Neuroendoscopic Techniques in the Treatment of Hydrocephalus

Abstract

Neuroendoscopic techniques have been used to treat hydrocephalus for more than 100 years. With the personalized design of surgical approaches, increased knowledge of ventricular anatomy, and improved neuroendoscopic equipment, the last 20 years have witnessed tremendous advances in the development of neuroendoscopic technology, especially in the treatment of hydrocephalus. Except for obstructive hydrocephalus, the application of neuroendoscopic technology in the field of hydrocephalus is also expanding and has received good results, mainly in the fields of pediatric hydrocephalus and communicating hydrocephalus. Additionally, many scholars have achieved satisfactory results in the application of ventriculoscopy to complex hydrocephalus. Among neuroendoscopic techniques, the third ventricular floor fistula and cyst wall fistula methods are commonly used in the treatment of hydrocephalus and are highlighted in this chapter. Undoubtedly, neuroendoscopic technology has become one of the key treatment methods for hydrocephalus, with its high success rate, few complications, and accurate long-term efficacy.

1. Introduction

The modern endoscope was first introduced by German urologist Maximilian Carl-Friedrich Nitze (1848–1906); in 1879, he introduced the Nitze-Leiter cystoscope, which opened the door for neurosurgeons to adapt endoscopy to their procedures [1]. However, neuroendoscopic techniques for treating hydrocephalus date back to the early twentieth century when, in 1904, Lespinasse, a urologist in Chicago, used cystoscopic cauterization of the choroid plexus to treat two infants with hydrocephalus; one infant died postoperatively, and the other died 5 years after surgery, and the cases were reported publicly at a medical conference in 1910 [2, 3]. In 1922, the neurosurgeon Dandy used endoscopic techniques to treat two patients with hydrocephalus, during which he successfully cauterized and removed the choroid plexus; in an article published in the same year, he described in detail the anatomy of the lateral ventricle, interventricular foramen, transparent septum, and choroid plexus. On this basis, Dandy first proposed the concept of ventricular inspection using an endoscope and the term “ventriculostomy” [4].

The first doctor to complete an endoscopic third ventriculostomy (ETV) was William Jason Mixter [4, 5]; in 1923, he published an article on his successful use of cystoscopy to enter the third ventricle via the lateral ventricle foramen of Monro to perform ETV. Subsequently, ETV has been widely used to treat patients with obstructive hydrocephalus. In the same year, Dr. Temple Fay and Dr. Francis Grant developed a method to capture clear black-and-white images of the ventricles using a cystoscope. Their case report of a 10-month-old Italian boy illustrated an attempt to fenestrate the corpus callosum to treat hydrocephalus; however, they were unable to divide the corpus callosum due to a malfunction in the cystoscope they were using. Although the procedure did not go as planned, they concluded that it was safe to visualize the ventricle using an endoscope without causing ventricular hemorrhage or other complications as follows [5, 6, 7]:

1. Ventricular imaging and neuroendoscopy are possible in enlarged ventricles.

2. Neuroendoscopy is an intervention with diagnostic significance and can directly observe changes in the ventricles and locate the causes of hydrocephalic lesions.

3. If performed correctly, the patient’s postoperative reaction will be mild.

Subsequently, this technique was widely used in the treatment of hydrocephalus. In 1934, Tracy J. Putnam reported on the technique of neuroendoscopic resection of the choroid plexus [8]. Additionally, he invented the electrocoagulation endoscope and designed a 7-mm ventriculoscope with a rod of optical glass, resulting in a wider optical aperture.

For some time, ETV, or endoscopic coagulation of the choroid plexus, was the mainstay of surgical treatment for hydrocephalus despite the associated high complication rate. In 1935, John Scarff (1898–1979) adapted an enhanced version of Putnam’s ventriculoscope with an irrigation system to maintain intraventricular pressure, thereby preventing ventricular collapse [2, 9]. In 1947, H. F. McNickle made a significant contribution to the principle of ventriculostomy itself rather than the technique; he was unconvinced that the procedure should be limited to non-communicating hydrocephalus. Additionally, he argued that choroid plexectomy was not the optimal treatment for hydrocephalus since decreasing the production of cerebrospinal fluid (CSF) would not clear the obstruction itself [10].

