OPERATIVE NUANCES
TRANSCALLOSAL TRANSCHOROIDAL APPROACH TUMORS OF THE THIRD VENTRICLE Hahnah J. Kasowski, M.D. Department of Neurosurgery, Yale University School of Medicine, New Haven, Connecticut
Brian V. Nahed, M.D. Department of Neurosurgery, Yale University School of Medicine, New Haven, Connecticut
TO
ALTHOUGH THE LITERATURE is rich with descriptions of the approach to the third ventricle, surgeons remain cautiously reserved. In this report, we demonstrate that the transcallosal approach can be easily performed provided that preoperative planning is adequate. Familiarity with the course of major cortical and deep draining veins grants the surgeon a wide exposure of the posterior third ventricle. We discuss the indications, surgical technique, and pitfalls to this approach while providing an accompanying video mirroring our discussion. KEY WORDS: Third ventricle tumor, Transcallosal, Transchoroidal Neurosurgery 57[ONS Suppl 3]:ONS-361–ONS-366, 2005
DOI: 10.1227/01.NEU.0000176652.59747.40
Joseph M. Piepmeier, M.D. Department of Neurosurgery, Yale University School of Medicine, New Haven, Connecticut Reprint requests: Joseph M. Piepmeier, M.D., Department of Neurosurgery, Yale University School of Medicine, TMP 4, 333 Cedar Street, New Haven, CT 06510. Email:
[email protected] Received, January 16, 2004. Accepted, May 19, 2005.
W
e present the surgical treatment of a tumor in the posterior third ventricle using a transcallosal approach. Third ventricular tumors are relatively rare, and few neurosurgeons approach these tumors regularly, although much has been written on this topic (1–9). The transcallosal approach provides excellent exposure to third ventricular lesions, but careful preoperative planning and knowledge of the anatomy is necessary to prevent complications. In particular, care should be taken to understand and predict the course of major cortical and deep draining veins with preoperative imaging. We discuss the indications, surgical technique, and pitfalls to this approach.
PREOPERATIVE PLANNING A clear understanding of venous anatomy is critical for the success of this operation. Adequate access to the corpus callosum and third ventricle requires that the surgeon establish an interhemispheric corridor that is 2 to 3 cm in diameter. Because the location of major cortical draining veins can limit this access and dictate the side on which the craniotomy is based, a magnetic resonance venogram, in addition to standard magnetic resonance imaging (MRI) sequences, proves to be quite useful. Although we usually plan to approach the ventricular system from the nondominant hemisphere, we may choose to enter on the opposite side if a wider corridor is available.
NEUROSURGERY
ILLUSTRATIVE CASE The patient was a 21-year-old woman who presented with a history of progressive headaches over a period of 1 month. MRI demonstrated hydrocephalus and a posterior third ventricular mass (Fig. 1). The patient initially underwent endoscopic third ventriculostomy and tumor biopsy, which was diagnostic of ependymoma. The third ventriculostomy was successful, and subsequent imaging demonstrated resolution of her hydrocephalus. Given the natural history of ependymoma, a complete excision of the tumor was planned.
PATIENT POSITIONING The patient was positioned supine, with the left shoulder elevated by a shoulder roll. The head was then rotated such that the coronal suture was oriented in the vertical plane and the right hemisphere was inferior. We prefer to have the right side down so that gravity aids in the retraction of the hemisphere. The head was fixed with a Mayfield frame, and a linear bicoronal incision was prepared and draped.
CRANIOTOMY (see video at web site) A bone flap that is centered two-thirds in front of and one-third behind the coronal suture is generally sufficient, but the bone flap may need to be moved slightly more anteriorly or posteriorly, depending on venous anatomy and location of the tumor (Fig. 2,
VOLUME 57 | OPERATIVE NEUROSURGERY 4 | OCTOBER 2005 | ONS-361
KASOWSKI
ET AL.
