Introduction

The resection of tumors of the aqueduct and third ventricle, especially the posterior part, is not without difficulties and potential complications, despite the applied approaches and techniques. After the first reports published by Gaab in 1998 [8], neuroendoscopy has increasingly focused the interest of neurosurgeons on the pure endoscopic resection of soft tumors smaller than 2 cm [2, 16, 24, 25]. Although this technique requires a strict patients selection and is only indicated in a limited number of cases, it can provide a number of advantages and only a few drawbacks compared to microneurosurgery. Several techniques and tools have been proposed for endoscopic removal of solid tumors: coagulation, dissection by forceps and laser, and aspiration by cannulas. We report our surgical method, providing both a technical description and some illustrative cases in which we have used the direct endoscopic aspiration procedure.

Procedure

A flexible scope (Karl Storz GmbH and Co., Tuttlingen, Germany) is used with a 3.7 mm diameter and a 1.5 mm working channel. A 14-French peel-away sheath is inserted, as usual, into the right frontal horn through a right precoronal burr hole positioned 2 cm from the midline. The flexible scope is then introduced with a free-hand technique. After quickly inspecting the lateral and third ventricles, the tumor surface is carefully examined, determining whether the lesion is pedunculated and its precise origin. Such close inspection of the external features can give an idea of its consistency. A vegetating aspect, pinkish color, and abundant surface capillaries are all signs that the tumor is amenable to aspiration (Fig. 1). The tumor is then removed with the working channel of the scope placed on the tumor surface and used as a surgical aspirator (Fig. 2). Despite the thin diameter of 1.5 mm, serial aspirations can be performed through the vacuum created by a syringe. This frees the posterior part of the third ventricle, restoring aqueduct patency. Bleeding is simply controlled by continuous irrigation with Ringer’s solution. In one case, after tumor aspiration, a blood clot was seen obstructing the aditus of the aqueduct. Accordingly, after aspirating it with the same technique, we also explored the fourth ventricle through the restored aqueduct. A third ventriculostomy is performed in all cases to improve the chances of preventing hydrocephalus.

Fig. 1
figure 1

Endoscopic images of Case 1. a Inspection of the third ventricle; the tumor is located in its posterior part. b The external features of the tumor are carefully examined; vegetating aspect, pinkish color, and abundant capillaries on the surface are all signs of good aspirability. c, d The scope is slid along the surfaces of the tumor to determine whether the lesion is pedunculated and its precise origin. The aditus aqueducti is identified under the tumor

Full size image
Fig. 2
figure 2

Illustration showing the trajectory of the flexible scope to the posterior aspect of the third ventricle. Some small, soft tumors can be directly aspirated through the working channel just by embedding the tip of the scope into the lesion

Full size image

Case series

Case 1

A 22-year-old male patient presented with a 40-day history of worsening headache, associated with nausea in the previous week. Cerebral MRI showed a 2.1 × 1.6 × 1.4 cm (2.4 cm3) mass in the pineal region, extending to the aditus aqueducti mesencephali (Fig. 3). Triventricular hydrocephalus was also present. Neurological examination was normal at admission. The tumor features allowed complete removal (Fig. 1). The postoperative course was uneventful, with resolution of symptoms. An immediate postoperative cerebral MRI confirmed the gross-total resection of the lesion, despite the presence of some blood clots (Fig. 3). Spinal MRI did not show any pathological enhancement after gadolinium injection. The examination was consistent with a diagnosis of anaplastic ependymoma with papillary features (grade III, WHO). The patient underwent conformational radiotherapy (60Gy in 30 fractions), which was completed three months after surgery. No chemotherapy was administered.

Fig. 3
figure 3

Magnetic resonance imaging of Case 1, showing a 2.1 × 1.6 × 1.4 cm hypointense tumor in the pineal region, extending to the aditus aqueducti mesencephali (a). Triventricular hydrocephalus was also present. b Postoperative MRI confirming the gross-total resection of the lesion. c 1-year follow up MRI reveals the absence of local recurrence

Full size image

Four months after surgery, the patient complained of headache and nausea. Cerebral MRI showed the presence of triventricular hydrocephalus. Cine-flow sequences revealed weak flow through both the third ventriculostomy and the aqueduct. The patient underwent a second neuroendoscopic procedure to repeat ETV. The former stoma site on the floor of the third ventricle, which had probably been incorrectly placed too laterally, was obstructed by a thin membrane, which was opened. The revised ventriculostomy demonstrated clear patency, with effective flow. Subsequently, the posterior aspect of the third ventricle was also accurately explored. No remnant of tumor tissue was observed. The bare choroidal tela and posterior commissure appeared almost anatomically normal, except for the slight ocher color of the posterior commissure caused by the described microbleedings (Fig. 4). The postoperative course was uneventful, with complete resolution of symptoms.

