Historical Landmarks

Pioneering neurosurgeons in the end of the 19th century and at the dawn of the 20th century, working without the benefit of imaging modalities, relied exclusively on anatomic knowledge and clinical findings to strategize their operative approaches. The first person to use a device for localizing brain structures was D. N. Zernov in 1889. A professor of anatomy in Moscow, Russia, he invented an apparatus that could be fastened to the head and adjusted to localize points that could then be identified on a map that showed the average positions of cranial structures.¹ Zernov first used this device to identify and drain an abscess in the interparietal sulcus. Although his work was continued by others in Russia, his method did not become well known worldwide.² Without being aware of Zernov’s method, Victor Horsley and Robert Clarke described a “steretotactic” method of localizing brain structures in animals with a frame based on a three-dimensional coordinate system in 1906 (Greek stereo meaning three-dimensional; taxis meaning to move toward).³ However, because subcortical structures could not be reliably localized, stereotaxy in humans was not widely adopted until intracranial imaging was available.²

It was not until 1947 that the first human stereotactic system, termed the encephalotome, was demonstrated by Spiegel and Wycis.⁴ By this time, x-ray technology combined with the development of ventriculography, a method whereby air or contrast material is introduced into the ventricles to facilitate clearer imaging on x-ray plates, made it practical to take films in the operating room and develop them in a timely manner.⁵ The underlying principle is to target an instrument toward a particular neural structure based on information regarding its location obtained on the preoperative images. Successful completion of this task required that the instrument involved be guided stably toward the target in a coordinate system common to both the preoperative images and the in vivo structures. Before image acquisition, a frame was firmly attached to the patient’s skull to produce images with markers that are referenced to a frame-based coordinate system.² By using the same coordinate system during the operation, surgical instruments could be aimed toward the target structure shown on the images.

The pioneering development of the encephalotome led to development of other stereotactic frames, and with the introduction of CT in 1973^6,7 and MRI in 1980,⁸ frame-based stereotactic guidance took a firm place in the conduct of neurosurgical operations.^2,5 These frame-based systems were initially used to generate lesions in different neural structures for treating movement disorders, pain, and epilepsy and then evolved for performing drainage of abscesses and biopsy of tumors.^2,9 Although frame-based systems have proved to be reliable and accurate targeting devices,^10,11 they have several disadvantages that limit their utility²: (1) attachment of the frame to the patient’s skull often needs to be done on the same day as the operation, especially when frame application requires general anesthesia, (2) posterior fossa lesions are difficult to target, (3) the frame can limit access to the surgical field, (4) guidance is limited to a single target or only a few targets, and (5) maneuvering the apparatus during surgery can be unwieldy.² For these reasons, frameless stereotaxy has been developed to improve the efficacy and accuracy of targeting systems.

Background

Frame-Based Stereotaxy

Numerous geometric systems exist on which stereotactic frame-based guidance systems have been created. Four systems have primarily been used in the past: polar coordinate, arc-radius, focal point, and phantom target. The polar coordinate system is based on the necessity that guidance of the probe be described with regard to the angle relative to a skull entry point. Usually, a minimum of two angles (in planes orthogonal to each other) is required to characterize a unique trajectory; in addition to these angles, the distance of the probe to the target lesion needs to be calculated.

The arc-radius system is probably the most commonly and currently used geometric system—it serves as the basis for the Leksell frame. Arc-radius systems are based on the concept that a probe equivalent to the radius of a semicircular arc will reach the center of the arc when introduced perpendicular to any point along the arc. Adjustment of the arc permits targeting of different points.

The Leksell frame (Fig. 114-1), introduced in 1949, is based on the arc-radius principle and is one of the more commonly used frame-based systems in modern neurosurgery. A frame is rigidly attached to the patient’s head to which the semicircular arc is mounted by pivoting shafts that in turn are connected to the graduated bars of the frame. Ultimately, the arc can be manipulated for vertical and anterior-posterior adjustments to permit differential targeting by moving the center of the arc.

FIGURE 114-1 The Leksell frame primarily consists of a rigid frame base attached to the patient before any prebiopsy imaging. Attached to the frame is a semicircular arc that can be manipulated for vertical and anterior-posterior adjustments to allow differential targeting.

(From Amundson EW, McGirt MJ, Olivi A. A contralateral, transfrontal, extraventricular approach to stereotactic brainstem biopsy procedures. Technical note. J Neurosurg. 2005;102:565-570.)

