Surgery of the Pons

Fig. 33.1

Illustration representing the different types of pontine tumors. Diffuse pontine gliomas (A) are invasive tumors that involve the pontine parenchyma and are the most common type. Focal intrinsic tumors (B) are well-circumscribed, intra-axial tumors. Dorsal exophytic tumors (C) arise from the dorsal pontine surface and expand into the fourth ventricle

Although pediatric brainstem tumors were historically lumped together into a single group, it is now understood that their behavior depends on their location and histopathology. The location and imaging characteristics now guide the treatment of pediatric brainstem tumors. In the same vein, tumors of the pons behave differently depending on their focality, histopathology, and location within the pons. Current surgical treatments include biopsy and open resection, while future surgical treatment likely will incorporate local drug-delivery techniques. In order to appropriately treat pediatric pontine tumors, an understanding of the tumor classification schemes, pathophysiology, and surgical options is necessary.

33.2 Classification Schemes

Several classification systems have been proposed to categorize brainstem tumors (Table 33.1). The earliest classification systems were based on the natural history without intervention combined with the tumor characteristics on computed tomography (CT) [5–7]. More recent systems have been primarily based on MRI appearance [8–11]. The simplest schemes classify tumors based on growth pattern, either focal or diffuse. The more detailed systems subdivide tumors based on location (midbrain, pons, medulla, cervicomedullary), whether an exophytic component is present, presence of hydrocephalus or hemorrhage, and contrast enhancement pattern.

Table 33.1

Classification systems for brainstem tumors

Author	Classification system
Epstein et al. [5]	Intrinsic
	1. Diffuse
	2. Focal
	3. Cervicomedullary
	Exophytic
	1. Anterolateral into cerebellopontine angle
	2. Posterolateral and into brachium pontine
	Disseminated
	1. Positive cytology
	2. Positive myelography
Epstein et al. [6]	Diffuse
	Focal
	Cervicomedullary
Stroink et al. [7]	Group I – dorsal exophytic glioma
	Group IIa – intrinsic brainstem tumors
	Hypodense, no enhancement
	Group IIb – intrinsic brainstem tumors
	Hyperdense, contrast enhancing, exophytic
	Group III – focal cystic tumor with contrast enhancement
	Group IV – focal intrinsic isodense, contrast enhancing
Barkovich et al. [9]	Location (midbrain, pons, medulla)
	Focality (diffuse of focal)
	Direction and extent of tumor growth
	Degree of brainstem enlargement
	Exophytic growth
	Hemorrhage or necrosis
	Evidence of hydrocephalus
Albright [8]	Diffuse
Albright [8]	Focal (midbrain, pons-intrinsic, dorsally exophytic, medulla)
Fischbein et al. [10]	Midbrain
	1. Diffuse
	2. Focal
	3. Tectal
	Pons
	1. Diffuse
	2. Focal
	Medulla
	1. Diffuse
	2. Focal
	3. Dorsal exophytic
Rubin et al. [11]	Cervicomedullary
	Exophytic
	Cystic
	Focal
	Diffuse
Choux et al. [12]	Type I – diffuse
	Type II – intrinsic, focal
	Type III – exophytic, focal
	Type IV – cervicomedullary

CT computerized tomography, MRI magnetic resonance imaging

The most recent radiographic classification, as proposed by Choux et al., is based on both CT and MRI data and classifies brainstem tumors into four types [12]. Type I tumors are diffuse and account for approximately 75 % of all brainstem tumors [8, 13]. Radiographically, they have poorly demarcated edges on CT and are usually hypointense on T1 MRI sequences (Fig. 33.2). They usually do not enhance with contrast. Histopathologically, they are most commonly malignant gliomas (WHO grade III or IV); however, diffuse tumors can be low-grade gliomas as well [12, 14, 15]. Type II tumors are focal, intrinsic masses that can be solid or cystic in appearance. The main distinguishing radiographic feature is their well-demarcated appearance relative to surrounding tissue (Fig. 33.3) [12]. Histopathologically, they are most commonly low-grade gliomas (WHO grade I or II); however, high-grade lesions have been reported to present in this way as well [4, 15]. Type III tumors are dorsal exophytic masses that originate from the subependymal glial tissue of the floor of the fourth ventricle [12, 16]. Radiographically, they have well-demarcated borders and demonstrate a dorsal and lateral growth pattern (Fig. 33.4). Histopathologically, they are commonly low-grade gliomas. It has been observed that the higher-grade exophytic brainstem tumors tend to have a lateral and ventral growth pattern compared to lower-grade exophytic lesions, which tend to extend dorsally [4, 17, 18]. Type IV tumors are cervicomedullary masses that present and behave very similar to intrinsic cervical spinal cord tumors [6, 12]. Histopathologically, most of these lesions are non-infiltrative, low-grade lesions whose growth pattern is limited rostrally by the white matter of the medial lemniscus and corticospinal tract [18].