Fig. 4
figure 4

Endoscopic appearance at 1-year follow up; a: aqueduct; pc: posterior commissure

Full size image

At 1-year follow up, the patient is asymptomatic. Cerebral MRI performed 1 year after initial surgery revealed no pathological enhancements (Fig. 3).

Case 2

A 48-year-old male patient presented with a one-month history of intense headache with bilateral tinnitus, and gait instability. Cerebral MRI revealed a partially solid, partially cystic lesion in the posterior aspect of the III ventricle, obstructing the aqueduct and causing triventricular hydrocephalus (17 mm maximum diameter). He was admitted to our department with a normal neurological examination. Due to the wide implant, the lesion was only partially removed, enough to free the aqueduct. The postoperative course was uneventful. The patient completely recovered, and MRI showed the resolution of hydrocephalus.

The pathological examination was consistent with a diagnosis of carcinoma with papillary architecture. The patient underwent radiotherapy (RT) and chemotherapy (CT). He is still alive after 9 years.

Case 3

A 57-year-old man presented with a 2-month history of asthenia, gait difficulties, and disequilibrium. Cerebral MRI showed a lesion of the tectum mesencephali, extending into the aqueduct and reaching the cranial portion of the IV ventricle. A triventricular hydrocephalus was also present. A near-total aspiration was performed. The postoperative course was uneventful. Pathologic examination was consistent with a diagnosis of rosette-forming glioneuronal tumor of the IV ventricle, with atypical characteristics. Despite adjuvant therapy with hyperfractionated radiotherapy and radiosurgery, the tumor rapidly regrew. The patient underwent RT and radiosurgery for recurrences, and died 3 years after diagnosis.

Discussion

The central location of the third ventricle makes it particularly challenging to reach in open surgery. Microneurosurgical approaches almost always require the dissection and retraction of eloquent and delicate brain structures, and can potentially lead to neurological deficits. This is particularly true of lesions that are deep-seated in the posterior third ventricle or pineal region, accounting for 0.6 % to 0.9 % of all brain tumors [1, 22]. For this reason, in an attempt to reduce mechanical stress on the brain, several neurosurgeons have proposed minimally invasive neuroendoscopic techniques to approach such lesions. Neuroendoscopy has already proved to be effective and safe in the treatment of purely cystic lesions as colloid cysts [10, 14], and in the resection of purely cystic choroid plexus tumors of the third ventricle [12, 18, 19], cysticercosis cysts of the posterior third ventricle [20], and ependymal cyst of the third ventricle [26].

The application of neuroendoscopy in the resection of solid tumors of the third ventricle is still a matter of technical debate and variants. Some authors have advocated the use of multiple endoscopes or entry sites for additional instrumentation [7, 11]. However, this is technically troublesome and not so minimally invasive. Others have used combined endoscopic and surgical approaches [3, 4], such as introducing wider operative sheaths to remove central neurocytomas [5] or adapting a thoracic port to resect tumors in the lateral or third ventricles [9]. Some authors have reported the use of a dual approach to remove posterior third ventricle tumors (forehead burr hole) and to manage hydrocephalus with ETV (standard frontal burr hole) [6, 23]. New tools as the ultrasonic aspiration tip have also been developed in order to expand the spectrum of endoscopic options when facing a solid intraventricular lesion [21].

A further step towards and effective, really minimally invasive approach to the third ventricle was made by Gaab in 1998, who partially or totally resected 4 tumors using just a burr hole and a rigid endoscope, with no complications. Bipolar diathermy and grasping and biopsy forceps were used to this end [8].

