Case Presentation
Case 1
A 47-year-old male was a known case of left putaminal CM. He underwent craniotomy
and excision of the lesion following an episode of hemorrhage and hemiparesis in 1996.
The residual right hemiparesis improved with time. He was on regular follow-up with
serial MRI which showed no residue/recurrence/any other lesions. The follow-up MRI
done in January 2018 showed a small lesion in the pons which was planned for conservative
management as the patient was asymptomatic. He presented to the emergency department
4 months later with sudden-onset altered sensorium. On neurological examination, he
was E3, V4, and M6 with ataxia.
The MRI brain revealed a large intra-axial pontine lesion, extending more to the left
with a characteristic “popcorn” appearance with a rim of signal loss due to hemosiderin,
with areas of fresh bleed and surrounding edema causing brainstem compression. The
SWI sequences demonstrated prominent blooming and the T2 signal was varied internally
due to multiple hemorrhages within the lesion. The T1 images were isointense to hyperintense
[[Figure 1]].
Figure 1: (a) Preoperative magnetic resonance imaging axial T2 image. (b) Preoperative magnetic
resonance imaging axial T1 image. Large lesion in the Pons with extension to the left
with perilesional edema and mass effect on the brain stem (c and d) Postoperative
T1 and T2 images confirmed complete excision
Microsurgical removal of the lesion was performed by a combined left retrosigmoid
and posterior transpetrosal approach with the patient in lateral position. Intraoperative
SSEP, MEP, and BAER were recorded. There was no discoloration noted on the pontine
surface. The lesion was approached through the lateral pontine zone/the peritrigeminal
area.[[10]],[[13]] Postcorticectomy, the hematoma was evacuated. The lesion was reddish brown in color,
well-marginated, firm, found adherent to the surrounding brain stem. Piecemeal complete
excision of the lesion was done. There were no changes in the evoked potentials intraoperatively.
Histopathological examination revealed the lesion to be a CM.
Postoperative period was uneventful. Patient's sensorium was normal with significant
improvement in ataxia. He had mild left trigeminal hypoesthesia postoperatively. Postoperative
MRI showed no residual lesion [[Figure 1]]. At 2-month follow-up, the patient was symptom-free.
Case 2
A 67-year-old male, who was a known case of right trigeminal neuralgia for 6 months,
presented with a history of right facial, persistent dysesthesia for 2 months which
did not subside with medications. On neurological examination, he had a right hemifacial
hypoesthesia involving the ophthalmic, maxillary, and mandibular divisions of the
trigeminal nerve. He had no other neurological deficits.
The MRI brain showed a right side intra-axial lesion, at the level of the root entry
zone of trigeminal nerve. The SWI sequences showed blooming. The lesion was hyper-
to iso-intense on T1WI and hyper to hypointense on T2 weighted image (T2WI).
Microsurgical removal of the lesion was performed by the right retrosigmoid approach,
with the patient in lateral position. Intraoperative SSEP, MEP, and BAER were recorded.
The lesion was approached through the safe entry zone in the lateral pontine or the
peritrigeminal area,[[13]] similar to the previous case. There was mild xanthochromic discoloration noted
on the pontine surface. Dark reddish brown fluid was found after corticectomy. The
lesion was reddish brown with a relatively good margin in most areas and was excised
piecemeal. Near total decompression of the lesion was done when the BAER showed increased
latency, which is when further excision was stopped.
Histopathological examination revealed the lesion to be a CM.
Postoperative period was uneventful. Patient's symptoms subsided completely with no
neurological deficits. There was no postoperative rebleed seen from the residual lesion.
Postoperative MRI showed a small residual lesion on the medial side of the operative
cavity.
Case 3
A 47-year-old male presented with a subacute onset left-sided facial numbness and
right-sided deviation of angle of mouth for 1 month. On examination, there was hypoesthesia
in the left V1 and V2 divisions and left UMN Upper motor neuron (UMN) type of facial
palsy. There were no other deficits.
MRI brain showed a left pontomedullary lesion with a dorsal component abutting into
the fourth ventricle with similar imaging characteristics as the other 2 lesions [[Figure 2]].
Figure 2: (a) Preoperative magnetic resonance imaging axial T2 CISS image. (b) Preoperative
magnetic resonance imaging axial T1 image. Large lesion left pontomedullary junction
abutting into the fourth ventricle. (c) Intraoperative view of the lesion abutting
into the fourth ventricle. (d) Intraoperative view showing complete excision of the
lesion. (e and f) Postoperative T1 and T2 images confirmed complete excision
Lesion was approached through a suboccipital craniotomy and a trans-fourth ventricular
approach. Intraoperative monitoring was similar to what was used in the previous cases.