FIGURE 1. Preoperative MRI images (A, axial; B, sagittal; C, coronal).
inset). In this patient, because the tumor was located in the posterior third ventricle, our bone flap was made slightly more anterior to provide better direct visualization of the posterior third ventricle. The bone flap should allow several centimeters of visualization of the right hemisphere and extend just past the sagittal sinus on the left. To prevent damage to the sinus, we prefer to take the bone flap in two pieces: first, taking a bone flap over the right hemisphere that extends just to the right of midline; next, dissecting the sagittal sinus from the bone just medial to the previous bone flap. The bone over the sagittal sinus is then removed. It is critical that the dural opening provide exposure up to the edge of the superior sagittal sinus. We begin opening the dura laterally and then carefully extend the opening toward the midline where the dural flap is based. It is not uncommon to find that dominant frontal cortical veins enter the dura lateral to the superior sagittal sinus. When this occurs, the dura can be incised on either side of the vein such that a dural roof is formed over the vein, sparing the vein. Dural tenting stitches are then placed such that the dura is carefully retracted across the midline to the sagittal sinus. Care must be taken not to place too much tension on the dural flap, because this can partially or completely occlude the sagittal sinus.
INTERHEMISPHERIC DISSECTION Brain relaxation is critically important to minimize retraction-induced injury and should occur before interhemispheric dissection. A combination of hyperventilation, osmotic dieresis, and ventricular drainage can be used. We routinely
ONS-362 | VOLUME 57 | OPERATIVE NEUROSURGERY 4 | OCTOBER 2005
FIGURE 2. Transcallosal approach to the lateral and third ventricles. Inset, we positioned the head and extended the scalp incision (solid line), which allowed us to turn the bone flap (broken line) across the superior sagittal sinus. A, using an incision in the anterior part of the corpus callosum, we exposed the body and frontal horn of the right lateral ventricle. B, there are several possible incisions used to reach lesions in the third ventricle: 1, the foramen of Monro may be enlarged by incising the ipsilateral column of the fornix at the anterosuperior margin of the foramen of Monro; 2, the transforniceal approach is completed using an incision along the body of the fornix in the midline; and 3, the transchoroidal approach is completed by opening the choroidal fissure by incising along the tenia fornicis. C, the transchoroidal approach requires an incision along the tenia fornicis rather than the tenia thalami to avoid the many veins and arteries that pass through the tenia thalami. Notice the internal cerebral veins that run along the roof of the third ventricle (from, Rhoton AL Jr: The supratentorial cranial space: Microsurgical anatomy and surgical approaches. Neurosurgery 51[Suppl 1]:S207–S272, 2002 [9]).
give 0.5 g/kg of mannitol at skin incision and maintain the end-tidal CO2 in the range of 25 to 30 mm Hg. If the brain is not adequately relaxed at this point, a ventricular drain can be placed ipsilateral to the craniotomy. However, large tumors may necessitate placement of a contralateral ventriculostomy. A self-retaining retractor system is brought onto the field and used to gently retract the right hemisphere. The operating microscope is then brought in and used for the remainder of the operation. Both blunt and sharp dissecting instruments are used to develop the plane along the falx and then between the cingulate gyri. The pericallosal arteries must then be identified. Once the surgeon is sure of their course, the corpus callosum can be identified by its relative hypovascularity and pale white color.
www.neurosurgery-online.com
TRANSCALLOSAL TRANSCHOROIDAL APPROACH
CALLOSOTOMY The callosal incision is planned between the two anterior cerebral arteries with the goal of entering the right lateral ventricle. We suggest incising the callosum just beneath the right anterior cerebral artery, aiming slightly toward the right hemisphere. The surgeon should avoid dissecting toward the left side because of the effect of the right-side-down head positioning and the resultant shift of the right hemisphere. Both of these techniques will help to keep the callosum from opening into the left ventricle or in line with the attachment of the septum pellucidum. Although this is not necessarily a problem, it can cause unnecessary confusion. The length and placement of the callosotomy is determined by the location and size of the tumor. Our callosal incision in this case measures approximately 2 cm in length and is made with bipolar coagulation and suctioning with a 6-French suction tip. The corpus callosum can be variable in thickness, depending on the underlying tumor mass and/or preexisting hydrocephalus, and can be stretched to just a few millimeters. In this patient, the corpus callosum was relatively thick, measuring approximately 1 cm. The exact placement of the callosotomy may vary, depending on the surgical window afforded by the cortical draining veins. The preoperative MRI scan may be helpful in localizing the preferred distance of the callosotomy from the genu of the corpus callosum. The ependymal lining is recognized by its deep gray coloring just before the ventricle is entered. Once the ventricle is opened, a self-retaining retractor blade can be advanced just into the ventricle.
