Training Models for Skull-Base and Vascular Micro-Neurosurgery

Abstract

This overview presents computer-based augmented reality (AR) and virtual reality (VR) tools, in-vitro and in-vivo models as useful teaching tools for neurosurgical training, especially in skull-base surgery. An easy set-up and practicable training model for ventricular drainage (VD) is demonstrated. The model allows to evaluate practices, pitfalls and traceability in a virtual but realistic set-up for simulating VD placement. Computer-assisted planning and simulation of skull-base approaches and integration within the daily neurosurgical routine with VR and AR models are discussed for neurosurgical education. A set-up for microvascular training on a plastic rat and a specific vascular anastomosis practice kit with different tube diameters of 1–3 mm of specific plastic vessels for the training of microvascular anastomoses is shown. End-to-end and end-to-side anastomoses were performed with different levels of difficulty, trying to simulate realistic conditions in bypass surgery. Additionally, the teaching strategy of experimental silicone aneurysm clipping in a 3D printed plastic skull and silicone brain model is demonstrated in video sequences. An experimental animal model with microsurgically created bifurcation aneurysms is discussed because this training model for clip occlusion of aneurysms on a living object, still has the greatest relevance to neurosurgical reality.

1. Introduction

Today the younger neurosurgical generation has a reduced possibility for practical training in nearly all fields of neurosurgery. Accordingly, neurosurgical training models show increasing popularity. However, skull-base surgery often has a higher level of difficulty and is therefore rarely integrated into the basic surgical training of younger neurosurgeons. Even sophisticated modern, computer-based 3D models cannot adequately replace the important practical and hands-on surgical training. For this reason, we need realistic and gradually coordinated practical scenarios to learn and practice basic neurosurgical skills in sequential steps. Here we present some easy-to-implement practical examples that have proven themselves for operative training in skull-base surgery and for vascular microsurgery.

The number of skull-base procedures as well as microsurgical clipping of cerebral aneurysms is continuously decreasing. Meanwhile, the endovascular treatment of cerebral aneurysms has markedly reduced the number of microsurgical clipping procedures. The result will be a reduced possibility for practical training in aneurysm surgery, especially for the younger generation [1, 2, 3]. Stereotactic radiosurgery on one hand, and the increasing number of patients successfully treated by endovascular techniques on the other hand, will further reduce the overall caseload in skull-base and vascular microsurgery in the next years. However, especially huge, calcified or wide neck aneurysms will remain for microsurgical clip occlusion and will be a challenge for cerebrovascular neurosurgeons in the future. On the other hand, revascularisation and bypass techniques will be more and more in demand [4, 5]. Due to these conditions, concepts and realistic models for practical microsurgical training are necessary to improve the practical skills in neurosurgeons during their training, especially in the field of vascular and skull base surgery [6, 7, 8, 9, 10, 11, 12].

2. Computer-assisted training tools

2.1 Ventricular drainage (VD) model

Various VD training models have been developed, tested and published [13, 14, 15]. Recently, we performed another prospective study to evaluate practices, pitfalls and traceability in a realistic but virtual set-up for simulating ventricular drainage (VD) placement as a practicable training model using our navigation system (Brainlab, Munich, Germany), the navigation pointer functioning as a virtual VD catheter to be placed in the cMRI of a healthy subject (Figure 1). We evaluated accuracy by repeated virtual freehand VD placement on the prone subject using anatomical landmarks by neurosurgeons-in-training and non-neurosurgical staff. The influence of the level of neurosurgical knowledge and training level were of overall interest, especially in narrow ventricles. It was found that accurate VD placement correlated with neurosurgical experience (Figures 2 and 3). These initial results are not yet published, however.

Figure 1.

Study set-up for virtual VD placement. The navigation pointer functions as a virtual VD catheter to be placed in the cMRI of a healthy subject.

Figure 2.

A selection of virtual VD trajectories by participants with neurosurgical experience including the ‘ideal trajectory’ in blue.

Figure 3.

A selection of virtual VD trajectories by participants without neurosurgical experience including the ‘ideal trajectory’ in blue.