Lesion was found to be abutting into the fourth ventricle with the surrounding brainstem-stained
xanthochromic due to the bleed [[Figure 2]]. Lesion was capsulated with partial liquefaction of the hematoma. There was a good
plane between the lesion and the surrounding brain. Complete excision of the lesion
in a piecemeal fashion was done.
Histopathologically, the lesion was proved to be a cavernoma.
Postoperatively, the facial numbness partially subsided. There was persistent left
UMN type of facial palsy. The patient also developed a left abducens nerve palsy postoperatively.
Postoperative MRI showed complete excision of the lesion [[Figure 2]]. At 3-month follow-up, there was complete improvement in the facial and abducens
nerve palsy along with the facial numbness.
Discussion
Epidemiology
Relapse and remission are the two most common words used in correlation with the natural
history of CMs. Cavernomas may be single or multiple; familial or sporadic in occurrence,
and congenital or de novo in evolution. The true natural history of these familiar
lesions is yet unclear despite so many studies on the topic.[[14]] They are low flow vascular malformations.[[3]] The most common location of CMs in the brain stem is the pons followed by the midbrain
and medulla. BSCs constitute 8.5%–35% of all symptomatic intracranial cavernomas.[[14]]
Clinical presentation
BSCs can have a varied clinical presentation. About 40% of the patients remain asymptomatic
till the first episode of bleed. Patients with hemorrhage may present with subjective
symptoms such as headache, vomiting, giddiness, nausea, altered sensorium or rarely,
and trigeminal neuralgia.[[1]],[[3]],[[10]] However, episodes of loss of consciousness or cardiorespiratory failure though
reported are rare. Focal neurological deficits can manifest in the form of cranial
nerve (CN) palsies, motor/sensory deficits, or cerebellar signs. The deficits may
fluctuate in their degrees of severity and combination depending on location, size
of the lesion, hemorrhagic episodes-single or recurrent, and extent of hemorrhage.
There is a direct interaction between the persistence of neurological deficits and
the intervening time duration of hemorrhage. Symptoms are usually subacute in onset
with gradual progression over hours to days. Neurological deficits usually improve
with time and some authors have reported up to 37% of complete recovery.[[1]],[[2]],[[4]],[[10]],[[15]]
Radiology
MRI is the optimal standard for the diagnosis of cavernomas. The radiographic appearance
is variable, depending on the stage of hemorrhage. The classical description of CMs
is known as the “popcorn,” with a central area of heterogeneous signal on T1 and T2WI,
surrounded by a ring of hemosiderin, which is hypointense on T2WI. T1 and fluid-attenuated
inversion recovery images are helpful in defining the boundaries of the CM and to
assess how close the CM is to the pial surface. T2 MRI should not be used for this
assessment as the “blooming” artifact of the peripheral hemosiderin content provides
a false and often exaggerated assessment of the CMs. The most sensitive sequence to
detect cavernomas is the gradient echo T2 or the SWI sequences because of the magnetic
susceptibility of products generated by degradation of hemoglobin. SWI sequences are
also used to screen the brain and cord to look for multiple lesions in familial cases.[[3]],[[10]] Diffusion tensor imaging allows for the visualization of white matter tracts, thus
improving the anatomical localization of corticospinal and sensory tracts preoperatively.
The relationship between the lesion and the dislocated fiber tracts can be displayed
in a three-dimensional manner, facilitating the preoperative planning of the surgical
approach.[[6]],[[7]],[[10]]
Hemorrhage rate
Hemorrhage rate in cerebral CMs is the most researched aspect of their natural history.
A thorough literature search on the natural history of CMs and BSCs has revealed variable
hemorrhagic rates with considerable controversy with regard to it. Unruptured CMs
have a relatively low prospective risk of hemorrhage (0.4%–0.6% per patient-year).
