Key-words:
Brainstem - cavernous malformation - hemorrhagic events - treatment strategy
Introduction
Cerebral cavernous malformations (CCMs) are proliferative hemorrhagic lesions containing
a cluster of vascular sinusoids caverns lined by endothelium and surrounded by gliosis
and blood degradation products.[[1]],[[2]],[[3]],[[4]] CCMs of the brainstem are particular forms of CMs associated with higher morbidity
and mortality rates than other CCMs [[5]],[[6]] due to the proximity of the lesion to critical neural structures and to the complex
blood supply in the region.[[7]],[[8]] Brainstem CMs (BSCMs) account for 15%–18% of intracranial CMs with annual re-hemorrhage
rates from 5.1% to as high as 30.8%. With regard to the rarity of BSCMs and the potentially
devastating consequences of bleeding, different therapeutic modalities have been proposed:
conservative, radiosurgery, or surgery. Even though indications for surgery remain
controversial, the surgical option appears the best option for treatment. Our series
analyzed 55 patients with BSCMs with the aim of studying the obtained results, viewing
them in terms of clinical and radiological outcome considering the surgical approach
chosen based on the lesion site, the surgeon's expertise, and data of pertinent literature.
Two groups were identified based on the approach performed: lateral and medial.
Methods
A total of 55 consecutive patients were reviewed retrospectively. The patients were
treated between January 2006 and March 2016 in three high-quality centers for the
treatment of CMs located in the brainstem or originating from the upper and lower
cerebellar peduncles and reaching the brainstem. The follow-up period was from 3 to
154 months (mean: 63.2 months). The patients included in our review had a histopathological
diagnosis of CM. Data collected included clinical assessments (on admission, the immediate
postoperative period, and at last follow-up), the number of hemorrhagic events, preoperative
imaging (magnetic resonance imaging [MRI] including tractography), operative technique
(the particularly surgical approach selection), and any complications (re-bleeding
and possible additional surgery). We examined early postoperative imaging (MRI within
3 days) and late control (1 year after surgery and every year during follow-up) to
define the resection quality (total/subtotal). The location of each BSCM was classified
as medullary (10 cases), pontine (22 cases), pontine/mesencephalic (12 cases) and
mesencephalic (11 cases), medial (6 medullary, 6 pons, 1 pontomesencephalic, and 8
mesencephalic), and lateral (4 medullary, 16 pons, 11 pontomesencephalic, and 3 mesencephalic).
We evaluated the patient's neurological status utilizing the modified Rankin Scale
(mRS) score. Demographic, clinical, and surgical data are summarized in [[Table 1]].
Table 1: Brainstem cavernous malformations: patient demographics, neurological aspects, lesion
location, and modified Rankin Scale score
Table 1: Contd...
[[Figure 1]], [[Figure 2]], [[Figure 3]], [[Figure 4]] show some illustrative cases.
Figure 1: Case report of a young patient with diplopia (VI nerve paresis). A T2-weighted-magnetic
image showed a cavernous malformation with lateral extension in pons-mesencephalon-junction
with hemorrhagic signs (a-c). A retrosigmoid approach was proposed, but the patient
preferred to delay. After 9 months, the patient presented progressive dysphagia, dysphonia,
and ataxia. A new magnetic resonance imaging showed a large cavernous malformation
pontomesencephalic (T1-weighted magnetic with gadolinium) (d-f). The patient had V,
VI, VIII, and IX severe cranial nerve paresis, VII mild cranial nerve paresis, sensory
disturbances, ataxia, and moderate-to-mild disability (modified Rankin Scale score:
2). Intraoperative neurophysiological monitoring was used and a median approach (transvermian)
was performed. A postoperative magnetic resonance imaging showed complete removal
of brainstem cavernous malformation and initially, the patient was stable, and at
last follow-up, presented an improvement of neurological status (g-i)
Figure 2: Case report of a middle-aged patient with onset of coma. Preoperative magnetic resonance
imaging (a-c) showed a large cavernoma in pons. A medial telovelar approach was performed
with intraoperative neurophysiological monitoring (motor, sensitive, and cranial nerve
function) and external pacemaker. The patient did not present improvement of neurological
status. Postoperative magnetic resonance imaging (d-f) showed complete removal of
cavernous malformation
Figure 3: Case report of a middle-aged patient with ataxia, right hemiplegia, and sensory disturbance.
Preoperative magnetic resonance imaging indicated a left pontomesencephalic cavernous
malformation. A subtemporal approach was performed with intraoperative neurophysiological
monitoring. Postoperative magnetic resonance imaging showed complete removal and no
left temporal lobe damage. Neurological status was completely recovered at 30 days
Figure 4: Case report of a 20-year-old patient with dysphagia, VII nerve palsy, and ataxia.