This simple set-up is an easy model for continuous neurosurgical training, with the aim of optimizing the medical care of our patients and constantly improving the level of education. Its set-up could be varied further, for example, as a good training module for simulating an intraoperative emergency puncture of the ventricle, exemplary in an atypical pterional positioning of the head. This poses an unusual situation that requires high surgical skills with an excellent three-dimensional spatial concept.

2.2 Augmented reality (AR) and virtual reality (VR) training tools

Navigation systems offer a multitude of technical options that can be used for training and further education [16, 17]. Brainlab has been offering the function of transparent reflection in a superimposed display of previously segmented anatomical structures for over 20 years (Figure 4). This head-up display, as an early AR tool, in which the pre-planned and segmented structures are faded into the ocular of the operating microscope or onto the screen, allows optimal orientation and is didactically very valuable. However, the accusation is justified that we generally use these technical possibilities of virtual reality or augmented reality, offered as multiplanar visualization or as 3D reconstruction view, far too little for teaching and training purposes [17].

Figure 4.

Intraoperative screenshot during a skull base procedure performed on April 12, 2002, shows then state-of-the-art technology with intraoperative AR tools that were already available at the time. The preoperatively segmented relevant anatomical structures are demonstrated and the navigation pointer shows the trajectory to the clival meningioma via a combined antero-sigmoidal/sub-temporal approach. The dominant sigmoid and transverse sinus (green) and the lateralized basilar artery (magenta) displaced by the tumor are visualized.

Recently, our department has been working with AR 3D models (UpSurgeOn S.r.l., Milan, Italy) to improve surgical and anatomical skills as well as presurgical positioning of the patient (Figure 5). For this, the department acquired partially reusable 3D models of the most common cranial neurosurgical approaches including certain pathologies (e.g., intracranial aneurysms) (Figures 6–8). The 3D models are made of synthetic materials which render them extremely lifelike in haptics and handling. They can be used as are or in conjunction with an AR app (UpSurgeOn Neurosurgery S.r.l., Milan, Italy), which walks the surgeon through the entire procedure by fusing a virtual image with the actual model, starting with optimal positioning of the patient for the specific surgery (Figures 9–11). The app also allows an image of the patient awaiting the planned procedure to be projected into any space, allowing the surgeon a 360° view. Training sessions were held for neurosurgery residents, offering them the opportunity to practice neurosurgical approaches safely, including craniotomies, drilling with various bits, brain retraction, basic intradural dissection and even aneurysm clip placement. We acquired a navigational data set for one of the models by placing it in a CT-scanner and uploading the data set it into our navigation system (Brainlab, Munich, Germany), thereby giving participants the additional chance to practice the procedure image-guided in a non-bloody and risk-free manner whilst getting more familiar with the navigation system itself (Figure 12). Our results concerning the efficacy of this kind of modern training model have yet to be published; however, anecdotally, residents report feeling more secure in their surgical approaches, in choosing and handling surgical drills as well as in the positioning of patients for surgery. Some residents now use the UpSurgeOn App to plan surgeries in advance or to double-check the preoperative positioning of the patient by overlaying the actual patient with an AR model.

Figure 5.

Our setup for a cadaver-free manual training session with AR 3D models (UpSurgeOn S.r.l., Milan, Italy).

Figure 6.

Our setup for a cadaver-free manual training session with AR 3D models (UpSurgeOn S.r.l., Milan, Italy). The models can be reused by purchasing additional craniotomy covers as depicted here.

Figure 7.

Close-up of the ‘aneurysm box’ including the craniotomy model, the synthetic brain and vessels and the associated QR code for use in conjunction with the Upsurgeon App for an AR component. The right-sided aneurysm is partially visible through the Sylvian fissure.

Figure 8.

Retracting the synthetic brain to expose the right-sided MCA aneurysm. The materials used render the models extremely lifelike in haptics and handling.

Figure 9.

The models can be used as are or in conjunction with an AR app (UpSurgeOn Neurosurgery S.r.l., Milan, Italy) via a QR code as pictured here. The materials used render the models extremely lifelike in haptics and handling. Depicted here is a right-sided pterional approach with craniotomy already performed. The underlying synthetic dura is visible.

Figure 10.

Here the 3D model is fused with a virtual image via the accompanying QR code, allowing the surgeon to explore deeper intracranial structures of the specific surgical approach.

Figure 11.

Hybrid view of one of the 3D models showing a right-sided ICA aneurysm.