Annual rates of hemorrhage range from 2.3% to 13.6% and rebleeding rates vary between
5% and 21.5% in various studies.[[2]],[[10]],[[14]] There is a significant range in the variability in the annual hemorrhagic rate
from 15% to 60.9% as quoted by various authors such as Taslimi et al.,[[4]] Horne et al., and others.[[1]],[[5]],[[16]],[[17]] However, there are various confounding factors that could have led to this large
variability, thus making it difficult to bank on any result till date. From all the
studies analyzed, we report an overall annual hemorrhage rate of 2.5% per patient-year
for cerebral CMs (95% confidence interval 1.3%–5.1%).[[1]],[[2]]
Definition of hemorrhage
The diverse hemorrhagic rate available in the literature could also be due to the
fact that there is no standardized definition on the term “hemorrhage” or “recurrent
hemorrhage” till date. Very few studies have attempted to bring about clarity amidst
this muddle. Al-Shahi Salman et al.[[18]] in his systematic analysis concluded that the available data were inconclusive
about the following aspects such as confirmatory imaging, whether the hemorrhage should
be clinically symptomatic and whether it could extend beyond the CM or not. He thus
defined a CM hemorrhage as “the one having acute or subacute onset of symptoms (any
of headache, epileptic seizure, impaired consciousness, or new/worsened focal neurological
deficit referable to the anatomic location of the CM) accompanied by radiological,
pathological, surgical, or rarely only cerebrospinal fluid evidence of recent extra-
or intra-lesional hemorrhage.” The definition includes neither an increase in CM diameter
without other evidence of recent hemorrhage, nor the existence of a hemosiderin halo.
Studies quoting rehemorrhage rates based only on the clinical parameters could be
erroneously overrating it as there could be other clinical factors such as edema or
thrombosis that can cause alterations in the clinical events. Only studies defining
rehemorrhage based on clinical parameters, with MRI confirmation of the hemorrhage
will provide a more accurate estimation which will understandably be smaller than
the current record.[[18]],[[19]]
Time of presentation
Time is the single most important factor in determining bleeding rates. Since CMs
can be congenital, radiation-induced, or de novo in origin, the assumption that the
bleed was from a congenital lesion thus neglecting the ones from de novo lesions makes
it not only an erroneous estimate but an underrated one too. Thus, making the current
literature quite unaccountable for this bias as there is no way in differentiating
the two from any of the studies available in literature.[[6]],[[20]],[[21]]
Patient selection
The studies quoted in literature have only considered patients with symptomatic hemorrhage,
neglecting the asymptomatic ones, thus making it a selection bias. Based on this bias,
it can be safely concluded that the rehemorrhage rates are inaccurate as well. Only
the selected cohort of symptomatic cases referred to tertiary care centers and institutes
are the ones considered for surgical series. The asymptomatic ones and patients with
surgical contraindications are not considered at all, thus making it a referral bias.
Evidently, these clustered cases have a higher hemorrhagic risk than the asymptomatic
patients contributing to higher rebleed rates. In prospective studies which determine
the natural history, asymptomatic patients and incidentally detected nonsymptomatic
patients are recruited, excluding the patients who require surgical intervention.
Thus, underestimating the bleeding risk, making it a selection bias. The time duration
and the sample size of a study are influential in the assessment of bleeding risk.
The risk of hemorrhage reduces with time, most often after the first 2 years of hemorrhage,
thus proving that short follow-up periods are likelier to cite higher hemorrhagic
risk. When the sample size is small, the values are erroneous as they can neither
represent the population in a statistically significant manner nor the natural history
of the disease.[[1]],[[2]],[[4]],[[9]],[[19]],[[21]],[[22]]
Hemorrhage risk factors
Various studies in literature quote a multitude of risk factors for hemorrhage in
CM/BSC [[Table 1]]. Female sex, patient age, associated developmental venous anomaly (DVA), perilesional
edema, large lesion size, history of previous ictus, deep location/brainstem, and
multiplicity of CMs have been reported to affect the risk of bleeding. The percentage
of risk among these are vacillating in their spread and report among the studies.[[1]],[[4]],[[18]],[[19]],[[22]],[[23]],[[24]]
Table 1: Risk factors for hemorrhage in cavernous malformations (cerebral/brainstem)
Anatomical location
Porter et al.[[25]] reported a 30-fold greater rate of hemorrhage in infratentorial cavernomas than
in supratentorial ones. Many studies report a significantly higher hemorrhagic rates
of CM in the brain stem and other deep locations like the Basal Ganglia.[[1]],[[2]],[[14]],[[26]] Porter et al.,[[27]] in a prospective study of cerebral CMs, reported the deep location to be a significant
factor influencing the clinical event risk. The reported rate of bleeding of 2.7%–6.0%
per patient-year in BSC exceeds that of cavernomas (2.4%) in the other intracranial
locations.[[2]],[[14]],[[19]],[[26]],[[28]] In a recent meta-analysis by Horne et al.[[17]] on the natural history of untreated cerebral CMs, 575 cases of BMC were included
in the study. Brainstem location was independently associated with the occurrence
of intracranial cerebral hemorrhage (30.8% 5-year risk bleed). The cause for this
increased risk in the brain stem could be due to its structure and eloquence that
makes it highly sensitive to even subtle changes in the lesion morphology which is
absent in the lesions present elsewhere in the brain. Thus, in theory, leading to
a discovery of a higher hemorrhagic rate.