Fig. 33.2

Diffuse pontine glioma as seen on sagittal (left) and coronal (right) T1 gadolinium-enhanced MRI sequences. The typical imaging characteristics are seen including heterogeneous tumor mass, poorly demarcated borders, and global expansion of the pons

Fig. 33.3

Focal intrinsic pontine tumor as seen on T1 gadolinium-enhanced sagittal MRI (left), T2 axial MRI (middle), and axial CT (right). It is notable that the tumor is fully within the pons and has well-demarcated borders in contrast to the global expansion of brainstem classically seen in diffuse pontine gliomas. In this case, the tumor has a cystic component, but focal intrinsic pontine tumors can appear solid as well. Pathological diagnosis was juvenile pilocytic astrocytoma

Fig. 33.4

Dorsal exophytic tumor as seen on sagittal (left), axial (middle), and coronal (right) T1 gadolinium-enhanced MRI sequences. The tumor is arising from the dorsal pontine surface along the floor of the fourth ventricle. It is notable that the tumor has well-demarcated borders and has grown dorsally and cranially. The mass effect exerted on the brainstem without parenchymal invasion is typical for these tumors. Pathological diagnosis was juvenile pilocytic astrocytoma

33.2.1 Presentation

Pontine tumors are distinguished by three characteristics: (1) signs and symptoms related to intra-axial location, (2) signs and symptoms related to hydrocephalus, and (3) timing of symptom onset. Whether a tumor is diffuse or focal affects each of these characteristics.

The most common signs and symptoms of diffuse pontine tumors are due to cerebellar dysfunction and cranial nerve palsies [19]. Cerebellar dysfunction can manifest as both appendicular and axial ataxia, although ataxia is not specific to diffuse tumors as it can commonly be seen in focal tumors [3]. Cranial nerve palsies are also common in both diffuse and focal tumors. Diffuse tumors tend to present with sixth and seventh nerve palsies, while focal tumors tend to present more commonly with diplopia and facial weakness [3, 14]. Interruption of the fibers of the corticospinal tracts resulting in the upper motor neuron syndrome is characteristic of pontine tumors [19]. These signs and symptoms include loss of fine motor skills, spasticity, initial hyporeflexia followed by hyperreflexia, and upgoing extensor plantar reflex.

Development of hydrocephalus is highly dependent on tumor location. Dorsal, exophytic tumors are found on the floor of the fourth ventricle and commonly produce obstruction of CSF flow as they grow. Although diffuse intrinsic tumors can produce enlargement of the brainstem, they rarely produce obstructive hydrocephalus [18, 20, 21].

The timing of symptom onset is dependent on aggressiveness of the tumor and whether hydrocephalus is present. Behnke et al. reported a series of 30 children with intra-axial, endophytic tumors of the pons and cervicomedullary junction. They noted that symptoms >6 months in duration were indicative of a benign brain tumor. With one exception, those patients did not have a PNET or malignant astrocytoma [14]. In patients with a history <6 months of duration, a malignant astrocytoma or PNET was much more likely. However, as has previously been demonstrated, patients with grade I and II astrocytomas can present with clinical symptoms of <6 months duration [14, 22]. In their study, 27 % of patients had hydrocephalus on presentation. Other studies report similar percentages of patients presenting with moderate to severe hydrocephalus [9, 14, 23, 24]. In contrast, only 26 % of patients in a series by Villani and co-workers presented with a history >6 months. However, the incidence of hydrocephalus was much higher (85 %) and possibly explained the differences in the two series [14, 25].