At this time, neuroendoscopic technology was imperfect, the technology was limited, and patient disability and mortality rates were high. In view of these, the equipment was not popular in neurosurgery due to surgical discomfort, the risk of eye infection from the endoscope, burns from the heat of the lamp in the endoscope tip, and lower optical quality compared with that of a stereoscopic microscope.

From the 1950s to the 1970s, with the advent of magnetron pressure-regulating drainage tubes, the ventriculoperitoneal (VP) shunt was widely recognized by the medical community. After Putnam’s initial success with a glass rod instead of lenses, Hopkins’ patents, which used several glass rods to fill the former air spaces between the lenses, enabled the development of today’s superb-quality rod-lens endoscopes. In 1959, Hopkins, a professor of physics at the University of Reading in the United Kingdom, made a modern optical fiber endoscope, and with the assistance of Karl Storz, they applied the column-lens system to the endoscope and combined it with optical fiber technology to greatly improve the illumination and resolution of the image [11]. In the subsequent 30 years, neuroendoscopic equipment continued to improve, and coupled with the promotion of the concept of minimally invasive treatment, neuroendoscopic technology was widely used in the field of neurosurgery, and many new neuroendoscopic techniques were developed.

A breakthrough occurred in Paris with the development of the cold-light generator by Vulmiere and colleagues, which Guiot subsequently used as part of a “universal endoscope” [2, 12]. The first modern description of intraventricular endoscopic biopsy using a ventriculofiberscope was provided by Fukushima et al. in 1973 [13]. Around the same time, in England, Hugh Griffith recommended the endoscopic procedure as a first-line treatment for childhood hydrocephalus. He used Hopkins’ rigid endoscope to perform ETV as well as choroid plexus coagulation (CPC) to treat hydrocephalus [14].

Advances in illumination and magnification and the refinement of endoscopic tools coupled with improved anatomical knowledge furthered the development of neuroendoscopy [15]. In 1980, Hoffman et al. reported a series of ETVs using stereotactic guidance, noting that the percutaneous stereotactic-guided approach yielded better results than with open craniotomy [16]. In 1991, Heilman and Cohen conducted endoscopic septostomy [17]; then, in 1993, K. Oka et al. performed both aqueductoplasty and ETV using flexible fiberscopes [18]. Finally, in 1996, Mohanty et al. reported the first foraminal plasty of the foramen of Monro [19].

From the 1960s to the 1990s, numerous articles were published that retrospectively described and summarized ventriculoscopic techniques. Scarff reviewed many cases of hydrocephalus in 1966 and 1970 [20, 21], comparing ventriculostomy, choroidectomy, and shunt surgery and highlighting the many advantages of neuroendoscopic techniques. In 1998, Duffer et al. comprehensively described the application of neuroendoscopic techniques on the treatment of hydrocephalus [22], including indications, the choice of surgical methods, complications, and prognosis. Since then, neuroendoscopic technology for the treatment of hydrocephalus has been widely accepted and popularized in academic circles, and neuroendoscopic technology and its theoretical basis have been progressively improved.

Another development came in 1996 and 2002. Endoscopy was respectively used in conjunction with ultrasound and stereotactic navigation to decrease vascular injury during the procedure, and in 2004, the first instance of a robot being used in ETV was reported [23].

At present, neuroendoscopic technology is widely used in neurosurgery, especially in the management of hydrocephalus, as it facilitates a wide range of possibilities for creating alternative CSF pathways (ETV), reducing CSF production, and restoring physiological CSF pathways. The advantages of endoscopy include minimal invasion, the avoidance of brain retraction, low blood loss, fast operation time, and reduced length of hospitalization. Neuroendoscopy provides a magnified view of the ventricular system from the inside and enhances the resolution of the surgical field. Additionally, it avoids the need to implant foreign bodies and reduces the demand for re-intervention, commonly observed in patients with shunts, with the potential to prevent shunt dependency.

2. Intraventricular anatomy

The importance of mastering intraventricular anatomy and landmarks is vital to ensuring successful surgery and limiting complications. The following is a brief description of the major anatomical features encountered in ETV.

Lateral ventricles: The lateral walls of the lateral ventricles abut the genu of the internal capsule at the level of the foramen of Monro. The choroid plexus is located at the bottom of the lateral ventricle from front to back and is situated at the anterior end of the interventricular foramen. The accurate identification of the choroid plexus helps to locate the interventricular foramen and enter the third ventricle. This marker is relatively constant, allowing accurate localization of the interventricular foramen by identifying the anterior segment of the choroid plexus, even in the presence of congenital encephalocele or ventricular deformation due to tumor compression. This may not be evident in enlarged ventricles, but finding the choroid plexus in the narrow ventricular cavity may be a quick and effective way to locate the interventricular foramen; particular attention should be given to locating the choroid plexus in the treatment of split-brain syndrome.