Previous ictus
Several prospective studies pertaining the natural history of CMs and BSCs have reported
that history of previous hemorrhage is a definitive risk factor for subsequent one;[[2]],[[18]],[[21]],[[22]] however, there are few authors like Kupersmith et al.[[29]] who have reported that there is no significant difference in the risk between the
bled and unbled lesions.
Sex
Many prospective studies of natural history of cerebral CM show a higher predilection
toward the female sex as a risk factor for bleeding. Li et al.[[21]] and Al-Shahi Salman et al.[[16]] in their prospective studies with 331 and 139 patients, respectively, along with
other large series have statistically proved that the female sex is at a higher risk
of bleed whereas there are a few studies that have opined that the female sex does
not influence the risk of hemorrhage.[[1]],[[2]],[[5]],[[11]]
Age of the patients
The mean age of patients who present with symptomatic hemorrhages ranges between 32
and 38 years.[[10]] Many retrospective as well as prospective studies have opined that younger age
(<40 years) has significantly higher risk of hemorrhage,[[5]],[[15]] although there are a few studies that have reported that age >50 years is a risk
factor for bleed.[[21]]
Size of the lesion
Most of the studies such as Al-Shahi Salman et al.[[16]] and Kupersmith et al.[[29]] in their prospective studies report that lesions beyond 10 mm carry a greater risk
of hemorrhage. Li et al.[[11]] in his retrospective analysis has concluded the same in lesions more than 20 mm.
However, there are a few authors like Li et al.[[21]] who in his prospective analysis showed that lesion size did not significantly impact
hemorrhage risk.
Developmental venous anomaly
DVAs are congenital anomalies of normal venous drainage, consisting of a number of
dilated medullary veins converging into a single large draining vein, typically presenting
with a caput medusae appearance. Some authors have reported that DVA has a higher
risk of bleeding. The pathological basis to this is said to be that DVA affects the
formation and clinical course of CMs by causing venous hypertension. Since DVAs are
naturally vulnerable to hemodynamic changes, there is a significantly higher risk
of bleed.[[1]],[[11]] However, not all studies of the natural history of CM have had similar results.[[1]],[[2]],[[16]]
Perilesional edema
It was found to be a significant predictor of hemorrhage in various studies. The pathological
basis for this was hypothesized to be impaired venous drainage and formation of vascular
connections between the lesions and the surrounding tissue. Edema also was related
invasion and infiltration of the lesions into the brainstem, thus causing a higher
risk of hemorrhage.[[11]],[[21]]
Other factors
Systemic arterial hypertension is said to be an attributable factor causing increased
hemorrhagic risk in CMs. The pathological basis is touted to be the changes in arterial
pressure that could cause meaningful alterations in the hemorrhagic propensity and
patterns within the CM.[[1]],[[21]] Although hypertension is a risk factor, it has been specifically found that pregnancy
is not a risk factor for bleeds in CM.[[1]]
Multiplicity of the lesions had no significant risk of hemorrhage on a per lesion
basis as consistently seen in most of the studies. It only results in a cumulative
increase in the hemorrhagic rate per patient.[[16]],[[21]]
Surgical intervention
In 1928, Dandy first resected a CM located in the pontomedullary region [[11]],[[30]] and since then, the advances in microsurgical techniques and technical aids such
as intraoperative monitoring and neuronavigation have had tremendous progress. Complete
surgical excision is the treatment of choice in BSCs. Surgical outcome with complete
excision and good clinical outcomes is seen in many studies but along with it comes
a high rate of immediate and long-term postoperative complications. Surgery in the
brainstem is more often than not associated with morbidity and mortality owing to
the compact nature and eloquence of the structure. Thus, understanding the natural
history, evaluation of the preoperative deficits, thorough preoperative planning of
the surgical approach, intraoperative adjuncts such as electrophysiological monitoring/neuronavigation
and the surgeon's expertise is of utmost essence, not only for a safe resection and
an acceptable postoperative outcome but also for a better quality of life.[[1]],[[10]],[[31]]
Indications for surgery
Surgical indications in BSC's have always been a controversial topic which has garnered
varied viewpoints from surgeons across the globe [[Table 2]]. Although CMs are benign, with the asymptomatic lesions having a low hemorrhagic
threshold, hemorrhage clustering with a higher rate of rebleeds ranging from 15% to
60% in the first 2 years following a bleed have been reported in all the major studies.