33.2.2 The Role of Biopsy

The use of biopsy in the management of tumors of the brainstem has come in and out of favor with technological advances and has changed with increased understanding of tumor behavior. Surgery as a whole was initially considered to play no role in the treatment of all brainstem gliomas. In 1939, Bailey postulated that “until some effective treatment other than surgery is devised, gliomas of the brainstem will be hopeless problems for treatment” [26]. Surgical techniques to approach eloquent brain structures were introduced prior to modern imaging techniques. In 1949, Spiegel and colleagues first reported the use of frame-based stereotaxy to perform a medial thalamotomy to reduce emotional reactivity [27]. This was the first description of the safe use of stereotaxis to approach deep-seated structures. The use of frame-based stereotaxy was then applied to approach the brainstem in a monkey model as reported by Ward in 1958 [28].

Despite early technical advances in approaches to the brainstem, surgical interventions, whether through biopsy or more open procedure, were not universally accepted. In 1967, Lassman and Arjona concluded that the diagnosis of pontine glioma could be made by ventriculography alone and that surgical management should be limited to control of hydrocephalus [23]. In 1969, Madson echoed Bailey’s early opinion when he stated that “regardless of specific histology, brainstem gliomas must be classified as malignant tumors since their location in itself renders them inoperable” [29]. Surgery would continue to not play a significant role in the management of brainstem tumors until after the advent of CT. The advent of CT allowed the brainstem anatomy and location of brainstem tumors to be studied with improved definition. The combination of CT and frame-based stereotactic techniques greatly facilitated biopsy of deep structures of the brain.

In 1978, image-directed biopsy of deep brain structures was first reported by Gleason and co-workers [30]. In the 1980s, multiple brainstem biopsy series reported a high diagnostic yield with low complication rate in both adult and pediatric patients [31–38]. The proponents of brainstem biopsy argued that a histopathological diagnosis provided important prognostic information and guided therapeutic decisions [39, 40]. Additionally, it was clear that brainstem tumors did not all exhibit the same behavior as once thought [5, 6, 13, 16, 41]. Epstein was the first to classify brainstem tumors based on location and focality, distinguishing more favorable focal lesions from less favorable diffuse lesions [5]. It appeared that with the combination of radiographic location and histopathology, empiric radiotherapy for a presumed brainstem glioma would no longer be necessary [42].

As MRI became increasingly available in the 1980s, it was noted that the resolution of normal anatomy and pathology of the brainstem was superior to CT. On MRI, the imaging characteristics finally reflected the heterogeneous nature of these lesions. As the experience with MR imaging of brainstem tumors increased, Barkovich et al. proposed the first classification scheme based on MRI criteria in 1990 [9]. Not only had MRI replaced CT as the primary imaging modality for brainstem tumors, certain groups were suggesting that it could replace stereotactic biopsy as the diagnostic tool of choice [43, 44]. Perhaps the most notable report favoring abandonment of biopsy in the workup of brainstem tumors came from Albright and co-workers [45]. Their report summarized the recommendations from the Children’s Cancer Group studying the effects of radiotherapy on brainstem tumors. Out of 120 patients, 45 (38 %) had either a stereotactic or open biopsy. All of these patients had either low-grade or high-grade gliomas on histopathology. Postoperative neurological complications occurred in five (11 %) of the operative cases. Due to the fact that MRI diagnosis correlated highly with diagnosis of glioma, it was recommended that MRI alone should be the diagnostic modality of choice for diffuse brainstem gliomas [45].

The shift in diagnostic criteria of brainstem tumors, from a combination of histological and radiological characteristics to purely radiological criteria, was reflected in several proposed brainstem tumor classification systems [5, 6, 8, 9, 12]. All of these classification systems made a distinction between focal and diffuse lesions. In 1986, Epstein and McCleary defined a diffuse lesion as any lesion greater than 2 cm. Their classification scheme was one of the earliest and perhaps the most influential. In their series of 34 patients, which served as the basis for their classification, 22 patients had lesions greater than 2 cm and were classified as diffuse. All of those patients had malignant gliomas (grade III–IV) [6]. Although CT was the primary imaging modality used to diagnose patients in this study, their definition of “diffuse” was more clearly defined than more recent studies.

