Neuroendoscopic septostomy in unilateral and bilateral ventricular hydrocephalus

This study describes the results of septostomy (SPT) in terms of success and analysis of follow-up in a series of pediatric patients diagnosed with unilateral ventricular hydrocephalus (UHV) and biventricular hydrocephalus (BVH). A total of 29 pediatric patients diagnosed with UVH or BVH were included in this study. In UVH, a neuroendoscopic SPT was performed (sometimes accompanied by NEFPFMO). In those diagnosed with BVH, an SPT combined with VPS was carried out. Demographic, etiological, clinical, and diagnostic variables and percentage of treatment success were collected. SPT was successful during follow-up when no VPS was required in UVH and only unilateral VPS was implanted in BVH. At the time of surgery, 16 patients needed a ventriculoperitoneal shunt. The first SPT was successful in 22 patients, requiring a second surgery in 7 patients, from the oncology group diagnosed with BHV. The surgical management of UVH and BVH still has some disclosure points to be reviewed. However, SPT seems to be a secure, non-traumatic, and efficient procedure.


Introduction
Pediatric hydrocephalus has been widely sub-categorized in ventricular with unilateral and bilateral impairment [1]. Unilateral ventricular hydrocephalus (UVH) was defined by Dott in 1927 as a condition where one lateral ventricle, in whole or part, becomes dilated due to obstruction of the cavity or its outlet [1]. Biventricular hydrocephalus (BVH) occurs because of a bilateral obstruction in the foramen of Monro. In 1985, Oi and Matsumoto defined the term progressive unilateral obstructive hydrocephalus [2]. They identified four categories based on the condition of the foramen of Monro that applies for both UVH and BVH: (1) congenital atresia (with or without membrane), (2) acquired obstruction (intraventricular hemorrhage/infection), (3) functional obstruction (because of shunt over drainage), and (4) patent foramina with asymmetrical parenchymal abnormalities (i.e., alterations in compliance). UVH is most commonly secondary to neoplastic or infectious causes in the pediatric series [3].
There are two therapeutic procedures strategies to treat UVH/BVH, either single or combined. First, the ventriculoperitoneal shunt (VPS) is considered a simple method, although high rates of complications and revisions have been reported, especially in infants [4][5][6][7]. Second, more recent neuroendoscopic procedures such as endoscopic third ventriculostomy (ETV), septostomy (SPT), and/or neuroendoscopic foraminal plasty of the foramen of Monro (NEFPFMO) have been proposed [8,9]. There is limited research describing UVH and BVH treatment. However, other surgical approaches have become relevant in recent years, such as SPT alone for UVH (sometimes combined with NEFPFMO) and SPT combined with VPS or NEFPFMO for BVH [10,11].
This study aims to describe the results of SPT in terms of success and analysis of follow-up in a series of pediatric patients diagnosed with UVH or BVH.

Methods
This study was conducted in the neurosurgical service of the Regional University Hospital of Malaga in Spain. A case series study of primary sources was carried out during the hospitalization and outpatient clinic follow-up, from August 2016 to December 2019.
Participants included were all patients between newborn and 15 years old who were diagnosed with UVH or BVH. In UVH, an SPT was performed (sometimes accompanied by NEFPFMO, if necessary). In those diagnosed with BVH, an SPT combined with VPS was carried out. During follow-up, neuroendoscopic SPT was considered successful when no VPS was required in UVH and when only unilateral VPS was implanted in BVH.
Participation was strictly voluntary for the parents by informed consent. Demographic information (including gender, age of the surgery), clinical background (including symptoms and previous shunt insertion), diagnosis, and treatment were recorded. Symptoms were categorized in two settings according to their characteristics: onset time, intensity, and ease of evaluation before and after surgery. The first group presented high-pressure symptoms such as headache, dizziness, vomiting, cranial nerve pathology, and papilledema. Meanwhile, the long-term symptoms were macrocephalic, altered psychomotor development, gait disorders, sphincteric alterations, and epilepsy. The diagnosis was grouped into three categories by etiology: oncology, congenital, and secondary encompasses infectious/hemorrhagic. In addition, surgical details included indications for surgery, type of surgery (septostomy (SPT) with/without additional procedure, ventriculoperitoneal shunt (VPS), ventriculocistostomy (VC), NEFPFMO, endoscopic tumor biopsy (EB), and/or Ommaya catheter placement (OCP)), complications, reintervention, and follow-up time in months.
The SPT technique was carried out with the patient in a supine position and with the head slightly flexed and fixed in a three-pin Mayfield clamp or horseshoe head. In that position, we located a 1 cm precoronal and 2 to 3 cm far from the midline mark to produce a single burr hole. After the dural section, we always access endoscopically through the widest ventricle. Neuronavigation was used when ventricle anatomical anomalies were evidenced in pre-surgery brain resonance. Inside the ventricle, we identify the septum pellucidum. Then, we fenestrated in the anterior vascular area, perforating layer by layer until creating a 5-10 mm stoma using a number 4 Fogarty catheter, bipolar coagulation, and tweezers. We withdrew the endoscopy and closed the bur hole with surgical sealer and the skin with sutures.
Categorical variables were summarized as frequencies and percentages, and continuous variables were described using median values and interquartile ranges (IQR) or mean and standard deviation, when appropriate. Data were analyzed with SPSS version 21 software.