Not only does the rebleed rate decline gradually after 2 years, the neurological deficits
are also seen to improve spontaneously after a hemorrhage. Up to one-third of the
patients make complete recovery with time. Thus, one of the most important indications
for surgery is the need for early intervention after a symptomatic bleed, so as to
prevent recurrence of hemorrhage. However, intervention must be deferred if the presentation
is after 2 years of a bleed, as that in itself could cause new deficits.[[1]],[[4]],[[16]],[[17]],[[32]] Many authors have opined that it is reasonable to wait until after the second hemorrhage
for a surgical intervention, as that puts the lesion under an “aggressive subset”
which has a higher tendency to bleed than the others. Hence, the surgical risks in
these patients are better accepted than the severity of the neurological risk following
a subsequent bleed which can have disastrous consequences.[[1]],[[7]],[[10]],[[15]],[[28]] Thus, surgical consideration in symptomatic patients after the first or the second
bleed would be the most appropriate time.
Table 2: Indications for surgery
Some authors are of the opinion that symptomatic lesions should be operated when they
are close to the pial surface or accessible through the safe entry zones to prevent
complications.[[8]],[[15]],[[24]],[[26]],[[28]],[[32]],[[33]] When the lesion has caused significant mass effect on the surrounding structures
resulting in altered consciousness or the need for life support, surgical intervention
should be immediate irrespective of the presence or absence of bleed. The ones in
the medulla must be positively treated according to some studies to avoid further
life-threatening events. Chen et al.[[34]] compared the initial and final neurologic states between conservative and surgical
treatment groups and found no significant differences regardless of patient age. However,
the surgical threshold must be lower in children as there is higher cumulative lifetime
risk of hemorrhage.[[4]],[[7]],[[9]],[[10]],[[22]],[[26]],[[28]],[[32]] In the geriatric population, intervention after a symptomatic bleed is essential
as the elderly are less tolerant to the functional damage, thus compromising on the
quality of life with worse outcomes.[[14]] Deep-seated lesions, notably those which are inaccessible to the safe entry zones,
carry a higher risk of immediate and long-term morbidity,[[35]] where only some have had impressive results after the removal of such catastrophic
lesions.[[1]],[[26]] It is safer to have a “wait and watch policy” till further bleeding episodes in
these patients as that might make the lesion more amenable to surgery by reaching
closer to a pial surface.[[32]]
Surgical intervention is not preferred in asymptomatic patients, incidentally detected
lesions, patients with mild/transient symptoms or patients with a single bleed having
mild symptoms. Surgery should be deferred in patients with mild symptoms, especially
when the intervention itself carries a risk of significant permanent symptoms.[[9]],[[10]],[[21]],[[26]],[[28]] The goal of surgery is to eliminate the risk of recurrent hemorrhage and to improve
and stabilize preoperative function while minimizing surgical complications. Thus,
complete excision is imperative to avoid renewed hemorrhage. The circumstances of
leaving behind a residue are acceptable is when its anticipated that complete removal
would have a high risk of permanent deficits.[[10]],[[25]],[[26]]
Timing of surgery
The objective of surgery in BSC is radical resection because partial removal is associated
with a persistent and higher risk of hemorrhage from the residual lesion. Despite
a few reports proposing surgery during the acute phase to decompress the brain stem,
many recommend delaying surgery for about 4–6 weeks after a symptomatic hemorrhage
unless the patient has a life-threatening need for intervention in the form of altered
consciousness, cardiorespiratory instability, or progressive neurological deficits.
Surgery during the first 2 weeks is not advisable since the hematoma is yet solid
with perilesional edema is at its maximum, thereby increasing postoperative deficits.
During the 4–6 weeks period, there is liquefaction of hematoma and the edema also
subsides (steroids may be used to reduce edema), thus providing a natural buffer against
surgery related trauma to the surrounding structures as the hematoma itself provides
a good plane for dissection. Usually, after hematoma evacuation, there is adequate
space that is obtained for the excision of the cavernoma without any need for retraction
of the brain stem.[[1]],[[6]],[[7]],[[8]],[[10]],[[13]],[[14]],[[21]],[[22]],[[26]],[[33]],[[36]] Authors such as Pandey et al.,[[33]] Garcia et al.,[[8]] and Zaidi et al.[[15]] who have large series (>100 patients) recommend intervention after 4 weeks but
before 8 weeks of bleed, coz a further delay in surgery as there is retraction and
organization of the hematoma along with gliosis, hyaline degeneration, and calcifications
leading to tight adherence between the CM and the surrounding parenchyma, making the
dissection plane obscure thus increasing the likelihood of mechanical trauma from
surgical manipulation.[[7]],[[8]],[[10]],[[13]],[[14]],[[21]],[[22]],[[26]],[[33]]
According to the foregoing opinions and our experience as reported in two cases, subacute-phase
surgery (i.e., when the hematoma is liquefied) is useful for the complete excision
of the lesion with a minimal damage to the surrounding structures.
