Computer assisted skull base surgery: a contemporary review

Virtual planning

CT is a critical CAS modality because it clearly defines bony margins and tumor infiltration. It is the primary modality for virtual surgical planning, 3D printing, and surgical navigation [25, 26]. CT imaging can be divided into two major categories: Fan beam CT which are often called “medical CT scanners,” and cone beam CT which are often called “in office” CT scanners [27]. Both modalities can be used for CAS applications [28]. Their advantages and disadvantages are discussed below (see intraoperative verification). Magnetic resonance imaging (MRI) has superior soft tissue resolution when compared to CT. However, poor bony resolution has made planning, navigation, and 3D printing more challenging. However, the pediatric skull base literature has recently described expansion of MRI for fabrication of 3D models and cutting guides, using the negative “black bone” sequences [29], [30], [31]. Refinement of this technique could significantly reduce radiation exposure while generating the necessary information for CAS applications.

Virtual planning tools include 1) planning software, 2) virtual environments, and 3) 3D models. Each of these will be discussed below:

Virtual Planning software: Virtual planning software converts DICOM (“DICOM” – Digital Imaging and Communication in Medicine) data, the common language used by medical devices [30], to a propriety format that generates a 3D patient representation. Several examples include Mimics (Materialise – Leuven, Belgium), Elements (Brainlab – Berlin Germany), Invivo6 (Anatomage, Santa Clara, CA, USA), and Dolphin (Dolphin Imaging & Management Solutions – Chatsworth, USA). Alternatively, open source software is available, but surgeons need to consider the clinical implications when using “non-CE or FDA” approved software tools. Planning software provides a myriad of functions including soft tissue/bony segmentation (Figure 2), measurement tools (distances/angles/volumes), object mirroring, and fusion of data sets (overlap of two CT scans or MRI and CT).

Figure 2:

Segmentation of an invasive squamous cell carcinoma (brown) of the left maxillary sinus and adjacent structures (red – maxilla, green – orbit, pink – zygomatic bone, purple – nasal bone and contralateral zygomatic bone).

Planning may be performed using a surgeon-based model which requires the surgeon to purchase, learn, and independently utilize the software; or an industry-based model where surgeons and clinical engineers meet in an online setting, incurring only the time commitment to the surgeon. Cost recovery from industry in this second model varies and may include fee for service or incorporation of engineer fees into the cost of implant production. Once planning is complete, data can be exported in proprietary or non-proprietary formats for use in virtual environments (see augmented or virtual reality below), 3D printing, implant fabrication, as well as intraoperative navigation.

Virtual Environments: Virtual environments allow surgeons, trainees, as well as patients to visualize surgical anatomy in true 3-dimensional immersive space. These virtual environments can be described on a “reality spectrum” ranging between the “real environment” and a completely “virtual environment” (Figure 3). These tools can be divided into three basic types: virtual reality , augmented reality , and mixed reality . Virtual reality tools completely separate the surgeon from his surrounding environment (Figure 4). This provides the user an extremely crisp and accurate view, however it makes interaction with other users more challenging and can lead to motion sickness [32, 33]. Augmented reality tools overlay information (generally data) onto the physical environment (Figure 5). Mixed reality tools allow the user to visualize their environment, superimposing virtual objects into the environment that can be manipulated (Figure 6). Objects appear more translucent (when compared to virtual reality), however users can easily interact with their environment, other users, and potentially the surgical field at some point in the future [34, 35]. Some examples include Magic Leap One Mixed Reality Viewer (Brainlab – Berlin Germany), Immersive Touch (Zimmer-Biomet – Warsaw, USA), and Precision VR (Surgical Theater – Cleveland, USA). Some tools include planning software which can be utilized by the surgeon (Magic Leap and Immersive Touch) while others (Precision VR) require an engineer to perform the surgical planning.

Figure 3:

Reality spectrum: 1) virtual reality completely separates the user from the real world 2) augmented reality overlays information (data) onto the physical environment, 3) mixed reality allows the users to see their environment and places objects that can be manipulated within that environment.

Figure 4:

Virtual reality environments completely separate the user from their physical environment while allowing the user to manipulate virtual objects.

Figure 5:

Augmented reality overlays information (often data) onto the environment as shown in this “heads up display”.

Figure 6:

Mixed realith tools allow the user to visualize their environment, superimposing virtual objects into the environment that can be manipulated.