The frame consists of four graduated bars that define the corners of a cube; on the left/right sides of the frame (in relation to the patient) there is a diagonal bar on each side connecting the corner of these posts to create an “N.” Once the patient is scanned (CT or MRI), the geometric center in the anterior-posterior plane is defined by looking at an axial image and drawing an “X” by connecting the corners of the frame. The center in the vertical plane is assessed by a line connecting the midpoint of the corner posts going through the “N.” After the center is defined, the location of the target in relation to the geometric center can be calculated and the semicircular arc manipulated (Fig. 114-2) accordingly to place that lesion at the center of the arc.

FIGURE 114-2 Use of a Leksell frame for biopsy of a brainstem tumor.

Frameless Stereotaxy

Principles of Neuronavigation

For both frame-based and frameless systems (Fig. 114-3), the relationship between the coordinate space for the preoperative images and that for the surgical field must be established. In frame-based systems, by using the same frame for both preoperative image acquisition and performing the surgery, the relationship between the two coordinate systems is known and no further transformation is required. For frameless systems, point-pair registration or surface contour registration can be used to establish the relationship between the preoperative images and that for the surgical field.

FIGURE 114-3 Image depicting a frameless stereotactic biopsy system that relies on neuronavigation systems for lesion targeting.

(From Woodworth GF, McGirt MJ, Samdani A, et al. Frameless image-guided stereotactic brain biopsy procedure: diagnostic yield, surgical morbidity, and comparison with the frame-based technique. J Neurosurg. 2006;104:233-237.)

Point-Pair Registration

In frameless systems, the relationship between the two different coordinate systems, one related to the image and the other related to the patient during surgery, requires a patient-to-image registration procedure to create a mutual relationship.² Point-pair matching was the first and has been the most common method to achieve transformation between the two coordinate systems.^12,13 This method requires a set of at least three non-colinear points to be defined in the coordinates of the images. These points are known as fiducial points and can consist of either natural anatomic landmarks (i.e., nasion, lateral canthus, or tragus), skin-applied markers, or bone-implanted markers.² These same fiducial points are defined again within the coordinate system of the surgical field in the operating room. Software is then used to establish the relationship between the coordinates of the fiducial points in the image space and their counterpart in the surgical space.²

Surface Contour Registration

As an alternative to fiducial point-pair registration, surface contour registration and inherent curvature matching of the head have also been used for coregistration. In surface contour registration, mapping a radiographic surface can be achieved by touching a number of multiple random points (termed cloud of points) or by scanning a surface with laser registration. These surface-based algorithms allow the use of imaging obtained before the intent to operate.

In a comparison of natural landmarks and skin-applied fiducial markers, navigation based on registration of natural landmarks alone has generally been slightly less accurate, in large part because they consist of less discretely defined points. Golfinos and coauthors reported that anatomic landmark and surface-fit algorithms have an accuracy of half that achieved with fiducial-based registration.¹⁴ To this day, the sophistication of surface registration strategies continues to evolve. Although surface matching can be used as an alternative to fiducial mapping, its use in combination with fiducial point-pair matching may be an optimal strategy to further improve reliability in stereotactic localization for biopsy.

Once the spatial relationship is established with either point-pair registration or surface contour registration, it is used throughout the remainder of the operation to map anatomic targets of the patient to that of the preoperative images. This thus requires that the patient’s head be immobilized with a head clamp and the operating table not be moved during surgery. To overcome the error that may be induced with patient or operating table movement, tracking units attached to the patient’s head¹⁵ or a surgical head clamp¹⁶ can be used. These sets of reference markers, termed dynamic reference frames, can be used to correct for any unwanted movement. Markers attached to a pointing device or, alternatively, a microscope can be tracked to locate position and orient the axis given these fixed dynamic reference frames.