Figure 12.

The navigational data set for the 3D model of an MCA aneurysm.

Due to the relative ease and cost-effectiveness of implementing the aforementioned tools, we consider simulation-based cadaver-free training with AR a promising option to bring skull-base surgery training to a new level.

The intense occupation with radiological images, necessary for the segmentation of a tumour or any other lesion and the marking of the anatomical structures, already brings great didactic benefit to the colleagues who are in training. During surgery, the recognition value of the anatomical structures rendered in the CT or MRT as a three-dimensional surgical site under the surgical microscope provides the greatest learning value. The use of navigation data for the dedicated operation planning and the daily discussion of the surgical cases are essential for this. Particularly, in the case of complex skull base operations, the position, craniotomy, microsurgical strategy and resulting operation-specific risks should routinely be discussed with the assistant physicians on the day before the surgical procedure in a pre-op conference. In particular, the discussion of surgery-specific anatomy on the basis of the 3D navigation data provides optimal preparation for these complex neurosurgical interventions [17]. So the following day, the display of the operation-specific anatomical structures via the head-up display during surgery will provide a much higher didactic benefit for complex skull base procedures (Figure 13).

Figure 13.

The operating microscope (Neuro NC4, Carl Zeiss) was linked and registered with the navigation system (Brainlab VectorVision). After the combined craniotomy and opening of the dura, AR overlay head-up display depicts the dominant sigmoid and transverse sinus (green), the basilar artery (magenta) and the clival meningioma below.

3. In-vitro microsurgical models

Here we present two sophisticated in-vitro artificial plastic models which closely mimic real vascular conditions and a more complex in-vivo experimental animal bifurcation aneurysm model in rabbits [18]. For initial and basic microvascular training, the poly-vinyl-chloride (PVC) rat model (Microsurgical Developments Foundation, Maastricht, The Netherlands) shows itself to be more or less sufficient [19]. For a more realistic haptic feeling and enhanced specific microvascular training, a model with highly elastic polyvinyl alcohol (PVA) vessels and an anatomical plastic head (Kezlex, Ono & Co., Ltd., Tokyo, Japan) is available [18, 20]. Additionally, our highly sophisticated in-vivo experimental animal model (rabbit carotid artery bifurcation model) is demonstrated and discussed [21, 22, 23].

3.1 PVC rat model

Using the PVC rat model for ‘unbloody training’ of microsurgical techniques and improvement of practical skills is a perfect example of the replacement of living animals (Figure 14). The number of live animals used for in-vivo training will likely be reduced in the future, therefore these in-vitro methods will be needed to make the transition to in-vivo models easier. The PVC rat model (Microsurgical Developments Foundation, Maastricht, The Netherlands) should be the first step in the practical education and allow various microsurgical training situations. The replacement of the plastic vessels is easy and relatively cheap [19]. Using a specific nozzle at the back of the rat model, the colored vessels can be filled with water, thus permitting an easy check of patency and quality of the anastomosis.

Figure 14.

In-vitro model with plastic vessels of the abdominal cavity of the PVC rat (Microsurgical Developments Foundation, Maastricht, The Netherlands).

In the prospective part of our published study [18], end-to-end and end-to-side anastomoses were performed with three different levels of difficulty in the PVC rat model. In total six surgeons with different expertises and different levels of vascular training and surgical skills performed these microsurgical procedures. Different sizes of plastic tubes of various lengths and diameters were used for reduction of the surgical approach and the workspace, to adapt the experimental set-up to a scenario with different degrees of difficulty (Figure 15). Those plastic tubes reduce the operative field and consecutively the surgical working space and determine the trajectory and the direction for the instruments. Due to this focused working channel, the degree of freedom for using the instruments is restricted and therefore, the level of difficulty to perform an adequate anastomosis increases significantly (Figure 15). The different sizes of the plastic tubes mimic intraoperative conditions in a narrow and deep surgical field. Plastic tube I (Advanced) has a diameter of 40 mm and a depth of 15 mm. Tube II (Expert) has a diameter of 30 mm and a depth of 35 mm (Figure 15). Tube III (Master) has a diameter of 25 mm and a depth of 45 mm. For the anastomoses, we used conventional microsurgical instruments and monofile polyamid sutures 8/0 (BV 2 needle), respectively, 10/0 (BV 100-4 needle) Ethilon (Ethicon, Johnson & Johnson MEDICAL GmbH, Norderstedt, Germany).