TRANSCHOROIDAL DISSECTION Once the ventricle is opened, it is important to first establish normal landmarks. The structure that is most obvious and easiest to identify is the choroid plexus, which can be followed anteriorly to the foramen of Monro. Further identification of the septal and thalamostriate veins can be used to determine which ventricle has been opened (Fig. 2A). In the right lateral ventricle, the thalamostriate vein is to the right of the choroid plexus. If callosal sectioning has led to the contralateral ventricle, the septum pellucidum can be opened to gain access to the contralateral side. Should this occur, however, the surgeon must take care to preserve the fornices at the base of the septum. Below the fornix, the roof of the third ventricle is composed of two layers of the tela choroidea and the velum interpositum, a vascular space that is located between these layers. The internal cerebral veins and the medial posterior choroidal arteries run within the velum interpositum. Our approach was through the tela choroidea and velum interpositum into the third ventricle. Dissection through the tela choroidea should be performed medial to the choroidal fissure in the tenia fornicis rather than in the tenia thalami, because the veins that drain the internal capsule and medial hemisphere pass through the tenia thalami and should not be sectioned (Fig. 2, B and C). We prefer to begin the dissection by gently lifting the choroid plexus, coagulating it, and
NEUROSURGERY
TO
THIRD VENTRICLE
then removing it along its insertion at the choroid fissure, allowing for visualization of the choroidal fissure. We begin our dissection along the medial edge of the choroidal fissure while gently elevating the adjacent fornix to avoid inadvertent damage. Once the tenia fornix is divided, the velum interpositum is entered and the internal cerebral veins and the medial posterior choroidal arteries are visualized. These vessels should be retracted laterally until the lower layer of tela is visualized, which may then also be divided (Fig. 3). The choroid plexus of the third ventricle, the massa intermedia, and the floor of the third ventricle are now seen. A small portion of the tumor can also be seen below the massa intermedia. At this point, we encountered two anatomic structures that limited our view and approach to the tumor. The first was a large massa intermedia. This patient was known from the preoperative imaging to have a large massa intermedia, and as expected, the massa intermedia significantly obstructed our view of the posterior third ventricle and tumor. As mentioned, this patient had had a previous endoscopic third ventriculostomy, which alleviated her hydrocephalus. We suspect that if the patient had still had significant hydrocephalus, the massa intermedia would have been significantly stretched and easier to work around. However, in this case, we were forced to divide it so as to better visualize and resect the tumor, a maneuver that we have not previously had to perform. Postoperatively, the patient experienced no appreciable neurological deficit from this maneuver. After the massa intermedia was sectioned, our entry into the third ventricle was also limited by two thalamostriate veins. We initially attempted to dissect the anterior vein from the lateral wall of the ventricle. However, when this proved difficult, we thought that it was preferable to sacrifice the vein rather than risk its rupture. Rarely do we sacrifice draining veins; however, when there is duplication of the venous drainage, as in this patient, we have not had complications when a single vein is sacrificed.
LESION EXCISION Tumor resection is facilitated by maintaining the dissection plane between the ependyma and the lesion. Because it is not uncommon for intraventricular tumors to grow slowly and become quite large, we often must first decompress the tumor and then identify the space between the tumor edge and ependyma. Determining the type of tumor is important in predicting the blood supply. Papillomas and meningiomas, for example, receive their blood supply from the choroidal vessels. With these tumors, early identification and transection of these vessels will significantly reduce bleeding. In contrast, tumors that arise from the ependymal surface and septum pellucidum, such as gliomas and neurocytomas, receive their blood supply from the small vessels of the ventricular walls. In this patient, the blood supply was encountered after primary debulking. These vessels tend to be smaller and cause less blood loss but still require meticulous dissection and good visualization.