Indications for Stereotactic Biopsy

The significant development of intracranial imaging over the past few decades has allowed much earlier diagnosis of brain tumors. Although some tumors have a characteristic appearance on imaging, no imaging modality is yet able to provide sufficient diagnostic information to direct subsequent therapy. As a result, tissue biopsy remains indispensible in many cases. The goal of biopsy is to provide a representative sample for pathologic diagnosis to guide subsequent treatment, which can include cytoreductive surgery, radiotherapy, or chemotherapy.¹⁷

The main stimulus for the adoption of stereotactic biopsy over an open operative procedure is to achieve a higher rate of diagnostic accuracy while minimizing potential adverse effects. Diagnostic accuracy is important for dictating appropriate adjuvant therapy. Particular characteristics of the tumor that favor the use of stereotactic biopsy over open biopsy include (1) lesions that do not exert a symptomatic mass effect or that are not curable by surgical excision, such as metastases or malignant intrinsic brain tumors (Fig. 114-4); (2) deep-seated lesions or those occupying space in eloquent cortex or deep nuclei (i.e., basal ganglia, thalamus), where open resection would lead to unacceptable morbidity/mortality; and (3) infiltrative lesions (i.e., gliomatosis cerebri) that do not have a clear brain-tumor margin and are unlikely to be excised completely without significant loss of normal brain parenchyma.¹⁷ Moreover, if the lesion’s appearance on imaging or the course of the disease suggests an alternative cause such as an infectious or demyelinating process rather than a neoplastic one, stereotactic biopsy is a more appropriate first step than a large open procedure.¹⁷

FIGURE 114-4 A heterogeneously enhancing lesion traversing the posterior corpus callosum for which frameless stereotactic biopsy was used; the final pathology was glioblastoma multiforme.

Nevertheless, although stereotactic biopsy is a relatively safe procedure, there are relative contraindications. For example, vascular tumors such as metastatic renal cell carcinoma, choriocarcinoma, or metastatic melanoma should not be approached stereotactically, if possible, because of their inherently higher risk for hemorrhage.¹⁷ In such cases of metastatic tumor, a stronger effort to biopsy and diagnose the primary neoplasm instead is generally recommended. Furthermore, location of the tumor close to a major blood vessel, the vessel-rich sylvian fissure, the cavernous sinus, or the brain-pial border should also alert the neurosurgeon to the more elevated risk for hemorrhage.

Technical Considerations

Complications of Stereotactic Biopsy

The most common complication of stereotactic biopsy is intracranial hemorrhage, which is frequently due to damage to blood vessels in the trajectory of the biopsy needle. In addition, damage to pathologically friable vessels within the targeted neoplastic lesion can account for a number of hemorrhagic complications. Reported rates of hemorrhage during stereotactic biopsy procedures range from 0% to 11.5%.^9,20,21 Although most cases of blood vessel damage are minor, catastrophic consequences can take place if a lesion such as a vascular malformation is accidentally subjected to biopsy.¹⁷ Some metastatic tumors such as melanoma, choriocarcinoma, and renal cell carcinoma are more likely to lead to hemorrhage and should be approached more cautiously. Preoperative MRI studies should be studied closely before surgery to identify the most optimal approach. In the case of a hemorrhagic complication with mass effect, craniotomy may be necessary for evacuation of the hematoma.

A second, less frequent complication of stereotactic biopsy is a new neurological deficit secondary to trauma to the surrounding brain parenchyma as a result of direct damage by the needle biopsy or the resulting edema.¹⁷ Brain penumbral damage can be limited by reducing the number of needle passes, and edema in some cases can be prevented by avoiding bridging veins, which in rare instances can be spotted on preoperative MRI. In some cases, preoperative corticosteroids can prevent exacerbation of edema in lesions already associated with significant edema.¹⁷

Other morbidities associated with stereotactic biopsy include seizures and infections. However, in view of the cost-benefit analysis and the extremely low risk for epileptic complications, we do not routinely use antiepileptic prophylaxis for our biopsies.

A review of 3866 stereotactic biopsies performed in the 1980s placed the overall mortality rate at 0.6% and morbidity rate at 3.2%, divided as follows: hemorrhage, 1.9%; neurological deficit, 0.8%; seizure, 1%; and infection. 0.5%.¹⁷