In contrast, the study by Fishbein and colleagues defined diffuse tumors as being “poorly marginated, involving more than one half of the involved brain stem segment, or infiltrating both the segment of brain superior to and the segment inferior to the segment of origin” [10]. The two most important prognostic factors in their study were focality and location. In concordance with previous studies, patients with focal tumors had higher long-term survival than patients with diffuse tumors [6, 12, 16]. Likewise, patients with pontine tumors did worse than patients with tumors of the midbrain or cervicomedullary junction [11, 41]. However, 6 out of 29 patients with diffuse, pontine tumors were alive with a mean follow-up time of 2.6 years [10]. This is in stark contrast to the series reported by Epstein and McCleary, in which all 22 patients with diffuse brainstem tumors died within 6–12 months [6]. One possible explanation for the differences in these studies is that the classification criteria defining “diffuse tumors” was significantly different so that “diffuse” tumors in each study were not directly comparable. Another explanation is that the sample size of each study was not large enough to give an accurate representation of tumor behavior.

Despite the fact that it was not clear how to select the few patients with diffuse pontine tumors that had longer survival, it was clear that patients with diffuse pontine tumors generally did not benefit from aggressive surgery and that most patients had a dismal prognosis [13, 46]. Given the reported experience at the time, Albright concluded that “there is probably no other intrinsic brainstem tumor with the same MRI appearance as diffuse brainstem glioma, with the possible exception of hamartoma, which is exceedingly rare and generally presents when the patient is younger than 2 months old” [3]. This led to a paradigm shift in the early 1990s in the management of children with brainstem tumors. For those patients who presented with signs and symptoms consistent with a brainstem glioma and who had a diffuse brainstem mass, biopsy was no longer considered necessary and stopped being standard of care [3, 45, 47].

However, others have shown that the appearance of a diffuse brainstem mass does not always yield a histopathological diagnosis of malignant glioma. Roujeau et al. presented a series of 24 children with diffuse pontine tumors not amenable to surgical resection. Of the 24 patients who underwent stereotactic, frame-based biopsy, 22 children had malignant gliomas while 1 had a grade II astrocytoma and 1 had a grade I astrocytoma. Given these results, the two patients with low-grade tumors did not receive up-front radiation therapy. Upon progression, the patient with the grade I tumor underwent surgery to attempt resection, while the patient with the grade II tumor was treated with radiation therapy [48]. Behnke and co-workers reported a series of 30 patients with intra-axial, endophytic tumors of the pons and medulla who underwent open surgery for excision. Although most malignant astrocytomas did demonstrate a diffuse pattern of growth on MRI, there were five cases that had a focal appearance. Likewise, three out of ten grade I and all six grade II astrocytomas had a diffuse growth pattern on MRI [14]. Unlike the early experience with brainstem tumors, histopathologically benign tumors clearly can mimic malignant tumors radiographically [14, 15]. In 2010, Kumar and colleagues presented a series of 100 patients harboring grade I astrocytomas with features mimicking high-grade gliomas on MRI. Of these patients, 76 were pediatric patients and 24 patients had brainstem tumors [15].

Currently, the role of biopsy in the management of brainstem tumors is being revisited [47]. For the past 20 years, it has been argued that sampling tissue was unnecessary and harmful to patients harboring diffuse brainstem tumors given that a diagnosis could be made on clinical presentation and imaging alone [12, 45]. At present, it is clear that the appearance of a diffuse brainstem mass on MRI correlates to a high-grade glioma in the majority but not all cases [14, 15, 48]. Additionally, advances in molecular genetics and drug-delivery mechanisms raise the possibility of improved therapies and outcomes for patients who would otherwise have a dismal prognosis [46, 47, 49]. Obtaining tissue is of central importance to study the molecular and genetic foundation and to better understand the biology of these tumors to develop therapies. The dilemma of whether to biopsy brainstem tumors lies in whether it is preferable to perform a biopsy on a large number of patients in order to find the few patients with low-grade tumors versus empirically treating patients with diffuse pontine tumors with radiation, knowing that the management would likely be different if there was a known diagnosis of low-grade glioma. Ideally, tissue diagnosis should be obtained in combination with testing a novel therapy in a clinical trial [49]. In routine practice, however, performing a biopsy on patients with diffuse, pontine tumors remains controversial.

Only gold members can continue reading. Log In or Register to continue