Figure 15.

A transparent plastic tube was brought into the operative field and fixed to a conventional flexible retractor system to reduce the working space and mimic surgical conditions in deeper approaches. In this narrow workspace, the level of difficulty to perform an anastomosis significantly increases.

In this experimental set-up, the increase of surgical complexity by reducing the workspace with the different plastic tubes clearly demonstrates that the time of surgery to perform the anastomosis increased significantly (Figure 15). In addition, the rate of incorrect sutures of the vessel wall increased, the narrower the surgical field became due to the decreasing diameter of the tube. Therefore, the overall patency rate of the anastomosis was clearly reduced with increasing grade of complexity [18].

3.2 PVA vessel with craniotomy site

The wet PVA vessels of the vascular anastomosis practice kit are transparent, highly flexible and soft [20]. However, they have to be kept moist and tend to dry out and lose their elasticity, especially under the high-energy xenon light of the operating microscope. At present, we have used the PVA vessels of the vascular anastomosis practice kit for various experimental anastomoses. Compared to the PVC vessels in the rat model, the preparation and handling of the PVA vessels, especially the grasping with a forceps or the insertion of the needle into the vessel wall is much more realistic and closely mimics human conditions.

An even more realistic scenario is provided by the plastic skull model with relatively soft and deformable silicone brain material and the vascular anastomosis practice kit with highly elastic plastic (PVA) vessels (Kezlex, Ono & Co., Ltd., Tokyo, Japan). The very soft and elastic plastic vessels of the vascular anastomosis practice kit are available in three different diameters: 1 mm, 2 mm and 3 mm. The vascular kit with the humid PVA vessels has a perfect haptic feeling during preparation, cutting and suturing of the vessels. The whole set-up with the deformable plastic brain, the realistic feeling of retraction and especially the optical impression under the microscope generate an overall aspect of a real microsurgical scenario closely mimicking human conditions.

The 3D model with pterional craniotomy and the deformable frontal and temporal lobe (Kezlex, Ono & Co., Ltd., Tokyo, Japan) is a perfect in-vitro model to simulate an opened Sylvian fissure for experimental bypass, as well as aneurysm surgery training.

This set-up allows a realistic retraction of the Sylvian fissure, its handling and preparation closely imitating human conditions (Figure 16).

Figure 16.

This plastic skull and silicone brain model shows an extended pterional approach with slight retraction and opening of the Sylvian fissure. An end-to-side anastomosis was performed using highly elastic PVA-vessels (Kezlex, Ono & Co., Ltd., Tokyo, Japan). This set-up allows for a realistic scenario and the handling of the vessels during the anastomosis, closely mimicking human conditions during bypass surgery.

We also used 3D printed aneurysms made from soft PVA tubing that could be clipped for training purposes. The 3D printed hollow aneurysms were positioned in the depth of the Sylvian fissure and so microsurgical clip application could be simulated adequately (Videos 1 and 2, https://bit.ly/3AeRgNk). Then the experimental plastic aneurysm could be removed from the site to assess the success of the clipping maneuver (Video 3, https://bit.ly/3AeRgNk). These realistic human skull and brain models are relatively expensive but could be used repeatedly [24, 25]. If the models are integrated into practical teaching and training, they help to establish a realistic microsurgical scenario and are definitively superior to computer-based animations alone [26].

4. In-vivo microsurgical models

4.1 Rabbit model for experimental bifurcation aneurysms

The materials and methods of the experimental aneurysm bifurcation model in rabbits (Figure 17) were described in great detail in previous publications [21, 22, 23]. The developed and previously described animal bifurcation aneurysm model is a perfect and highly realistic vivisection model for education and practical training for microsurgical handling and preparation of cerebral vessels. However, this model should be used as a final education tool in an advanced stage of training. The blood flow, the vessel diameter, the haptic feedback, even the induced vasospasms by manipulating too roughly, and the tension of the vessel walls, all of this can be compared to vascular microsurgery in humans (Figures 17 and 18). Therefore, it is an optimal training tool for all cerebrovascular reconstructive surgical procedures and maintains expert status in bypass surgery [27].