VOLUME 57 | OPERATIVE NEUROSURGERY 4 | OCTOBER 2005 | ONS-363
KASOWSKI
ET AL.
FIGURE 3. Transchoroidal approach to the third ventricle directed along the forniceal side of the choroidal fissure. A, superior view of the frontal horn and body of the lateral ventricle. The body of the fornix forms the upper part of the roof of the third ventricle. The left thalamostriate vein passes through the posterior margin of the foramen of Monro, and the right thalamostriate vein passes through the choroidal fissure a few millimeters behind the foramen. Anterior septal and anterior caudate veins cross the wall of the frontal horn. Posterior septal and posterior caudate veins cross the wall of the body of the lateral ventricle. The thalamus sits in the floor of the body. The choroidal fissure, located between the thalamus and fornix, is opened by dividing the tenia fornix, which attaches the choroid plexus to the lateral edge of the fornix, leaving the attachment of the choroid plexus to the thalamus undisturbed. B, enlarged view. The columns of the fornix form the anterior and superior walls of the foramen of Monro. The massa intermedia is seen through the foramen. Anterior and posterior septal veins cross the septum pellucidum and fornix. C, the tenia fornix, which attaches the choroid plexus to the fornix, has been divided and the body of the fornix retracted medially to expose the internal cerebral vein and medial posterior choroidal arteries. The lower layer of tela, which attaches to the striae medullaris thalami and forms the floor of the velum interpositum, is intact. D, the separation of the fornix and choroid plexus has been extended posteriorly to the junction of the atrium and body of the ventricle. The lower layer of tela remains intact. E, the lower layer of tela has been opened to expose the massa intermedia, the posterior commissure, and the floor of the third ventricle. The ependymal covering of the anterior septal vein has been opened so that a short segment of the vein can be mobilized. The possibility of damaging the thalamostriate vein is reduced by allowing the choroid plexus to remain attached to the thalamus and the upper surface of the vein (from, Rhoton AL Jr: The supratentorial cranial space: Microsurgical anatomy and surgical approaches. Neurosurgery 51[Suppl 1]:S207–S272, 2002 [9]).
The tumor was small, and initial tumor removal was easily accomplished with suctioning. As the tumor was aspirated, the cerebral aqueduct was visualized. The previous third ventriculostomy was also appreciated more anteriorly. As we
ONS-364 | VOLUME 57 | OPERATIVE NEUROSURGERY 4 | OCTOBER 2005
approached the posterior part of the tumor beneath the posterior commissure, many small tumor vessels were encountered. Visualization of the most posterior portion of the tumor was limited by the overlying posterior commissure. The resection continued until we felt that the risk of further resection outweighed the risk of some small residual tumor. Although we strive for a complete tumor excision, extending the resection into the underlying midbrain can be devastating and should not be attempted if the tumor plane is unclear. After tumor resection, it is important to ensure complete hemostasis. To prevent ventricular obstruction and hydrocephalus, blood that may have accumulated should be gently irrigated out of the ventricular system. The ventricles are then filled with saline to displace any air that may have become trapped. A ventricular catheter is left for 24 to 48 hours to measure ventricular pressure and to demonstrate that the ventricular system is patent. A computed tomographic or MRI scan is obtained on the first postoperative day to check for obstruction and to evaluate the resection. Postoperatively, the patient was neurologically intact and did not experience any complications related to the procedure.
LIMITATIONS OF THIS APPROACH This case demonstrates that the transcallosal approach is easily performed, provided that the surgeon is familiar with the anatomy and achieves wide exposure of the posterior third ventricle. However, postoperative imaging revealed small residual tumor in the posterior third ventricle (Fig. 4). In this patient, our exposure was limited in the pineal recess of the third
www.neurosurgery-online.com
TRANSCALLOSAL TRANSCHOROIDAL APPROACH
ventricle, which might have been avoided if the surgery had been performed before resolution of the patient’s hydrocephalus, because this approach is greatly facilitated by a dilated ventricular system. Alternatively, the occipital transtentorial approach might have provided better access to the posterior third ventricle; however, a discussion of this approach is beyond the scope of this article and is useful for leFIGURE 4. Postoperative axial MRI sions in the pineal recess of scan. the third ventricle. Although a gross total resection would have been preferred, we stress that the risks of inadvertently entering the underlying midbrain far outweigh the benefit of a gross total resection if the margins are difficult to identify. In the future, we may also use the endoscope in the open transcallosal approach. Several authors have recently described the use of the endoscope in open surgery to provide better visualization of deep structures not easily seen with the standard operating microscope.