Principles of surgery in brainstem cavernomas
Surgical management of developmental venous anomaly
There is a clear association between DVA and CM. About 16%–100% of CMs are found in
association with DVA.[[36]] A few authors have had a positive correlation between DVA and bleeding risk of
CM. Significant difference in the hemodynamic alterations around DVAs with and without
CM was found. CMs associated with DVAs had a significantly higher bleeding and rebleeding
risk compared to the ones without the association. Thus, hypothesizing that the abnormal
hemodynamics of DVAs might induce the formation of CMs. The pathological basis to
this is the chronically increased intraluminal pressure and the resulting reduced
tissue perfusion leading to tissue hypoxia, stimulating a local increase in angiogenic
factors, which would induce the formation of vascular malformations.[[10]],[[37]] Surgical management of DVA is yet another controversial topic that remains unclear
till date. A few studies advocate complete removal of DVAs as they are promoting factors
in the development of cerebral CMs and resection of the associated DVA components
may prevent regrowth of a partially excised CM,[[1]],[[38]] whereas the others are of the opinion that complete removal can cause hemorrhagic
infarction as these drain normal brain as well. However, leaving the DVAs intact carries
a risk of residual CM which might result in rebleed.[[1]],[[10]],[[11]],[[39]] Zhang et al.[[39]] in his study has recommended complete excision of the CMs combined with the coagulation
of the distal radicles in association with the CMs and preservation of the caput medusae
and main trunk of the DVAs.
Choice of surgical approach
A good surgical approach must minimize the brain retraction and violation to the normal
structures. The shortest distance from the pial surface to the lesion need not be
the safest. The presence and position of DVAs also influence the choice of approach
and trajectory.[[10]],[[37]] Cavalcanti et al.[[40]] and Giliberto et al.[[13]] have described various microsurgical safe entry zones and approaches based on the
location of the lesion in great detail with both cadaveric pictures and exquisite
figures. We have combined the work of these authors to give a summary of all the approaches
and safe entry zones according to the locations [[Figure 3]] and [[Figure 4]].
Figure 3: Schematic drawing illustrating the most common surgical approaches used for different
areas of the brainstem (Source: Giliberto et a/.)[13]
Figure 4: Posterior view of the brainstem showing the various safe entry zones (Source: Giliberto
et a/.)[13]
Midbrain
The ventral and central areas of the midbrain can be reached through a transsylvian
route with the classic pterional or the fronto-orbitozygomatic craniotomy with one
of its numerous modifications and the midbrain is approached through the transsylvian
route. The safe entry zone in this area is a narrow corridor lateral to the emergence
of CN III between the superior cerebellar artery (SCA) and the posterior cerebral
artery and medial to the pyramidal tract. Ventrolateral lesions of the midbrain can
be reached either through the transsylvian route or though the subtemporal transtentorial
approach. Ventrolateral lesions with a more caudal extension can also be approached
through more complex skull base transpetrosal approaches that afford a wider and more
lateral exposure for the lower midbrain, pons, and higher medulla.
Posterior midbrain is approached through median, lateral, and extreme lateral approaches
depending on the location of the lesion. The lateral mesencephalic sulcus is considered
the limit between the anterolateral midbrain and the posterior midbrain. Midline lesions
are approached through median supracerebellar infratentorial route, which allows an
adequate view of the posterior and posterolateral surface of the midbrain, quadrigeminal
plate, as well as the posterolateral surface of the upper pons. This approach includes
median, paramedian, and extreme lateral variants which provide access to different
parts of the posterior midbrain. The occipital transtentorial approach is an alternative
for patients with a steep tentorial slope. The supracerebellar infratentorial approach
requires a craniotomy exposing the entire width of the transverse sinus as well as
the confluence of sinuses to increase the angle of view by upward retraction of the
sinus. The lateral supracerebellar infratentorial approach requires a paramedian craniotomy,
again exposing the entire width of the transverse sinus. This provides access to the
posterior portion of the ambient cistern, including the proximal portion of the trochlear
nerve, the SCA, and the posterolateral aspect of the midbrain. The extreme-lateral
supracerebellar infratentorial variant is performed through a retrosigmoid craniectomy,
with full exposure of the transverse/sigmoid sinus junction. It allows for a more
lateral view of the posterolateral midbrain than the lateral approach. The lateral
mesencephalic sulcus is the safe entry zone in the posterolateral aspect. The lateral
mesencephalic vein runs into the lateral mesencephalic sulcus, thus representing an
easily identifiable surface landmark for this structure. In the medial posterior midbrain,
two safe entry zones at the level of the supracollicular and infracollicular areas
are identified. These are two narrow horizontal lines immediately above and below
the lamina quadrigemina.[[8]],[[10]],[[11]],[[13]],[[32]],[[33]],[[41]] [[Table 3]].