Figure 17.

In-vitro model in a rabbit. An end-to-side anastomosis of both carotid arteries was created using 10/0 sutures to generate an arterial bifurcation.

Figure 18.

The insertion of a venous pouch into the bifurcation finally results in an experimental bifurcation aneurysm, comparable to a 6 mm MCA aneurysm with a broad neck. The in-vivo bifurcation aneurysm model is a perfect training tool for clip application, especially if a plastic tube (schematic drawing) creates a narrow workspace with limited access for the clip applicator. Due to the determined trajectory, the clip occlusion of the aneurysm mimics human conditions. This set-up is a sophisticated in-vivo training model for active teaching and practical training for bypass surgery, as well as aneurysm clipping.

Additionally, this experimental aneurysm model allows an optimal practical training of clip application and is a realistic teaching model for optimizing clip occlusion of cerebral aneurysms. As described in the PVC rat model study conditions, the different sizes of plastic tubes were also integrated into our experimental animal model (Figure 18). The tubes were fixed to a conventional and flexible retractor system and could be removed easily if difficulties arise, especially inadvertent bleeding intraoperatively. The transparent plastic tubes create a narrow and deep surgical approach by restricting the angle of view and determining the trajectory of the clip occlusion of the aneurysm as in real aneurysm surgery. After clipping of the experimental bifurcation aneurysm, the plastic tube was removed and the aneurysm could be inspected easily from all sides and the clip position could be checked adequately.

For repeated training, the clip was removed from the experimental aneurysm and the procedure could be repeated, for example, with a narrower, longer or differently angled plastic tube, creating a completely new situation with a different view and access to the experimental bifurcation aneurysm. This high-end in-vitro animal model is an excellent and realistic set-up for intensive practical training and teaching of aneurysm clipping. However, it takes a great deal of logistical and technical effort to produce such an experimental animal aneurysm.

5. Conclusion

VR and AR are currently established in many areas of medical education and should increasingly become the standard in modern neurosurgical advanced teaching and training as well. Therefore, these tools should also be used regularly for surgical training and further education of young neurosurgeons. With modern navigation systems, diverse software and hardware components are generally available and should consequently be used and strictly integrated into our daily clinical routine. This technology thus forms the basis for highly qualified practical training in skull base surgery. In addition, it facilitates the necessary interdisciplinary cooperation between faculties and offers the opportunity for lifelong learning for all surgically active colleagues in skull base surgery.

In-vitro models, like the AR 3D models, the PVC rat model or the PVA vascular model combined with the realistic brain and craniotomy site, allow for a perfect set-up for the advanced training of microsurgery and microvascular anastomoses. The main advantages of these artificial plastic models are their overall availability, the low price and the lack of a specific OR set-up or instruments, compared to training in in-vivo models. The costs and logistical considerations, as well as the ethical and legal aspects involved in maintaining living animals for education and training, make in-vivo models a relatively impractical tool.

These in-vitro models are easily adaptable to the respective circumstances and allow unhindered practical training under almost realistic operating conditions. The surgical complexity with end-to-end and end-to-side anastomoses could be adapted in models and the success rate is easy to check. Parameters like the time of surgery, the rate of incorrect sutures of the vessel wall and the overall patency rate of the anastomoses can be clearly monitored, as well as the learning curve. Therefore, these in-vitro models form the basis for the first step in basic practical training and are a prerequisite for a successful career in vascular neurosurgery and skull base surgery.

In-vivo models should be the last step of practical education. Like our experimental animal model with the insertion of a venous pouch within the microsurgically created arterial bifurcation represents an advanced training model very close to realistic human conditions. In the first step of this model, microvascular anastomoses are trained and secondly, the resulting bifurcation aneurysm is a perfect training tool to learn clip application. Especially, if a plastic tube is positioned over the surgical field and creates a narrow approach with restricted workspace and limited scope for manipulation for the correct clip occlusion of the aneurysm. Our experimental animal model represents a higher level of surgical vascular expertise and additionally is a perfect model to practice bypass surgery, as well as the appropriate handling of clip application and clip occlusion of cerebral aneurysms.

Conflict of interest

The authors have no financial relationship with the organizations mentioned in the paper. All authors declare that they have no conflict of interest.