Preoperative magnetic resonance imaging (a-c) showed a small medullary cavernous malformation.
Intraoperative external pacemaker and neurophysiological monitoring were used. A far-lateral
approach was applied and complete removal and resolution of symptoms were obtained
All procedures performed in studies involving human participants were in accordance
with the ethical standards of the institutional and/or national research committee
and with the 1964 Helsinki Declaration and its later amendments or comparable ethical
standards. For this type of study, formal consent is not required.
Surgical technique
Indications for surgery were as follows:
-
Large acute and subacute hemorrhage or large diameter of BSCM
-
Severe progressive or worsening of neurological deficits due to repeated hemorrhagic
events or enlargement of the BSCM
-
More than one hemorrhagic event
-
Lesion close to the pial surface, exophytic, or surgically accessible following the
preoperative planning.
We did not consider patients for surgery if asymptomatic, or if they had experienced
only one hemorrhagic event. We recommend delaying surgery, where possible, for 2–3
weeks, following a symptomatic hemorrhage.
We considered all available approaches to the brainstem to remove CMs. We selected
six main approaches based on the shortest trajectories to the BSCM and the senior
surgeon's experience. We performed the following approaches in our series:
-
Anterior (orbitozygomatic and pterional) for lateral ventral mesencephalic BSCM
-
Retrosigmoid for ventrolateral pontine or medullary lesions
-
Far lateral for ventral pontine or medullary lesionsSuboccipital telovelar for medial
dorsal pontine (through the fourth ventricle) and medial dorsal medullary lesions
-
Supracerebellar infratentorial for medial dorsal mesencephalic lesions
-
Subtemporal approaches for lateral pontine, ponto-mesencephalic, and mesencephalic
lesions.
In 15 cases, neuronavigation and image guidance provided real-time anatomic localization
of the BSCM and its relationship to eloquent tissues, particularly for lesions that
did not reach the surface. A cardiac pacemaker was used in 13 cases. Neurophysiologic
monitoring, including somatosensory and motor-evoked potentials, was used in all cases.
Moreover, neurophysiologic mapping was used to check cranial nerve response. The aim
of this surgery was to plan the most appropriate approach, minimize possible surgical
complications, and choose the safest entry route.
The relationship between the cavernoma and pial or ependymal surface of the brainstem
helps to define the choice of surgical approach. The supra- and infra-facial triangles
at the floor of the fourth ventricle, the lateral mesencephalic sulcus, the peritrigeminal
area, and the inferior olivary nucleus are described as entry zones [[9]],[[10]] which allow an anatomical route to preserve neurovascular structures.
Since all of our cases were localized lower than 3 mm below the pial surface, we did
not have to select a real safe entry zone because the thin layer did not involve any
neurovascular structures, so the lesion itself dictated the entry point. The microsurgical
technique performed in all of our cases was the 2-point rule, introduced by Brown
et al.,[[11]] to choose the optimal angle of entry for each lesion already planned with the aid
of high-resolution preoperative images. All patients underwent computed tomography
scan postoperatively. Patients were maintained on mechanical ventilation in the Intensive
Care Unit for a minimum of 24 h following the surgical procedure.
Statistical analysis
The data values are summarized as median, mean, and range for continuous variables
and as frequency and percentage for categorical variables. Statistical analyses of
categorical variables were carried out using the Fisher's exact tests for linear association
as appropriate and nonparametric tests (median values and ranges as well as numbers
and percentages). P ≤ 0.05 was considered statistically significant. We compared the
mRS score for each surgical approach used and formed them into two main subgroups
according to lateral and medial locations of BSCM. When the ΔmRS was higher than 0,
it indicated neurological improvement; ΔmRS = 0 means stability of neurological status
and ΔmRS <0 means worsening of neurological status.
Results
Fifty-five patients (23 males and 32 females) were included in our series. The mean
age was 40 years (range: 16–70 years). The principal localizations were medulla in
18% of cases (10 patients), pons in 40% (22 patients), ponto-mesencephalic junction
in 21% (12 patients), and mesencephalon in 20% (11 patients).
The onset of signs and symptoms included cranial deficits in 74.5% of cases, motor
deficits in 40% of cases, sensitive deficits in 22% of cases, ataxia or dysmetria
in 12% of cases, and an altered level of consciousness in 1.8% of cases. The main
subjective symptoms were headache, dizziness, nausea, and/or vomiting.