Fornix: The fornix is located at the front and top of the foramen of Monro, where it courses from the hippocampus to the mammillary bodies. Attention should be given to intraoperative protection, especially when entering the third ventricle through the interventricular foramen in ventricular endoscopy, in which it is very easy to damage the fornix.

Thalamus: The thalamus on both sides constitutes the lateral wall of the lateral ventricle. Within the third ventricle, the lateral walls comprise the anterior two-thirds of the thalamus and hypothalamus. The superior optic nucleus and the paraventricular arcuate nucleus of the hypothalamus are the structures most easily damaged in ETV, during which the supraoptic nucleus can secrete an antidiuretic hormone, and the paraventricular nucleus can secrete oxytocin. During surgery, care should be taken not to damage the thalamic structure to avoid corresponding postoperative complications.

The floor of the third ventricle: The floor of the third ventricle is the thinning and migratory part of the thalamus on both sides, and it overlies the interpeduncular fossa. The anterior floor is formed by the optic chiasm, optic recess, infundibulum, and infundibular recess. The floor posterior is the papillary body. Within this range, there is a relatively safe area for fenestration between the infundibular recess and the papillary body. When fenestration is performed anterior to this area, the fenestration is directly opposite the clivus, so it is relatively safe to choose a position at the midpoint of this area. Usually, in the middle area of the thinned floor, there is a bowed translucent blue area that provides a safe zone in which fenestration can be performed [24]. Fenestration along the anterior floor risks injury to the pituitary, while fenestration posteriorly risks injury to the mammillary bodies and brainstem.

Liliequist membrane: The Liliequist membrane is composed of arachnoid membrane and wraps the top of the basilar artery. It is a plate-like structure that separates the suprasellar cistern of the middle cranial fossa from the prepontine cistern of the posterior fossa. After fenestration, the Liliequist membrane wrapped around the basilar artery can be seen downward, and the prepontine cistern can be observed through the membrane. Opening the Liliequist membrane during ETV is critical to the open circulation of CSF between the third ventricular and the anterior pontine pool. Clear visibility of the clivus, basilar artery, and pontine is an important marker for successful ETV.

3. Endoscopic third ventriculostomy

3.1 Patient selection for endoscopic third ventriculostomy

In the last 10 years, high-resolution endoscopic images have further extended the usefulness of neuroendoscopy in the treatment of hydrocephalus. There are indications for ventricular endoscopy for hydrocephalus. Obstructive hydrocephalus is widely considered suitable for treatment with neuroendoscopy. However, the use of neuroendoscopy in children with hydrocephalus, especially those under 2 years of age, is controversial. Satisfactory results in normal-pressure hydrocephalus and post-infection or traumatic communicating hydrocephalus have been achieved in recent years with neuroendoscopy [25, 26, 27].

Although ETV is widely used in the treatment of hydrocephalus, its success rate is below 100%, and about 25 to 40% of patients undergoing ETV require further ventricular-peritoneal shunting [25, 28, 29]. In a study by Iantosca, after analyzing the relationship between the etiology of hydrocephalus and the prognosis of ETV [30], patients with hydrocephalus were divided into three groups according to surgical success rates. The first group had a success rate of >75%, including hydrocephalus due to aqueduct sclerosis; hydrocephalus due to third-ventricle outflow obstruction, resulting from tumors, cysts, or infections; hydrocephalus due to lesions in the quadrigeminal corpus and pineal gland areas; and hydrocephalus due to ineffective shunting. The second group had a 50–75% success rate and included hydrocephalus due to occlusion of the fourth ventricular outlet, including tumors, cysts, inflammation and bleeding, and other causes; Dandy-Walker syndrome; and split-brain syndrome. The third group had a < 50% surgical success rate and included post-infection or hemorrhagic hydrocephalus and meningocele-induced hydrocephalus [without shunting].

A major point is the concept of the ETV success score (ETVSS) proposed by Kulkarni et al. [10, 31]. Based on this concept, neurosurgeons can now perform ETV surgery with confidence. The ETVSS is an estimate of the percentage probability of ETV success, ranging from 0 to 90%. There are three primary components: age, etiology, and history of shunt surgery. Although ETVSS does not determine the indication for surgery, it is an extremely practical rule that can predict surgical success rates based on only three factors.