REFERENCES 1. Apuzzo MLJ (ed): Surgery of the Third Ventricle. Baltimore, Williams & Wilkins, 1998, ed 2. 2. Apuzzo MLJ, Chikovani OK, Gott PS, Teng EL, Zee CS, Giannotta SL, Weiss MH: Transcallosal, interfornicial approaches for lesions affecting the third ventricle: Surgical considerations and consequences. Neurosurgery 10:547–554, 1982. 3. Apuzzo MLJ, Litofsky NS: Surgery in and around the anterior third ventricle, in Apuzzo MLJ (ed): Brain Surgery: Complication Avoidance and Management. New York, Churchill Livingstone, 1993, pp 541–579. 4. Kasowski HJ, Piepmeier JM: Transcallosal approach for tumors of the lateral and third ventricles. Neurosurg Focus 10:Article 3, 2001. 5. Lavyne MH, Patterson RH: Subchoroidal trans-velum interpositum approach, in Apuzzo MLJ (ed): Surgery of the Third Ventricle. Baltimore, Williams & Wilkins, 1998, ed 2, pp 453–469. 6. Piepmeier JM: Tumors and approaches to the lateral ventricles: Introduction and overview. J Neurooncol 30:267–274, 1996. 7. Piepmeier JM, Spencer DD, Sass KJ, George TM: Lateral ventricular masses, in Apuzzo MLJ (ed): Brain Surgery: Complication Avoidance and Management. New York, Churchill Livingstone, 1993, pp 581–600. 8. Rhoton AL Jr: The supratentorial cranial space: Microsurgical anatomy and surgical approaches. Neurosurgery 51[Suppl 1]:S207–S272, 2002. 9. Wen HT, Rhoton AL Jr, de Oliveira EP: Transchoroidal approach to the third ventricle: An anatomic study of the choroidal fissure and its clinical application. Neurosurgery 42:1205–1217, 1998.
COMMENTS
K
asowski et al. review the transcallosal transchoroidal approach to tumors of the third ventricle. Although the authors present little new information, they have created a good resource for residents or practicing neurosurgeons to review this elegant approach. A concise, pertinent video is provided, as are anatomic drawings.
NEUROSURGERY
TO
THIRD VENTRICLE
We agree that positioning the head so that longitudinal access is parallel to the floor is preferable to positioning it so that longitudinal access is perpendicular to the floor. As the authors note, an advantage of this horizontal position is that gravity is used to allow the ipsilateral hemisphere to fall away, thus naturally providing a larger interhemispheric corridor. The second advantage to this position is that the surgeon’s hands can work side by side. This position is much more comfortable and convenient than the superoinferior hand position required when operating with the head in the vertical position. We have found that when a wider corridor is needed to access the third ventricle via the transcallosal approach, the septal vein can be sacrificed and cut. The opening through the velum interpositum can then be extended to incorporate the foramen magnum. The result is a relatively large corridor into the third ventricle. In our experience, sacrifice of the septal vein has minimal or no negative ramifications for the patient. Jeffrey Klopfenstein Robert F. Spetzler Phoenix, Arizona
K
asowski et al. present one approach for the removal of tumors located in the third ventricle. As an example, the authors demonstrate a case of a 21-year-old woman with a third ventricular ependymoma that became symptomatic with obstructive hydrocephalus. After a third ventriculocisternostomy, the authors performed a transcallosal approach in a second surgery and removed the lesion incompletely. This is an excellent case for the presentation and indication of this approach. The article gains from its detailed description of surgical steps, including positioning, attention to brain relaxation, and dealing with bridging veins and the sagittal sinus. Pitfalls and important planning details, such as respecting the cortical draining veins in the planning of the side of the craniotomy, are underlined. In our opinion, surgery in this area benefits greatly from neuronavigation, especially in the planning of the craniotomy with respect to cortical draining veins. We completely agree with the authors regarding the planning of the craniotomy, if possible, in the nondominant hemisphere. Not only cortical draining veins but also the location and lateral growth of the lesion can change the planning of the craniotomy in this approach. The authors show their experience with this approach in their description of how to avoid opening of a wrong-sided ventricle or a