In our series, 22 patients complained of 1 hemorrhagic event. Surgical procedure was
considered for these patients when there was a complaint of neurological deficit and
when the size of the CM, or hemorrhage, was significant. Two hemorrhagic events occurred
for 29 patients, while >2 events occurred for the remaining 4 cases (3, 4, 5, and
7, respectively). A surgical procedure was selected for those cases, in accordance
with our indications.
Clinical, demographic, and surgical data relative to location are summarized in [[Table 2]].
Table 2: Clinical, demographic, and surgical 10.325 data according to location in brainstem
of cavernous
A gross total resection was achieved in 46 patients (83.6%). The most faceable sites
to reach this result were the ponto-mesencephalic-junction (100%) and the medullary
(90%).
Subtotal resection was achieved in nine cases. Three patients out of the nine underwent
a second surgery: 1 patient to remove an important residual lesion through a surgical
access, 1 patient because of the persistence of preoperative symptoms, and 1 patient
because of a second hemorrhagic event. We observed an improvement in the neurological
status for all patients who underwent an additional surgical procedure. Six patients
of the nine with residual CM were not considered for a second surgery due to the small
size of the residual lesion or because they were asymptomatic. Three of these did
not re-bleed at last follow-up and did not show any neurological change. The remaining
three patients re-bled: one showed worsening in neurological status, while the two
remained stable.
In summary, we observed improvement in the subtotal removal group (nine patients)
of neurological status and mRs score at last follow-up in three patients who underwent
an additional surgical procedure. The five patients who were not re-operated on remained
stable at follow-up: two notwithstanding re-bleeding and one patient worsened after
re-bleeding and conservative treatment.
The approaches were chosen depending on the localization of the lesion.
The BSCMs were medial in 21 patients, and the approaches selected were telovelar,
transtentorial, infratentorial, and transvermian. In 34 cases, the lesion was lateral
and the approaches performed were anterior, retrosigmoid, far-lateral, and subtemporal.
We observed a neurological improvement in 70% of patients who underwent lateral approach
and a neurological worsening in 8.8% of cases. Regarding the medial approaches, we
observed neurological improvement in 47% of cases and neurological worsening in 14%
of cases.
We compared lateral and medial approaches to the BSCMs in relation to the mRS score
(postoperative and at clinical follow-up) [[Table 3]].
Table 3: Statistical data of two subgroups: lateral and medial location of brainstem cavernous
malformations
Significant statistical data were obtained with P = 0.0086, which indicates that the
lateral approach, in our series, is associated with a better neurological and clinical
outcome.
At the last follow-up (3–154 months), we observed an improvement in the mRS score
in 34 cases, a decline in the neurological status in 5 patients, while 16 patients
remained stable.
Discussion
The correct management of BSCM requires careful knowledge of the epidemiology, natural
history, and clinical presentation. Patients affected by BSCMs are usually symptomatic,
mainly because of hemorrhagic event or due to the size of the CM.[[5]]
Despite our increasing knowledge regarding natural history, there is currently no
available treatment algorithm for cavernomas. Three treatment modalities (observation,
microsurgery, and radiosurgery) have been discussed in the literature, but their indication
criteria are yet to be defined. Frischer et al.[[12]] reported on a treatment model in which the microsurgical resection is suggested
for symptomatic lesions with suitable operative corridors, gamma knife radiosurgery
for surgically high-risk deep-seated lesions,[[13]] and conservative management of asymptomatic lesions. In our opinion, as Almefty
and Spetzler [[5]] reported, there is no available literature on the use of gamma knife radiosurgery
for CMs. Hence, we recommend conservative management for all the lesions considered
unsuitable for resection. With regard to microsurgical treatment, one of the most
important aspects related to the natural history of BSCMs is their re-hemorrhagic
rate following an initial hemorrhage. Contrasting data are reported in the literature
regarding re-hemorrhagic rates. In 2010, Abla et al. reported a series of 260 patients
with a rate of re-hemorrhage >30%.[[14]] In 2015, Starke reported a re-hemorrhagic rate for BSCMs which varied from 5% to
21.5%.[[15]] In 2016, Walcot et al. observed an estimated 5-year recurrent hemorrhagic risk
of 30.8%, with a 50.7% risk of developing either a recurrent hemorrhage or a new neurological
deficit unrelated to hemorrhage.[[8]]
Of the 55 patients reviewed, the bleeding rate in female patients was higher than
that for male patients (female-to-male ratio – 32:23), probably indicating a hormonal
effect on the cavernoma.[[16]] We observed a single hemorrhagic event in more females than males (female/male
ratio – 14:8), while multiple hemorrhagic events were found equally in both sexes
(female/male ratio – 18:15). These data suggest the same chance of re-hemorrhage rate
after an initial hemorrhage despite a predominant female population with a single
hemorrhagic event. The frequency and number of hemorrhages are essential in the decision-making
process. In fact, re-hemorrhaging of a BSCM increases the rate and severity of neurologic
deficits or the worsening of preexisting symptoms and occasionally, it is the cause
of death. The hemorrhage rate should be compared with clinical symptoms and signs
when deciding upon surgical strategy. The surgical procedure should be scheduled as
soon as possible so as to prevent re-bleeding and reduce perioperative risks.