Results
A total of 29 patients were included in the study. More than half were male, with a mean age at the time of surgery of 73 months. In addition, 9 patients were carriers of VPS. Three patients had a history of valve infection. Almost a third-quarter of the patients presented symptoms of intracranial hypertension. The group of tumoral hydrocephalus etiology presented almost all BVH (it was mainly III-IV tumors with bilateral Monro obstruction), while in cases of congenital etiology or secondary to infection/hemorrhage, UVH was the most frequent group. The characteristics of the series and the differences by etiological groups are described in Table 1.
Septostomy was combined with different endoscopic maneuvers, depending on the type of hydrocephalus (UVH/ BVH) and its etiology ( Table 2). At the time of surgery, 16 patients have implanted a ventriculoperitoneal shunt. These 16 patients had BVH (13 from oncologic, 2 from secondary, and 1 from congenital), for all of which SPT + DVP was carried out. In 2 patients diagnosed with congenital UVH, SPT was combined with NEFPMO, as both presented membranous stenosis of the foramen of Monro. In 8 cases, neuronavigation was used during the surgical procedure, mainly in the congenital group, due to distorted ventricular anatomy evidenced in the preoperative image. After the surgical procedure, no relevant complications related to the procedure were found. The mean follow-up time was 46 months (SD, ± 29).
In the oncology group composed of 14 patients, an endoscopic biopsy was performed in 12 children with BVH, and due to bilateral obstruction of both foramina of Monro, a VPS was implanted in 13 of them. In the same group, in one patient diagnosed with cystic craniopharyngioma, SPT was combined with an Ommaya reservoir, and in another 2 cases, SPT was accompanied by VC for communication of intraventricular cysts or other septate cavities with the lateral ventricle.
The first SPT procedure was successful in 22 patients (75.9% of the total). In contrast, 7 patients underwent a second surgery to resolve the hydrocephalus, 4 of these patients had BVH secondary to a tumor, and a new SPT was carried out with valve replacement. Meanwhile, the other 3 patients had a congenital UVH and were treated with an SPT initially, which failed, and a new SPT was performed with good clinical and radiological follow-up results. Also, a third surgery was necessary for 3 oncologic patients (10.3% of the total) with BVH, to whom a DVP was replaced in 1 patient and added in the other 2, with excellent clinical results. According to etiologic subgroups, the oncology had the highest frequency of the second and third surgery in 35.7% (n = 5) and 21.4% (n = 3), respectively. Also, there were 3 deaths in this group during the follow-up due to a progression of their oncology condition, unrelated to the neuroendoscopic procedure (Table 2).