Comparison of Frame-Based and Frameless Methods

With the increasing popularity of frameless biopsy, several studies have sought to compare frame-based and frameless biopsy techniques.^22–25 Studies assessing the technical accuracy of frame-based versus frameless stereotactic localization have found acceptable ranges of accuracy when summing localization vector errors in models.²⁶ Clinical studies have shown no difference in the diagnostic yield or diagnostic accuracy (concordance between initial frozen section diagnosis and final open resection diagnosis) of either technique.^24,27,28 The most recent estimate of 391 procedures done over a 10-year span (227 frame based and 164 frameless) reported a diagnostic yield of 89.4% with no significant differences between frame-based and frameless methods.²² The reported diagnostic accuracy for either technique ranges between 51% and 78%.^19,24,28,29 Series reporting lower diagnostic accuracy generally involved patient populations with larger and more complicated tumor morphologies.²⁸ We have seen that increasing volumetric tumor size correlates with decreased concordance between biopsy and resection specimens. In one series performed at our institution, for a third of gliomas larger than 50 cm³, biopsy specimens were insufficient for accurate grading altogether.²⁵ An increased volume of central necrosis and the higher degree of tissue heterogeneity seen in more aggressive gliomas may be responsible for such outcomes.

The nonconcordance noted in studies comparing biopsy and open-resection pathologic diagnosis was most often secondary to misinterpretation of radiation-induced necrosis from recurrent glioma specimens.²⁴ In such series, undergrading of high-grade astrocytic tumors was the most common occurrence. It is suspected that targeting stereotactic biopsy away from hypointense areas of necrotic tissue yields a higher chance of encountering tissue of lower grade appearance. Conversely, nonviable (and therefore nondiagnostic) tissue is often encountered if the biopsy is limited to highly necrotic areas of the neoplasm. Diagnostic yield and accuracy may increase with the number of tissue samples.²⁴ However, the increased sampling has to be balanced with the potential for brain shift and increased risk for hemorrhagic complications.

On comparison of complication rates, transient or permanent morbidity, or mortality after either type of procedure, negligible significant differences have been reported between frame-based and frameless techniques.^25,30 One series examining 76 cases of frameless and 79 cases of frame-based biopsies found that operating time and time under anesthesia were significantly shorter in frameless series.²³ The frameless series also had a significantly lower rate of complications (14%) than did the frame-based series (22%).²³ However, some of these findings might be related to the fact that frame-based techniques were used earlier when image-guidance support was not as accurate as in recent years and the surgical technique not as refined. Another series of 110 frame-based and 52 frameless procedures showed equal efficacy and similar complication rates.³¹ Those who advocate frameless stereotaxy cite the reduced operating time needed and the tendency for earlier discharge.²² Overall, these studies demonstrate that both frame-based and frameless techniques are efficacious in obtaining a tissue biopsy specimen, with the resulting complications probably secondary to the experience of the neurosurgeon and the instrumentation at that particular institution.

Stereotactic Biopsy of Brainstem Lesions

Lesions within the brainstem (Fig. 114-5A and B) have long been considered challenging to diagnose and treat. Although brainstem tumors represent only about 2% of all intracranial tumors in adults as compared with about 10% to 15% in the pediatric population,³² radiologic diagnosis of brainstem lesions in adults is inaccurate 10% to 20% of the time. Most brainstem tumors in children, in contrast, are gliomas, which arguably can be diagnosed with MRI alone.³² Anatomically, brainstem lesions are often classified as subsiding within either the midbrain or pons and occasionally reaching into the medulla. In general, because most adult brainstem tumors are not amenable to surgical excision, stereotactic biopsy is important for obtaining a pathologic diagnosis. Histologic diagnosis often enables replacement of empirical treatment modalities with more specific therapies, as well as determination of a more accurate prognosis.

FIGURE 114-5 A and B, Example of a lesion in the pons/cerebellar peduncle for which a frame-based stereotactic biopsy system was used.

Image-guided frame-based stereotactic biopsy of brainstem lesions has been performed since 1975,³³ and the high diagnostic and low complication rates are comparable to those seen with supratentorial lesions.³⁴ As with supratentorial lesions, there have been several reports demonstrating variable concordance between the actual findings on biopsy and preoperative assessments from imaging.^34,35 For example, in the case of nonenhancing diffuse astrocytoma, half of those presumed to be low-grade glioma based on their characteristics on MRI were found to be anaplastic gliomas.^34,36 In another study of 26 patients, MRI combined with positron emission tomography was concordant with biopsy findings in only 63% of cases, thus suggesting that stereotactic biopsy is necessary to obtain the correct diagnosis.^34,37

A recent literature review of 469 cases of stereotactic brainstem biopsy in adults performed between 1975 and 2004 showed a diagnostic rate of 94.8%.³⁴ In terms of complications, 6.6% have transient or mild symptoms and 1.8% have permanent deficits.³⁴ Although the only noted mortality from this brainstem biopsy review occurred in a patient with capillary telangiectasia,³⁸ it is notable that biopsies of many other vascular malformations, with most of them being cavernomas, have been performed without hemorrhage or complication.³⁴ When hemorrhagic complications in the biopsy bed occur, prevention of postoperative swelling and close neurological observation are indicated.