3.2 Technical considerations of endoscopic third ventriculostomy

The desired trajectory for ETV allows for passage through the foramen of Monro and visualization of the midline floor of the third ventricle. A correct burr hole position is critical. The burr hole should be 3 cm away from the midline and anterior to the coronal suture. Planning the trajectory with frameless neuronavigation is helpful, and the burr hole is generally made on the side with the enlarged lateral ventricle and the foramen of Monro. Warm lactated Ringer’s solution is connected to the irrigation port. The use of a pressure bag is not recommended because little pressure is required. Using low pressure to dilate the ventricles gently enables both better visualization and better navigation of the endoscope into the ventricles. The confluence of the thalamostriate vein, septal vein, and choroid plexus marks the foramen of Monro.

Generally, the surgeon navigates the endoscope through the foramen of Monro into the third ventricle. When navigating the endoscope through the foramen of Monro, care should be taken to avoid injury to the fornix. A large foramen of Monro enables safe endoscopic entry, so in cases with a small foramen of Monro, a flexible thin scope is recommended. When the endoscope is well positioned within the third ventricle, the anatomical landmarks along the floor of the third ventricle should be defined. The infundibular recess, the mammillary bodies, and the basilar artery are critical structures. In patients with a thinned, bowed translucent floor, often, the basilar artery may be observed. A thick massa intermedia and interhypothalamic adhesions increase the risk of hypothalamic injury, and sometimes, ETV has to be abandoned [28, 32].

The transparent site anterior to the basilar artery and between the mammillary bodies and the infundibular recess is the ideal site for fenestration. We prefer to use grasping forceps or scissors for fenestration via the floor of the third ventricle. As critical basilar artery and posterior cerebral artery branches may be at risk of thermal injury, monopolar diathermy should be used carefully. After the floor of the third ventricle has been fenestrated, a Fogarty balloon is centered and inflated to widen the opening. Cauterizing the stoma margins may prevent the opening from reclosing. Occasionally, mild bleeding may be observed from the tuber cinereum, which may be tamponaded with the inflated balloon. Additional bleeding generally resolves with copious irrigation.

Once the third-ventricle floor is perforated, CSF enters the cistern, producing a pulsatile beating of the third-ventricle floor. The absence of such movement indicates an increased chance of ETV failure. After careful inspection of the interpeduncular cistern, the Liliequist membrane must be removed to allow the egress of CSF from the third ventricle into the ambient or prepontine cistern. The naked basilar artery in association with good stomal pulsation indicates successful ETV.

3.3 Complications of endoscopic third ventriculostomy

The overall incidence of complications after hydrocephalus with ETV varies from 0 to 31.2%, but most scholars report rates ranging from 5 to 15% [33, 34, 35, 36, 37, 38, 39]. In a meta-analysis by Madsen and colleagues, overall perioperative mortality was 0.2 ± 0.1%, and the mean ETV failure rate was 34.7 ± 18.0% [40]. Bouras and Sgouros performed a systematic literature review of 34 studies reporting 2985 ETVs performed in 2884 patients. The early postoperative mortality rate was 0.21%, and the overall complication rate was 8.5%. The rate of intraoperative hemorrhage was 3.7%, and the rate of severe intraoperative hemorrhage was 0.6% [39].

Complications such as intraoperative hemorrhage, postoperative intracranial infection, and nervous system injury seriously affect both the surgical success rate and the prognosis of patients. Familiarity with common complications after ETV surgery, timely prevention, and correct treatment of these complications play an important role in improving the success rate of surgery.

The following are descriptions of common complications after ETV:

1. Intraoperative hemorrhage: Intraoperative hemorrhage is not uncommon as an ETV complication, and its incidence varies from 0 to 8.5% [41, 42, 43, 44], although in most reports, it tends to be about 4% [39]. Intraoperative hemorrhage can occur at many stages of ostomy and may cause bleeding by damaging small vessels in the cerebral cortex and ependymal membrane during ventricular entry into the lateral ventricles. Such bleeding tends to be mild, easy to manage, mostly self-limiting, and small amounts of bleeding can be controlled by irrigation. Bleeding can also occur when the neuroendoscopy passes through the interventricular foramen into the third ventricle; this can be severe and often indicates damage to important veins, such as the transparent septum and thalamus. In such cases, hemostasis should be washed and electrocoagulation avoided as much as possible to prevent thrombosis during the electrocoagulation process and basal ganglia venous embolism. A small amount of bleeding from the floor of the three ventricles, which is not uncommon, may also occur during ETV, but it can be stopped with irrigation. The most severe bleeding during ETV is due to damage to the basilar artery. According to statistics, its incidence is about 0.21% [39], although different scholars have reported it to be about <2% [42, 45].