3D Models: 3D printing allows the surgeon to fabricate physical objects that can be used for surgical planning and pre-surgical simulation. This may include osteotomies, bony advancements/setbacks, resections, as well as trajectory planning and patient specific hardware bending.

Surgical execution

Surgical execution tools allow the surgeon to perform surgical steps with improved accuracy and efficiency. Some examples include 3D Models, cutting/drill guides, surgical implants.

3D Models: Sterilization of 3D models for intraoperative use requires selection of materials that will withstand the autoclave process or selection of other low temperature sterilization processes such as Sterrad (Advanced Sterilization Products, Irvine, Ca, USA). Once sterilized, models can be used for intraoperative hardware bending (Figure 7), verification of anatomic landmarks, and resident education [18, 38].

Figure 7:

3D model used for template bending in the preoperative setting.

Cutting/Drill Guides: Intraoperative cutting and drill guides can be printed (or milled) for fibula free flap reconstruction (Figure 8), orthognathic surgery, and dental implants. Most cutting and drill guides are fabricated by industry with CE/FDA certification. They are very expensive, which has limited the total number of end users worldwide. Some institutions fabricate cutting and drill guides in house at a significantly reduced cost, however this incurs other potential costs including 3D printers/software, medicolegal risks, and surgeon time. Cutting and drill guides are particularly useful in planned (secondary) reconstructions, because there are clearly defined margins and the implant size is predictable. However, use in primary reconstruction is limited due to the unpredictablity of tumor resection margins.

Figure 8:

3D printed cutting guide for mandible reconstruction.

Surgical implants: The use of 3D printing and milling allows for fabrication of patient specific implants (PSI). Some examples include orbital (Figure 9), TMJ, mandible (Figure 10), and cranial reconstruction. Due to strict manufacturing guidelines, the need for long term structural stability, and medicolegal risk, industrial partners generally fabricate these implants and few if any hospitals are entering into this arena. Potential drawbacks, especially in large implants (Figure 9) are imaging artifacts and implant exposure after radiation therapy.

Figure 9:

Commercially fabricated patient specific implant for orbito-midface reconstruction.

Figure 10:

Patient specific implant used to shape fibula in situ for mandibular reconstruction.

Intraoperative verification

After completion of many maxillofacial reconstructive procedures, visualization of the entire repair is challenging (ex. Orbital reconstruction, complex mandible reconstruction, zygomaticomaxillary reconstruction, etc.). Intraoperative navigation and CT can be efficacious to verify the accuracy of the reconstruction [41].

Intra-operative navigation: DICOM as well as proprietary data can be imported into surgical navigation systems. Once the data is imported, accurate “registration” of the patient is critical. The surgeon must provide the navigation system with precise landmarks on the patient’s skin, bone, oral appliance, or rigidly fixated hardware. Ideally, anatomic landmarks should be non-coplanar and non-collinear, collecting a sphere of points with the area of greatest interest at the center. This will allow the navigation system to accurately localize the patient in space, with the area of greatest precision being the area of interest (i.e. tumor, injury, anatomic deformity). The accuracy of most navigation systems ranges from 1 to 2 mm [42]. Registration modalities include electromechanical, electromagnetic, and optical [42]. Optical registration has several advantages for skull base reconstruction including: (1) it allows for free mobility around the patient without wires and (2) the instruments have no metallic distortion which can result in inaccuracies with electromagnetic systems [22, 43].

Once patient registration is complete, navigated instruments can be used to localize patient anatomy in difficult to visualize areas. It should be noted that all anatomic information provided by intraoperative navigation is “inferred.” In other words, it is only as precise as the imported data and accuracy of the registration process. Therefore, the information that is provided to the surgeon should be used to confirm anatomic information, not to guide surgical dissection.

Another advantage of navigation in oncologic surgery is the ability to place virtual “markers” within the 3D data set to identify the resection margins [22]. This information can then be used in combination with final pathology and radiation therapy to determine the optimal strategy for post-operative radiotherapy treatment.

Intraoperative CT : Unlike intraoperative navigation, CT incurs radiation exposure. Both fan beam and cone beam CT offer excellent intra-operative visualization, providing “real time” view of patient anatomy. However, there are some important differences to consider. Fan beam scanners provide excellent bone/soft tissue resolution and can be utilized with intravascular contrast material. However, they are much more expensive and incur a significantly higher radiation dose. Cone beam scanners have lower radiation dose, adequate bony detail, narrow slice thickness, and can be purchase for a much lower cost [27].