Review of current surgical disclosures about technical surgery
The septum pellucidum (SP) is a translucent structure placed in the midline. It is part of the limbic system and is connected with the amygdala, habenula, and brainstem reticular formation [12]. The SP is composed of two parts: the upper SP contains only fiber tracts and the lower part called septum verum (anterior to the foramen of Monro (FOM) and above the rostrum of the corpus callosum(CC)) which contains various nuclei including the septal nuclei, the diagonal band of Broca, the bed nucleus of the anterior commissure, and the bed nucleus of the stria terminalis [13] The SP contains no arteries but one to three veins draining in the thalamostriate vein, with an asymmetric distribution in greater than 50% cases. In most cases, the large septal veins are located anterior to the foramen of Monro [14].
Currently, there is an ongoing controversy about the technical specifications for endoscopic septostomy. The first sticking point is the position of the patient's burr hole to access the lateral ventricle. Most authors advocate performing a single 1 cm precoronal drill, 2 or 3 cm from the midline, to get a trajectory more perpendicular to the septum pellucidum [10,13] However, Tamburrini et al. consider a standard precoronal burr hole to be disadvantageous, arguing that it does not allow a sufficient view of the septal anatomy, making it difficult to perforate the septum. Therefore, they propose to make a more lateral frontal horn approach to a standard Kocher burr hole (1 cm precoronal, 3-5 cm from midline) [8]. Oertel et al. recommend an even farther lateral burr hole, up to 7 cm lateral to the midline and slightly anterior to the coronal suture. If a simultaneous ETV is planned, they recommend placing the burr hole up to 4 cm lateral to the midline to allow a straight approach to the floor of the third ventricle [15]. In the series, we carried out a single precoronal burr hole of 1 cm and 3 cm away from the midline. Entering the smaller or larger lateral ventricular cavity is another issue [8,16,17]. There is opposition among several authors, while some defend the approach by the smallest to avoid the decrease in the pressure gradient between the enlarged and normal ventricle, with the possibility of affecting the permeability of the septostomy [17]. In the situation of the contralateral slit ventricle, Tamburrini et al. prefer to enter the affected lateral ventricle with the aid of navigation to minimize the risk of injury to the contralateral thalamus when carrying out septostomy [8]. In contrast, other studies prefer to approach the largest ventricle due to its ease (even without neuronavigation) and greater workspace [10,15]. In our series, we executed the septostomy, accessing the widest ventricle and use of neuronavigation in 8 patients (27.5%) when anatomical distortion was evidenced. Neuronavigation exhibits the required depth and target point when anatomical orientation is lacking. The successful combination of frameless neuronavigation with neuroendoscopy has been previously reported [18,19]. Rohde et al. found that neuronavigation was beneficial in 52.7% of septostomy cases performed with tumor biopsies, by allowing better anatomical orientation despite distortion caused by tumor growth [20]. According to Oertel et al., a septostomy can be carried out without neuronavigation, but they mention two situations in which it would be useful. The first situation is when we need to calculate the ideal approach for septostomy if it is combined with tumor biopsy or ETV, and the second is when the septum pellucidum is pushed contralateral with possible adherence to the contralateral thalamus causing a thalamic lesion with fenestrations [15].
The ideal site to fenestrate the septum pellucidum is not uniformly described in the literature [10,11,15]. Schroder et al. mention that the area of fenestration depends on the individual anatomy [16]. The septum pellucidum is divided into a frontal segment, which is the anterior region to the foramen of Monro, and an atrial segment, which is the posterior region to the foramen of Monro. In the frontal segment, there is a triangular avascular area delimited by the inferior septal vein, the floor of the frontal horn, and the corpus callosum. This triangular area is sufficiently translucent to allow the visualization of the contralateral part of the septum pellucidum, and it is described as safe for fenestration. The recommended point site corresponds approximately 1.0 cm superior and 2.0 cm anterior to the superior margin of the foramen of Monro, also used by Aldana et al. [10] and Vinas et al. [12]. Moreover, there are two other avascular areas in the posterior atrial segment. The first, one avascular area, is between the superior septal vein and the posterior margin of septum pellucidum, and another avascular area is identified in the atrial segment between the two septal veins [14]. However, Mohammad et al. considered both of these areas inconsistent in their size and shape [14]. Meanwhile, Oertel et al. carried out the fenestration of 5 to 10 mm posterior to the foramen of Monro [15]. Hamada and Tamburrini et al. performed the fenestration between the anterior and the posterior septal vein [8,11]. Finally, in contrast, we did the fenestration in the anterior avascular area, perforating layer by layer until creating a 5-10 mm stoma, using a number 4 Fogarty catheter, bipolar coagulation, and tweezers.

Role of SPT in the treatment of UVH/BVH
There is currently a debate on several surgical procedures to manage UVH/BVH in the pediatric population. There are two therapeutic procedures and strategies to treat them, either single or combined. First, the ventriculoperitoneal shunt (VPS) is usually considered the simplest method, although higher rates of complications and revisions have been reported, especially in infants [4]- [7]. Second, more recent neuroendoscopic procedures alone or combined have been proposed as septostomy (SPT) and/or neuroendoscopic foraminal plasty of the foramen of Monro (NEFPFMO) [8].
Few published reports have addressed the results of SPT among other endoscopic procedures, and only eight series deal with the technical aspects of this procedure [10,11,15,16,21,22]. At the moment, our report is the third longest in the literature after Aldana and Oertel et al. [10,15]. Concerning all the series (n = 8), cases are described with a range of age between 5 months and 44 ages, and up to half of the series (n = 4) have included the pediatric population (Table 3). In the same way, surgical management in four series was SPT alone [10,17,21,23], Schroeder et al. only combined procedures [16], and three series both single and combined [11,15,22]. In our series, half of the patients were managed with a unique surgical procedure while the other half with combined procedures. Of these viable combinations, Gabb et al. described an experience with 75% of SPT plus ETV and 25% SPT plus VPS; meanwhile, Oertel et al. mixed 88% of SPT plus ETV and 12% SPT plus VPS, Oi et al. described a case of SPT plus NEPFMO, and also Hamada presented one case with SPT plus VPS [11,15,16,22]. These findings were like to our study, in which an SPT plus VPS was reported in almost half of the patients. In the overall series, Aldana reported the longest follow-up period with a mean of 31 months [10]. Interestingly, our study could reach a mean of 46 months. Hayashi reported a reoperation proportion in half of the patients, while Oertel and Aldana reported a proportion of one-third and Hamada only 10% of cases [10,11,15]. Our report reached a reoperation in almost one-quarter of patients.
In conclusion, surgical management of UVH and BVH still has some disclosure points to be reviewed. However,  SPT and NEFPFMO seem to be secure, non-traumatic, and efficient procedures. Larger pediatric population series have to be analyzed to establish clinical care guidelines.