Common injured blood vessels are located mainly in the interpeduncular cistern. The basilar artery, the P1 segment of the posterior cerebral artery, and the posterior choroidal artery pass through the anterior half of the cistern; the perforated branches of the basilar and posterior cerebral arteries run through the posterior half, both of which are located below the floor of the third ventricle. In the case of thinning of the dilated base of the three ventricles, the surgeon can observe these vascular structures through the floor of the third ventricle. At the same time, if the stoma is created strictly according to the safe zone in front of the midpoint of the third-ventricle floor, the chance of damaging these blood vessels is very small. However, in some pathological states, the shape, position, and thickness of the three ventricular floors may change, potentially damaging the above-mentioned important vascular structures. To avoid damage to the basilar artery, attention should be given to the position of the three-ventricle floor ostomy; this should be positioned at the infundibulum crypt and the midpoint of the papillary body, as here, the bottom of the three ventricles faces the saddleback, which avoids the basilar artery to the greatest extent. Additionally, the size of the fistula should be strictly controlled at 2.5–5 mm in children and 5–10 mm in adults [46]. If there is bleeding during surgery, CSF turbidity, anatomical structure variation, or other conditions that affect surgery, the surgical procedure should be abandoned promptly [44].

1. Bradycardia: Bradycardia is reported to be a common complication during surgery, but because it is often paroxysmal, it is also the most easily overlooked. El-Dawlatly et al. reported a 41% incidence of bradycardia in ETV [47]. The cause of its formation is not fully understood. Van Aken et al. believe that the increase in intracranial pressure during surgery is the main cause of arrhythmia, so the flow rate of ventricular flushing should be strictly controlled and measured [48]. It has also been suggested that intraoperative lavage, mechanical stimulation, or temperature changes that stimulate the preoptic area can lead to slowed heart rate and decreased blood pressure [41]. Additionally, it has been proposed that after the opening of the three-ventricle floor fistula, changes in pressure in the basal pool affect the pressure in the basilar artery, thereby interfering with the blood supply to the brainstem and eventually causing arrhythmias [49]. Mild arrhythmias are often transient, but severe arrhythmias can lead to surgical termination [50]. Gentle manipulation to avoid damage to thalamic structures, control intraventricular flushing velocity, and balance flush fluid intake should be the main preventions of such complications [41].

2. Neurological injury: The incidence of neurological injury varies from 0 to 18.7% [49, 51, 52]. Almost all such injuries are related to improper endoscopic access to the lateral ventricle or endoscopic access to the third ventricle via the interventricular foramen. In particular, the incidence in pediatric patients is higher than in adult patients, which may be related to anatomical differences between children and adults. For example, the interventricular foramen is narrow in children, the papillary body is not easy to identify, and a meningocele with congenital malformations increases the difficulty of surgery. For such patients, pre-surgery magnetic resonance imaging of the patient’s head should be read carefully to observe the morphology of the ventricles and correctly assess the difficulty and risk of surgery.

Hypothalamic injury is associated with its anatomical location. The hypothalamus is located on the lateral wall of the three ventricles, and endoscopic mechanical injury, thermal injury, and irrigation perfusion injury can easily cause hypothalamic dysfunction during ETV and should be avoided as much as possible during surgery. Hypothalamic injury can trigger endocrine dysfunction, of which diabetes insipidus is the most common symptom, reportedly occurring in approximately 0.5% of ETV cases [41, 53]. Most cases of diabetes insipidus tend to self-limit, and only three cases of long-term diabetes insipidus after ostomy have been reported [53, 54].

