The design of an anthropomorphic head phantom for neurosurgical planning, education and training

Neurosurgery is technically challenging because of the inaccessibility and sensitivity of the brain. The risk of complications for the patient is high and simulation-based training is therefore preferred. However, surgical training in cadavers and virtual reality is not satisfactory. A phantom is a model of organs or tissues consisting of tissue mimicking materials. Anthropomorphic phantoms could solve most of the issues related to VR or cadaveric models and could be used for neurosurgical planning, education and training purposes. For this master thesis, it was chosen to focus on one specific type of surgery, namely endoscopic endonasal neurosurgery. The goal of this study was to find out if an anthropomorphic head phantom could improve endoscopic endonasal neurosurgical planning, education and training. This research project is partly initiated by Philips, as they benefit from a head phantom for two different research projects concerning neuronavigation systems. By reviewing relevant literature and attending an endoscopic endonasal surgery, important anatomy is selected for the design of the head phantom. Also, different phantom characteristics described in literature were reviewed in order to select the correct materials and methods for the production of the head phantom. A head phantom is created from a high-quality CT scan of an anonymous female adult. 3D slicer is used to segment a patient-specific model consisting of three parts, namely the skull, brain and face. In order to segment all structures correctly, a radiologist helped with inspecting the scan. The head phantom is produced in three parts: the skull, the brain and the face. The skull phantom is 3D printed using an Ultimaker 3. It is made of PLA with calcium for the bone parts and MP Flex for the cartilage parts. A mould is 3D printed for the brain phantom using an Ultimaker 2. The brain phantom itself is made of water and coolant with 6 wt% PVA and 1 wt% barium sulphate. The face is 3D printed using an Ultimaker 2 and is made of PLA and MP Flex. The three phantom parts can be fitted together easily and securely. Each part is scanned using an XperCT scan at Philips. By measuring the Hounsfield units, it has been shown that there is significant contrast difference between the skull and the brain phantom. An indention test has shown that the brain phantom does not exactly match the mechanical properties of real brain tissue. The produced skull and the brain phantom have been compared with the original 3D models and the error in size difference between the two scans ranges from 0-1 mm. Neurosurgeons and residents inspected the head phantom, after which they filled in a questionnaire. The questionnaire has been used to evaluate if the anatomical structures are represented correctly and if the head phantom can improve neurosurgical planning, education and training. The anatomic accuracy and appearance of the nasal cavities, sphenoidal sinus, sella turcica and pituitary gland are found realistic by the surgeons. The results of the questionnaire show that surgeons found the use of a patient-specific head phantom for surgical planning purposes useful. The overall impression was that a patient-specific head phantom will not only improve surgical planning for endoscopic endonasal neurosurgery but would also be desirable other types of surgery. The qualitative study also suggests that the head phantom may be valuable for neurosurgical education and training. Surgeons from different hospitals would like to use the head phantom for a training day, to let residents practise the use of an endoscope. An augmented reality neuronavigation system is tested with the head phantom at Philips in Best. The tests showed that the head phantom can also be used for research purposes and testing of neuronavigation systems.