Table 3: Safe entry zones and surgical approaches to midbrain
Pons
The ventrolateral and lateral areas of pons are accessed through the retrosigmoid
approach and usually enter the brainstem between the trigeminal and facial nerves.
For more ventral lesions, this approach can be extended by anterior mobilization of
the skeletonized sigmoid sinus. The alternative routes to this area include the subtemporal
transtentorial route (for lesions with more rostral extension), the presigmoid route,
which provides a more lateral and direct view to the lesion or the transpetrosal approach.
The safe entry zone into the lateral pons is the so-called “peritrigeminal area” between
the emergence of CNs V and VII. This is an area located medially to the trigeminal
nerve and laterally to the pyramidal tract [[Figure 5]].
Figure 5: Peritrigeminal safe entry zone in the ventrolateral pons (Source: Giliberto et a/.)[13]
Dorsal pontine area is approached by either a Telovelo tonsillar or a vermian split
approach. The safe entry zone is through the floor of the fourth ventricle. They are
the median sulcus above the facial colliculus, the suprafacial triangle (located immediately
above the facial colliculus between the MLF and the cerebellar peduncles) and the
infrafacial triangles (located immediately below the facial colliculus, lateral to
the MLF, and is bordered inferiorly by the striae medullares and superolaterally by
the facial nerve). However, these safe entry zones are useful only if the lesion is
abutting the pial surface. In the cases, where the lesions are not approaching the
surface, intraoperative electrophysiological monitoring and mapping of the floor are
indispensable armaments to identify a safe corridor [[8]],[[10]],[[11]],[[13]],[[32]],[[33]],[[41]] [[Table 4]].
Table 4: Safe entry zones and surgical approaches to pons
Medulla
The ventrolateral medullary lesions can be resected through a far-lateral approach
through a lateral suboccipital craniectomy. For more ventral lesions, additional drilling
of the occipital condyle may be required to achieve optimal exposure. A safe entry
zone has been described in this region at the level of the retro-olivary sulcus or
between CN XII and C-1 in the anterolateral sulcus.
The upper part of the dorsal medulla is approached through the floor of the fourth
ventricle and is the same as the ones used to approach the dorsal pontine area. The
lower dorsal medulla is approached by a median suboccipital craniotomy. The three
safe entry zones for the posterior medulla: the posterior median fissure below the
obex, the posterior intermediate sulcus between the gracile and cuneate fascicles,
and the posterior lateral sulcus between the cuneate fascicle medially and the spinal
trigeminal tract and nucleus laterally [[8]],[[10]],[[11]],[[13]],[[32]],[[33]],[[41]] [[Table 5]].
Table 5: Safe entry zones and surgical approaches to medulla
Microsurgical technique-keypoints
Repeated hemorrhagic episodes cause enlargement of the lesion and pushes it toward
the pial surface causing a xanthochromic discoloration of the surrounding brain tissue
and a dark blue area corresponding to the area of bulging hematoma at the pial surface.
Such areas are the safest entry point for evacuation of BSC as there is no parenchymal
covering over them.[[14]] The hemorrhagic event can be divided into two categories: extralesional and intralesional
bleeding. The type of bleeding bears an effect on the surgical decompression. In extra-lesional
bleeding, decreasing the mass effect of the hematoma with the excision of cavernoma
itself and preserving the surrounding brainstem, is the easier than the intra-lesional
bleeding, where complete removal may induce injury of the surrounding brainstem. Therefore,
the extent of excision is controversial.[[14]]
The classic 2-point method can be used as an objective means to guide selection of
the surgical approach.[[40]],[[42]] In this technique, 1 point is placed in the center of the CM and the second, closest
to the pial surface/safe entry zone. The line connecting these 2 points is extended
to the skull; this trajectory is used to select the most optimal surgical approach.[[7]] If there is a discrepancy between the shortest trajectory and approach through
the safe entry zone, then, the latter should be preferred.