Papilloma and fornix damage predispose a patient to memory dysfunction because both the papilla body and fornix are associated with hippocampal structures. A very small number of patients have psychiatric symptoms after surgery, which may be related to corticostomy injury to the frontal lobe. It has also been suggested that psychiatric symptoms are associated with intraoperative ventricle flushing with cold saline, so it is recommended that the rinsing fluid is preheated before intraventricular lavage [41]. If the patient has postsurgical motor dysfunction, such as limb paralysis or oculomotor nerve palsy, it may injure the nucleus of the thalamic nerve, cranial nerve, and midbrain. When the diencephalon is damaged, hallucinations or Horner syndrome may occur. Cleft injury can cause visual field defects.

In general, the occurrence of postoperative neurological dysfunction is often related to the location of the corticostomy and the rough surgical methods of the intraoperative provider. Therefore, appropriately selecting the corticostomy’s location to enable the neuroendoscopy to pass through the interventricular foramen smoothly; controlling the speed, flow, and perfusion fluid temperature of ventricular perfusion; and avoiding mechanical and thermal damage to important nerve structures of the ventricle wall have become important preventive measures to reduce postoperative neurological dysfunction.

1. Cerebrospinal fluid leakage: The leakage of CSF is a very common postoperative complication, with reported rates ranging from 0 to 5% and recent bulk meta-analyses suggesting an average incidence of 1.7% in all ETV patients [39]. This complication is particularly common in children, with CSF leakage occurring in 2–18% of cases [39, 41]. The mechanism of CSF leakage may result from a large amount of CSF entering the subarachnoid space in a short period after surgery; at this time, the absorption function of CSF is still in the adaptation period, and the secretion and absorption of CSF cannot reach an effective balance quickly, resulting in a potential fistula during surgery, an increase in intracranial pressure, and finally, CSF leakage. The reason for the higher incidence in infants and young children may be that the absorption function of pediatric arachnoid granules is not fully developed. When a large amount of CSF in the ventricle or cyst cavity enters the subarachnoid space, it cannot be absorbed by arachnoid particles; coupled with the weak healing ability of the pediatric scalp, the CSF gushes out along the ventricular endoscopic channel, resulting in CSF leakage [28].

In some cases, CSF leakage is an early manifestation of ETV failure. We believe that gentle handling during surgery to minimize damage to cerebral cortical channel and the use of artificial materials, such as gelatin sponges, to block the cerebral cortical channel after the end of the stoma can reduce the incidence of CSF leakage to a certain extent. In most patients, postsurgical CSF leakage can be effectively treated by re-suturing the skin fistula; however, for some patients with CSF leakage, it is necessary to reduce intracranial pressure by lumbar puncture or ventricular drainage.

1. Intracranial hemorrhage: Almost all types of intracranial hemorrhage can occur after ETV, for example, acute and chronic subdural hematoma, subarachnoid hemorrhage, and intraventricular hemorrhage have all been reported. However, the overall incidence is less than 1% [49]. Intraventricular and subarachnoid hemorrhages are often triggered by intraoperative injury to the corresponding vessels [28]. Some researchers have concluded that the main cause of postoperative acute subdural hematoma is the reduction in brain volume caused by excessive drainage during surgery, which further damages the bridging vein [33, 44, 55]. A small amount of subarachnoid hemorrhage tends to be self-limiting, and intraventricular hematomas can be treated with ventricular drainage. However, the treatment of subdural hematomas is relatively complex and contradictory, and although the treatment principles are contrary to those of hydrocephalus, a relative increase in intracranial pressure is necessary.

2. Central nervous system infection: The latest statistics show that the incidence of central nervous system infection after ETV is 1.81%, including meningitis at 1.6% and ventriculitis at 0.21% [49]. However, varying overall rates of central nervous system infection have been reported, ranging from 1 to 6.1% [29, 33, 56]. Nevertheless, most scholars believe that neuroendoscopic treatment of hydrocephalus for central nervous system infection is closely related to a patient’s previous history of shunt infection. There is also a view that the infection rate is age-dependent, and the postoperative infection rate in infants under 1 year of age can be as high as 11% [57]. This complication is one of the more serious complications of ETV. Fukuhara et al. identified central nervous system infection as an independent risk factor for ETV failure [28, 58, 59]. Despite the occurrence of infection, the prognosis is significantly better than that of patients with infection after shunt surgery, and in most cases, the infection can be controlled with targeted medical therapy [28, 29, 59].