Another factor favoring the decision to perform biopsy is the benefit of potential treatment. This benefit is significant given the great diversity of brainstem lesions. High-grade gliomas represent only 31% of cases. Low-grade gliomas and metastases account for 25% and 11% of cases, respectively.³⁴ The remaining lesions consist of hematomas, lymphomas, abscesses, demyelination, radiation necrosis, infarcts, and other lesions.³⁴ Patients most likely to benefit from stereotactic biopsy are those who are given a better prognosis based on biopsy results or who are spared a course of debilitating therapy.³⁴ Such patients include those with radiation necrosis, chemotherapy-sensitive metastasis (breast, choriocarcinoma), a lymphoma, or an abscess rather than a malignant glioma.³⁴

A number of approaches are available for stereotactic biopsy of an infratentorial brainstem lesion, including the transtentorial approach, the ipsilateral transfrontal trajectory, and the suboccipital transcerebellar route.³⁹ Although all these approaches have been used successfully to reach lesions located within the midbrain, pons, and medulla, they all have limitations.

Approaches

The Transtentorial Approach

The transtentorial approach, which requires entry from the posterior cerebrum to reach the upper pons, has virtually been abandoned because it necessitates crossing the pia superior and inferior to the tentorium and places vital vasculature and cranial nerves at risk. Additionally, biopsy through this route may cause pain with tentorial puncture.¹⁰

The Suboccipital Transcerebellar Approach

The suboccipital transcerebellar approach has been described as a means for obtaining a biopsy sample of lesions of the lower midbrain, pons, and rostral medulla.^40,41 Positioning often requires intubation and general anesthesia, although a method for using a local anesthetic has been described.⁴² The route is straightforward and projects through the ipsilateral middle cerebellar peduncle to the desired midbrain or brainstem lesion. One drawback to the suboccipital transcerebellar approach is the pain associated with muscle dissection before making the twist drill hole in the skull. A more decisive contraindication to this approach is the risk for posterior fossa subdural hematomas caused by blind puncture of the cerebellar cortex. Postoperative posterior fossa hematomas pose a higher risk for neurological sequelae than do similarly sized supratentorial subdural hematomas caused by a transfrontal approach.

The Ipsilateral Transfrontal Approach

The ipsilateral transfrontal approach is effective for biopsy of the midbrain, upper pons, and medulla.^43–47 For midline lesions, an anterior coronal entry point is chosen such that the trajectory enters through the midpoint of a gyrus approximately 15 to 25 mm from the midline. For lower pons or medulla biopsies, an ipsilateral transfrontal approach necessitates a trajectory traversing the lateral ventricle before entering the anterior thalamus, cerebral peduncle, and then the brainstem. As a result, lesions of the brainstem below the lower pons are typically approached by the contralateral, extraventricular approach (described in the next section). Finally, to reach the lateral midbrain, a trajectory lateral to the ventricle suffices.⁴⁶ This technique, however, has a few potential disadvantages. First, approaching the pons and medulla with this method often requires entering the frontal horn of the lateral ventricle, which violates two ependymal surfaces and places the patient at risk for intraventricular hemorrhage. Second, ventriculostomy may lead to intraoperative loss of cerebrospinal fluid, which may contribute to postoperative headache and shifting of intracranial contents, thereby possibly altering the location of the target. Third, this route is limited to midline regions of the pons and midbrain by the tentorial incisura.⁴⁸

The Contralateral, Extraventricular Transfrontal Approach

We recently reported on a contralateral, transfrontal, extraventricular approach (Fig. 114-6) as a safe and straightforward alternative for biopsy of infratentorial brainstem targets.³⁹ We have in particular found this approach advantageous for more lateral pontine lesions. After placement of the Leksell stereotactic frame and multiplanar MRI reconstruction, a point of entry through the midpoint of a gyrus approximately 3 cm lateral to the midline is chosen with special attention paid to avoiding large cortical vessels. An intraparenchymal trajectory is projected to avoid the lateral ventricle, the tentorium, major vessels, and the cerebral aqueduct. Biopsy specimens are sent for immediate intraoperative review by an attending neuropathologist. In cases requiring more than one sample to obtain adequate diagnostic material, we change the orientation of the needle and biopsy without performing additional passes. In our experience, a diagnosis was achieved in all cases, including metastatic adenocarcinoma, anaplastic astrocytoma (two cases), glioblastoma multiforme, and progressive multifocal leukoencephalopathy (two cases). Detailed intraoperative and postoperative neurological examinations showed no decline in function.