cavum septi pellucidi or cavum vergae. Regarding the transchoroidal dissection, they point out the importance of anatomic identification and respecting the fornix and deep draining veins. Regarding the preoperative ventriculocisternostomy, we agree with the authors that the surgical procedure might be easier and more successful if they had not performed it preoperatively. Even the division of the adhesio interthalamica (massa intermedia) might have been avoided. If really necessary, we would recommend the introduction of a temporary ventricular drainage that can be closed before surgical tumor removal. This will render microsurgery in the third ventricle much easier, and obstruction of the massa intermedia and thalamostriate veins would be diminished. We have to point out that third ventriculocisternostomy is a more dangerous procedure than a ventricular drainage, especially because the procedure might be unnecessary after a successful
VOLUME 57 | OPERATIVE NEUROSURGERY 4 | OCTOBER 2005 | ONS-365
KASOWSKI
ET AL.
surgery. This article offers a lot of insight into the application and indications of this anatomically interesting approach. Wolf Lu¨demann Madjid Samii Hannover, Germany
T
his is a very nice description of the transchoroidal approach to the third ventricle, with excellent illustrations. The advantage of this technique is to spare the veins. The authors are very cautious in performing six burr holes for the craniotomy in two flaps. One or two burr holes on the midline over the sagittal sinus might be enough in young patients. The transchoroidal approach is particularly suitable for tumors in the midportion of the third ventricle. The tumor in the present patient is in the posterior third ventricle, and I would have preferred an occipital transtentorial approach. Nicholas de Tribolet Geneva, Switzerland
T
he authors present their treatment of tumors involving the posterior third ventricle using a transcallosal approach. They emphasize that the transcallosal approach provides excellent exposure when coupled with diligent and careful preoperative planning. Knowledge of the salient venous anatomy, variants of the choroidal fissure anatomy, and three-dimensional relationships of the fissure to the pineal region and quadrigeminal plate cistern is requisite to prevent complications. Preoperative understanding of midline draining veins is essential in tailoring the approach to minimize venous sacrifice and compromise. The use of
ONS-366 | VOLUME 57 | OPERATIVE NEUROSURGERY 4 | OCTOBER 2005
gravity to aid in the retraction of the hemisphere is also an important factor. We use the three-quarter lateral position to maximize this. In addition, we prefer soft retraction with rolled patties placed at the anterior and posterior limits of the interhemispheric corridor. Although the authors suggest an interhemispheric corridor that is 2 to 3 cm in diameter, we have had good results with smaller approaches, given that the anteroposterior dimension of the corridor is more important than the lateral dimension (which only promotes the potential for excessive retraction). We also prefer to take the flap as a single piece and actually use the drill to place burr holes over the midline. We have not had a problem with sinus injury using this approach in children, given that the ability to separate the dura from the overlying bone is easier in younger children. In addition to dural tenting stitches, as noted, we also place dural sutures in the falx for additional retraction. The use of ventricular drainage to allow for relaxation of the brain cannot be overemphasized. A 2-cm callosal incision will usually allow for any approach in this region. In addition, as the authors note, the surgeon must take care to preserve the fornices at the base of the septum. It is also important that dissection through the tela choroidea is performed medial to the choroidal fissure in the tenia fornicis rather than in the tenia thalami. Finally, although complete excision is an appropriate goal, extending the resection into the underlying midbrain can be devastating and should not be attempted if the tumor plane is unclear. The only other consideration is the use of concurrent endoscopy, as initially described by Perneczky, which is now common in a number of institutions. Michael L. Levy San Diego, California
www.neurosurgery-online.com