3. Subdural effusion: The overall incidence of this complication has been reported as less than 2% [41, 55, 60, 61], but it can reach 5% in pediatric patients [62]. The sudden reduction in brain volume after ETV and the neuroendoscopic trajectory contribute to the subdural accumulation of CSF, which eventually leads to subdural effusion [55]. Preventing a large outflow of CSF during surgery, perfusing intraventricular irrigation fluid, and blocking the cerebral cortical sinus tract before the end of the ostomy have a certain preventive effect on the occurrence of subdural effusion. The treatment of subdural effusions is as complex as that of subdural hematomas, and large subdural effusions should be treated with external drainage and appropriate elevation of intracranial pressure [55, 63].

4. Late endoscopic third ventriculostomy failure: The rate of delayed failure is reported in the literature to range from 2 to 15% [35, 36, 56, 64]. Late failure resulting in rapid clinical deterioration is rare but has been described [25, 28]. Age appears to be a strong predictor of ETV success, with infants less than 6 months old having the highest risk of failure, in some cases up to a fivefold increase in risk compared with that of older patients [28, 65, 66]. The mechanism of ETV failure may include the following factors: an inadequately sized stoma; impaired flow through the stoma; hemorrhagic obstruction facilitating closure or narrowing of the stoma; elevated CSF protein and fibrinogen; impaired CSF absorption by the arachnoid granulations, particularly in the context of infection; and tumor progression resulting in stomal blockage [28, 63].

3.4 Fenestration of loculated hydrocephalus

The use of the endoscope has been explored for other complicated forms of hydrocephalus. Septostomy can be performed endoscopically to treat isolated lateral ventricles. Recently, aqueductoplasty was reported in the treatment of trapped fourth-ventricle syndrome. Neuroendoscopic techniques have been extended to foraminoplasty of the foramens of Monro and Magendie as well as to endoscopic fourth ventriculostomy [67, 68, 69]. The fenestration of loculated hydrocephalus is another endoscopic technique widely used in the treatment of hydrocephalus [38, 70, 71, 72]. The presence of one or more non-communicating isolated compartments within the ventricular system that tend to enlarge despite a functioning shunt system is defined as loculated hydrocephalus [73]. Intraventricular septations or obstructions between the site of CSF production and the tip of the ventricular catheter can present a barrier to the flow of CSF and result in an accumulation of fluid in the excluded compartments [70]. Intraventricular septations develop secondarily from meningitis or ventriculitis, germinal matrix intraventricular hemorrhage, congenital brain anomalies, and postoperative gliosis [38, 70]. Patients with loculated hydrocephalus and intraventricular arachnoid bands or arachnoid cysts with extension into the ventricles are ideal candidates for intraventricular neuroendoscopy.

Historically, the management strategy for multiloculated hydrocephalus has been to place multiple catheters in isolated compartments. However, the use of multiple shunts has been associated with high morbidity and mortality owing to the cumulative risk of shunt infection and mechanical failure [38, 72]. Nowadays, the goal of surgery has evolved to create as wide and as many communications as possible among the isolated intraventricular compartments to achieve a simplified ideally unique ventricular environment. The fenestration procedure aims to create communication between the maximal possible number of isolated parts of the ventricular system by making openings in the septal walls to achieve CSF drainage using the minimal number of shunts; this, in turn, will lessen the frequency of shunt revisions [33, 71, 74, 75].

The first large research series on the endoscopic treatment of loculated hydrocephalus was published by Lewis et al. in 1995 [76]. Endoscopy reduced the annual rate of shunt revision from 3.04 to 0.25%. Subsequently, neuroendoscopy became popular [77]. Nowosławska et al. published the only study where the neuroendoscopic treatment of loculated hydrocephalus was compared with the conventional implantation of shunt systems (43 endoscopy/80 control group) [78]. They concluded that neuroendoscopic cases experienced a significantly higher clinical improvement, a better outcome (86 vs. 60% with a good outcome), fewer postoperative complications (23 vs. 66%), and fewer reoperations (average of 1.76 vs. 7.05 operation/patient, 3.95 vs. 1.02 revisions/year) compared with patients in a shunt control group.