FIGURE 114-6 Image depicting the contralateral-transfrontal-extraventricular approach, our preferred technique for brainstem biopsies.

(From Amundson EW, McGirt MJ, Olivi A. A contralateral, transfrontal, extraventricular approach to stereotactic brainstem biopsy procedures. Technical note. J Neurosurg. 2005;102:565-570.)

The contralateral, extraventricular transfrontal approach is a good option for biopsy of the lateral pons and medial middle cerebellar peduncle. One goal in choosing a trajectory is minimization of the number of pial and ependymal surfaces crossed to decrease the risk for hemorrhage. With an entry point approximately 4 cm off midline in the coronal plane, the needle’s trajectory remains completely intraparenchymal, thereby avoiding the ventricle and limiting the number of crossed pial surfaces to one. This approach thus limits the potential for intraventricular hemorrhage, shift associated with loss of cerebrospinal fluid, and subdural hematomas. Advancing the biopsy needle slowly also theoretically helps minimize trauma by allowing tracts, nuclei, and small vessels to be displaced rather than sheared.⁴⁶ Intraoperative pathologic review by an attending neuropathologist may be helpful in minimizing the need to take multiple biopsy specimens. We believe that it also improves diagnostic yield and minimizes the potential morbidity and trauma associated with obtaining multiple samples, especially in a region as eloquent as the brainstem. This route also allows the use of a local anesthetic together with light intravenous sedation, which permits continuous patient participation in the neurological examination. Finally, although the length of the needle’s passage and therefore the number of structures displaced are greater when reaching the pons via the transfrontal route versus the transcerebellar route, we believe that the risk is negligible. In our previously published series of infratentorial brainstem biopsies, both ipsilateral and contralateral combined, there has been no significant differences in rates of complications, including symptomatic hemorrhage or hydrocephalus associated with aqueductal stenosis.⁴⁸ This result is consistent with findings in other series featuring this approach—no significant correlation between the length of the needle tract and morbidity has been found.⁴⁹

Conclusion

Stereotactic brain biopsy remains an indispensable method that may provide ample diagnostic information to guide oncologic treatment, including cytoreductive surgery, adjuvant chemotherapy, and radioablation. Frameless image-guided and frame-based stereotactic navigation are techniques that provide high yield and accurate brain biopsies. Studies have shown comparable technical and clinical diagnostic accuracy with both procedures. Additionally, complications associated with both techniques are infrequent. Indeed, a number of complications can be prevented by appropriate selection of lesions that warrant biopsy, careful planning of the trajectory, avoidance of bridging veins and critical structures seen with multimodality integration of functional imaging, and the judicious use of prophylactic antibiotics and anticonvulsants for those at risk. Factors that have been shown to limit the diagnostic accuracy of biopsy samples include tumor size and heterogeneity. Additionally, anatomic shifts that occur intraoperatively can cause shifting of intracranial contents and lead to inaccuracy in navigation. The evolution of iMRI represents an increasingly realizable solution for improving intralesional sampling accuracy and for surveillance of hemorrhagic complications.

With modern stereotactic techniques, there is growing evidence that lesions from the midbrain to the lower pons are becoming increasingly accessible for diagnostic biopsy. In adults, a diversity of benign and malignant lesions can occur in the midbrain and brainstem. Additionally, imaging alone is not sufficient to make an accurate diagnosis of lesions. Histologic diagnosis can therefore aid by narrowing treatment options so that patients are not subjected to unwarranted empirical therapies. In children, the majority of lesions in the brainstem are intrinsic, diffuse astrocytomas, for which resection is not a viable option. Given that indolent nonmalignant brainstem lesions occur in only a minority of children, the advisability of biopsy in children with midbrain or brainstem lesions is still a topic of debate.