pregnancy or after delivery (22). Surgery during pregnancy is not usually necessary unless there is evidence of hydrocephalus or serious mass effect from a spinal lipoma. If there is a need to operate during pregnancy, the risks to the fetus are those of any intracranial or intraspinal operation requiring general anesthesia.

Anesthesia

The same concerns and precautions exist for lipomas as for any intracranial or intraspinal surgery.