All the authors agree that the factors potentially limiting this approach are size of the tumor, its consistency, and vascularization. Small, soft tumors no greater than 2 cm are the ones most amenable to neuroendoscopic removal. Unfortunately, tumor consistency is not always clearly predictable, although in our opinion it represents the main feature. Souweidane et al. showed CT scan to be beneficial in excluding calcified and hyperdense tumors, which are usually more cellular than hypodense lesions [25]. In our opinion, however, direct inspection is the definitive way to assess tumor aspirability. For this reason, we always plan an initial endoscopic biopsy, and intraoperatively decide whether tumor aspiration would be feasible.

In all the published cases, the authors report coagulating the tumor surface, followed by the use of several instruments, as the biopsy rongeur, the forceps, and in some cases the diode laser (Table 1). In our view, however, some tumors lend themselves to the use of a flexible scope alone, which can be used as an aspirator by simply embedding its tip into the lesion and carefully applying a pulsating suction action with an external syringe. Although visualization is prevented once the endoscope is placed on the tumor, the tumor can be removed piecemeal due to the small dimensions of the working channel and because the degree of aspiration can be controlled through the syringe. As the aspirated fragments are very small, inadvertent aspiration of surrounding structures is avoided by carefully and accurately positioning the tip of the endoscope on just the tumoral portions. Nonetheless, such a manoeuvre requires an adequate neuroendoscopic training, and should be performed by expert endoscopists only. We have gained extensive experience in the endoscopic aspiration of intraventricular hemorrhagic clots in hematocephalus totalis [13]. Direct aspiration with the working channel was also used to empty all our cases of colloid cysts and even brain abscesses [14, 15]. We applied the same technique for the aspiration of particularly soft tumors with high water content. Aspiration can be a simple, useful manoeuver for removing small, soft tumors. With this technique we were able to resect three tumors (1 gross total removal, 2 partial removals) of the third ventricle/aqueduct. In another 12 cases, either size or stiff fibrotic consistency prevented us from performing anything more than a biopsy. Luther and Souweidane also reported the complete resection of 6 tumors of the third ventricle by aspiration. They did, however, stress the need to perform aspiration through a 6-French endotracheal suction catheter which had been previously angled-cut by the operators themselves, in order to more successfully break the tumor capsule [16, 25]. We agree on the importance of aspiration in such cases. In our experience, a flexible scope with a smaller diameter allows minimal invasiveness but still guarantees effective, direct aspiration.

Table 1 Endoscopic removal of third ventricular tumors. Literature review
Full size table

Tumors connected to the ventricular wall by just a pedicle are the most amenable to this kind of resection, which is particularly indicated in the subgroup of pedunculated tumors plunging into the aqueduct. In these cases the best results can be achieved as described by Rosenstengel et al., and confirmed by our own cases [24]. Some authors consider hypervascularization to be a contraindication for endoscopic resection of ventricular lesions. However, superficially vascularized tumors are usually softer and therefore good candidates for neuroendoscopic aspiration. Bleeding during aspiration is mainly due to laceration of capillaries or microarteries. In our experience, although it is annoying and delays surgery time, bleeding can usually be controlled by continuous irrigation with Ringer’s solution. In general, we used a flexible scope in more than 600 procedures, and never failed to manage intraoperative bleeding. Nonetheless, thorough endoscopic inspection is advisable even in the fourth ventricle, as it permits aspiration of potential clots obstructing CSF flow.

All the authors consider it mandatory to perform a third ventriculostomy during the same procedure. Third ventriculostomy is almost always necessary to cure or prevent hydrocephalus in such cases.

The opportunity to take a second look at the surgery site in one of our patients, in order to redo a third ventriculostomy four months after tumor aspiration, allowed us to observe the complete absence of the tumor and restoration of the normal anatomy of the posterior third ventricle. Although impressive and esthetically rewarding for the surgeon, this evidence could be due at least in part to the radiotherapy performed two months beforehand.

Conclusion

When a tumor of limited size grows in the posterior third ventricle or plunges into the aqueduct, it can be reasonably reached through a neuroendoscopic procedure, during which not only biopsy, but also partial or gross total soft tumor aspiration are achievable. Tiny soft tumors have been endoscopically aspirated in a few cases reported in the literature. To our knowledge, our case provides for the first time late, direct, visual evidence of the effectiveness of endoscopic gross total aspiration of a soft tumor in the third ventricle.