Bertalanffy et al.[[17]] reported that symptomatic lesions on the pial or ependymal surface can be considered
for resection, while asymptomatic lesions should be managed conservatively. Petr and
Lanzino [[18]] agree with the literature in considering surgery for patients who have experienced
at least two symptomatic bleeds and for lesions that come close to the pial surface.
In 2015, Horne et al. made a very important contribution to our knowledge of the natural
history of CMs [[19]] reporting on a large meta-analysis providing strong evidence for long-held suspicions
that brainstem location and hemorrhagic lesions portend a more aggressive clinical
course.
In our series, a surgical procedure was performed after the careful analysis of each
single case, based on size, location, multiple hemorrhagic events, and consequently
signs of worsening of neurological status.
In 1999, the Barrow Neurological Institute reported on a consecutive series of 100
patients with BSCMs. The authors suggest that in symptomatic hemorrhage patients a
standard skull base approach should be performed when lesion reaches the pial surface.
In this series, the standard skull base approaches were practiced in 86% of the cases.[[2]] Nearly 87% of the patients were stable or better at follow-up, 10% were worse,
and 4% died. Similarly, in our series, 29% of patients remained stable, 62% were better
at the follow-up, 9% were worse, and none died. We agree with the literature that
CMs localized on the surface of the brainstem are more easily approachable. In our
series, we did not observe a distance between ependymal surface and CM >3 mm that
offered a direct and safe entry route due to a thin parenchymal layer. When critical
neural structures are sparse and perforating arteries are involved along surgical
corridors, a safe entry zone has to be used, as well described in the literature.[[9]],[[10]],[[14]],[[20]] We selected skull base approaches in 62% of the cases, our analysis compared lateral
and medial approaches in terms of outcome. A statistical significant value evidenced
that the lateral approaches are associated with a better outcome.
Another controversial aspect involving BSCMs is an indication of a second surgical
procedure. In 2003, Wang et al.[[21]] reported a series of 137 cases in which, at follow-up, a total cavernoma resection
was obtained in 96% of cases and a subtotal removal was achieved in 4%. Moreover,
the rate of re-bleeding was 2.3% and three patients underwent a secondary operation.
The authors suggested that the high bleeding and re-bleeding rate of BSCMs would have
resulted from its biologic uniqueness.
In our series, a total resection was achieved for 46 patients (83.6%), a subtotal
removal in 9 cases (17.4%), and the re-hemorrhage rate was 7% at follow-up (48 months).
We considered a second operation for four patients. Surgical criteria to second treatment
were relevant residual lesion, persistence of preoperative symptoms, and re-bleeding.
We were conservative in nonsymptomatic and nonhemorrhagic cavernomas or residual in
brainstem, poor clinical status, and when the risks of surgery outweighed the benefits.
We analyzed the number of hemorrhagic events and the timing (days) from last bleeding
to surgery, both resulted similar to data reported in the literature [[Table 4]].
Table 4: Relation between hemorrhagic event and timing (days) from last bleeding to surgery
Garcia et al. (2015)[[22]] highlighted that the number of BSCM hemorrhages is beyond the surgeon's control
as well as being difficult to assess, whereas the timing of surgery relative to hemorrhage
can be arranged by the neurosurgeon.
Numerous technological advances have been carried out over the past two decades to
reduce the intraoperative risks in this kind of surgery.[[23]],[[24]],[[25]] Intraoperatively, we utilized the navigation system with fusion of preoperative
magnetic resonance, neurophysiological monitoring, and, in selected cases, a pacemaker
was placed preoperatively.
Limit: This is a retrospective study, notwithstanding our results are now confirmed
by the new cases.
Conclusions
As our analysis suggests, we recommend preference for a lateral approach where possible
to reach BSCM. Moreover, according to evidence for a better outcome and lower re-bleeding
risk, we suggest performing a second surgical procedure when a residual BSCM is present.
However, each patient needs to be evaluated individually, in view of their clinical
history, lesion's location, and the surgeon's technical expertise.