In most cases, neuroendoscopy is considered the treatment of choice for loculated hydrocephalus, but if it fails, shunt placement is often required. Consequently, the second-best treatment is considered to be the implantation of a simplified shunt system. Most neuroendoscopic procedures appear to achieve these goals in loculated hydrocephalus [38, 70, 79]. Following endoscopic fenestration, fewer shunt revisions per year are necessary [33, 73, 79]. In 2007, Spennato et al. reported a reduction from 2.07 to 0.35%, with a single shunt or no shunt present in 83.3% of cases [77]. In El-Grandour’s study of uniloculated hydrocephalus including 31 children cases, the annual shunt revision rate fell from 2.7 to 0.25% [79]. Alice Noris’s study seems to confirm these findings, with 75.6% of patients having no shunt or only one shunt implanted after neuroendoscopy [70]. Despite these encouraging findings, the rate of patients without shunts is low. In their study, the shunt-free rate was 28.9% [70], which is in line with data reported by other authors (26, 28, and 10% in Spennato’s, Teo’s, and Peraio’s series, respectively) [33, 77, 80].

3.5 Technical considerations of the endoscope for loculated hydrocephalus

Neuronavigation is useful in maintaining orientation during surgery and avoiding damage to neural structures, even though the risk of a substantial shift in anatomical landmarks after fenestration of the isolated fluid compartments remains a major problem [64, 70, 71, 81]. The reported disadvantage of neuronavigation, namely the occurrence of an intraoperative brain shift that reduces the accuracy of the navigation for cerebral surgery, can be overcome during ventricular endoscopic neuronavigation; during this procedure, the most important consideration is the direction to follow. Gradual step-by-step opening of the isolated compartments to minimize CSF loss and continuous irrigation throughout the procedure are effective methods for minimizing the brain-shift effect during neuronavigation-aided ventricular endoscopy. These steps maintain the existing anatomy and dimensions of the cysts and the ventricular system as far as possible throughout the entire procedure [38, 81].

Surgical procedures are tailored according to the pathologic entity and the site of septations. In most cases, surgery can be performed through a coronal burr hole, including septostomy, cyst fenestration, and aqueduct stenosis. For fenestration, we prefer a rigid endoscope rather than a flexible one. The advantage of this system is that it enables a better view and better light. Compared with flexible endoscopes, the limited maneuverability of rigid instruments requires more careful placement of the burr hole and widening of the outer edge of the burr hole in selected cases [38].

Care should be taken during fenestration and resection to coagulate the cyst wall along its thickened portions before resection, as these portions of the wall may be vascular. Many instruments can be used to open cyst walls. The author used a thulium laser to fenestrate the septations. Because it is designed to operate in wet environments, and due to the very contained diffusion of energy from its tip (as low as <1 mm in depth), this instrument is much safer and more efficient than others, such as those used in monopolar cauterization. Fenestrations should be as wide as possible to reduce the risk of closure of the fenestration sites. Subsequent enlargement should be performed using grasping forceps and a balloon, with portions of the cyst wall removed using scissors and grasping forceps. If a ball valve exists at the basilar artery near the base of the clivus from an arachnoid cyst, complete resection of that aspect of the cyst may be required, unlike in fenestration, to prevent the symptoms from recurring.

3.6 Complications

The endoscopic treatment of loculated hydrocephalus has been associated with a reduced rate of perioperative morbidity due to its minimal invasiveness compared with that of open surgery via craniotomy. The endoscopy complication rate ranges from 10 to 20% [70, 71, 72, 77, 82, 83]. Complications following the endoscopic approach are usually minor and are mostly treated conservatively. The most common complications include CSF infection, thalamic hematoma with transient akinetic mutism, minor intraoperative bleeding, bilateral subdural hematoma, and seizure.

4. Limitations of the endoscope in the treatment of hydrocephalus

While neurosurgeons increasingly advocate the use of the endoscope in the treatment of hydrocephalus, the development of instruments for endoscopic surgery has not followed the same pace. Neuroendoscopy can be improved by the use of even smaller optics with higher resolution than are currently used, providing more surgical space. The endoscope itself occupies space in an already limited surgical corridor. The ideal endoscope is thin and sturdy, does not generate heat, and provides high-resolution images. In addition, a self-irrigating feature could minimize the need to remove and reinsert the endoscope for cleaning.

In addition to the inadequacy of the instruments used, ETV surgery itself has certain limitations. The procedure is inferior to shunt placement in terms of acute infection, premature infants, and post-infective and post-hemorrhagic hydrocephalus. When ETV alone is ineffective, a CPC or VP shunt will be required. Moreover, multiloculated hydrocephalus is a complex and challenging condition, and endoscopic fenestration alone is not effective in controlling it. Therefore, endoscopic-assisted VP shunt insertions are useful for reducing shunt complications and should be considered a therapeutic option.

Conflict of interest

The author declares no conflict of interest.