In most of the cases, the CN nuclei and white matter tracts are displaced by the lesions
and in many cases the lesions do not surface to the pia. Hence, neuronavigation and
electrophysiological monitoring techniques should be used to accurately locate the
critical structures intraoperatively. Intraoperative monitoring includes evoked potentials
such as SSEP, MEP, BAER, and direct CN monitoring. These provide real-time feedback
about the progression of surgery. A baseline reading after position before the start
of surgery is a must.[[10]],[[43]]
Lesion is approached through a parenchymal incision smaller than the lesion. Lesion
should be internally decompressed first, followed by an attempt to develop a cleavage
plane between the gliotic brain and lesion circumferentially. Piecemeal excision should
be done with extremely gentle traction. If any resistance is felt, then, further dissection
between the lesion and brain is resumed.[[43]]
Surgical outcomes
Surgical outcomes of BSCs with series more than 30 patients since 2010 have been analyzed
and summarized in [[Table 6]], of which the largest study of 397 patients was reported by Zaidi et al.[[15]] These 15 studies included a total number of 1666 patients. The complete excision
rate was 95% which was comparable with the meta-analysis. The early complication rate
was about 32%, with improvement noted in 52.3%, worsening in 10%, and mortality in
2% of the patients. These values too were comparable with the other studies.[[2]]
Table 6: Surgical outcomes of brainstem cavernous malformations with series of more than 30
cases since 2010
The surgery of BSCs is always fraught with morbidity as seen in the outcome summary.
The predictors of poor outcome that have been identified from the long-term studies
are age >40–50 years, poor preoperative status, large lesion size, multiple hemorrhages,
ventrally located lesions, presence of a DVA, postoperative rehemorrhage or second
operation and time of surgical intervention after 6–8 weeks of bleeding.[[15]] In a recent publication, Garcia et al.[[8]] proposed a grading system for BSCs to help predict good long-term outcomes for
patients undergoing resection. Proposed factors in the grading scale include patient
age, lesion size, presence of a deep venous anomaly, extension across the midline,
and hemorrhage, with a lower score indicating a greater likelihood of a good long-term
outcome. The proposed factors in a comparison study between the long-term outcome
between observation and surgery could give us a better indication about the correct
therapeutic management. However, the results of such a study cannot be taken into
account due to the selection bias involved between the cohorts. In the current case
series, two patients had immediate improvement postoperatively. One patient developed
a postoperative abducens nerve palsy which improved completely after 3 months.
Radiosurgery
Several studies report a decrease in hemorrhage rates after 2 years of radiosurgery.
It is a nonsurgical method to control hemorrhage in CM. However, the use of radiosurgery
is yet a controversial topic because the hemorrhage clustering in aggressive lesions
also abates after 2 years according to the studies in natural history. Many authors
believe that radiosurgery should be an alternative to observation but not to surgery,
especially in aggressive lesions. Moreover, SRS might induce the developments of de
novo CMs as well. Patients with deep-seated lesions which are surgically inaccessible,
patients who have surgical contraindications and aggressive lesions can be considered
for radiosurgery, but at marginal doses of 12–14 Gy to reduce procedure-related complications.[[10]],[[14]],[[28]],[[41]],[[46]]
Newer treatment modalities
Genetic analysis has revealed that CMs are linked to loss-of-function mutations in
the genes encoding any of three structurally distinct proteins: KRIT1 (aka CCM1),
OSM (aka CCM2), and PDCD10 (aka CCM3). Further studies have shown that these mutations
resulted in RhoA hyperactivation and endothelial instability. Thus, leading to the
hypothesis that reduction of RhoA hyperactivation with drugs, statins or fasudil,
ameliorated the pathobiology caused by these mutations in genes for CCM1 and CCM2
(evidence for a role of PDCD10 [CCM3] in a similar pathway is not as strong as for
the other 2 CCM genes), which has been shown in many animal studies. No human studies
have been done to validate this information till date.[[47]],[[48]],[[49]]
Another new treatment modality is propranolol. There are a few case reports that have
mentioned the usage of propranolol in patients with aggressive CMs. Apparently, this
drug controls recurrence, de novo evolution of lesions and rehemorrhage, however,
the pathophysiology behind the mechanism is yet unclear. Further clinical studies
are required to validate the efficacy of this pharmacological agent.[[50]],[[51]] Thus, these alternatives such as SRS or propranolol can be used in patients with
high surgical morbidity or any other significant surgical contraindications.
