Virtual reality for the training of ear surgery

© 2016 Dr. Patorn PiromchaiIn current residency surgical training programs, residents learn by observing their mentor and practicing under his/her supervision. For patient safety, the mentor needs to ensure that residents have sufficient practice before they perform surgery on patients. Simulation, ranging from low fidelity (e.g. suture practice on rubber models) to high fidelity (e.g. cadaveric temporal bone dissection), is ideal for this purpose and cadaveric bone dissection is the gold standard of simulated training for temporal bone surgery. However, it has a few setbacks. For example, it is limited by the shortage of cadaveric temporal bones and the bones have to be discarded after drilling and cannot be reused. Also, it is difficult to ensure that residents are exposed to a wide range of temporal bones, specifically ones with rare anatomical anomalies. Another concern is that illnesses may be transmitted through contact with diseased tissues and fluids. Virtual reality surgical simulation is an alternative form of high fidelity simulation that offers risk-free, relatively low cost training environments that can be used repeatedly if necessary at the convenience of the user. Reviews have found that practicing surgical procedures on virtual reality simulators improves trainees' operative performance. One major factor contributing to the success of virtual reality surgical simulation systems is their ability to capture clinical variation. Current simulators in Ear Nose and Throat (ENT) surgery mostly use the core training methodology: starting from easier procedures and moving onto more difficult ones. However, this configuration has traditionally been based on the same anatomical model. To the author’s best knowledge, there do not exist virtual reality simulators in the field of ENT that capture anatomical variation as a basis for surgical education. The aim of this thesis was to determine whether exposure to anatomical variation in a virtual reality simulated environment leads to better surgical outcomes, to establish the foundations of a more efficient model for surgical education. To achieve this goal, several investigations were carried out. Firstly, a systematic review and meta-analyses were conducted to examine the role of virtual reality simulation based training in ENT surgery, and to investigate whether it can improve the acquisition of surgical skills. This has revealed that there is currently no literature available in this field to inform whether anatomical variation is important for improving outcomes in simulation based surgical training. Secondly, methods were developed to identify internal anatomical structure variation of temporal bones by examining the bony surface landmarks in an effort to guide the selection of specimens that exhibit differences in anatomy. To this end, correlations between internal and external landmarks were determined and mathematical models developed to predict internal structure variation given combinations of multiple external landmarks using a library of CT scans. It was observed that these correlations and mathematical models were able to predict some anatomical variations with a high level of accuracy. Thirdly, to standardize the evaluation of competence in cochleostomy, an objective assessment tool was developed and validated. Its construct validity and inter-rater reliability were established to be high through evaluations of resident and expert performances on a standardized virtual temporal bone conducted by expert assessors using the proposed tool. Lastly, a randomised controlled trial of ENT residents performing cochleostomy was conducted to determine whether anatomical variation in virtual reality simulation based training improved surgical performance. Participants were randomly allocated to two groups: the experimental group who were exposed to a six different temporal bone anatomies over a period of three weeks and the control group who were trained on one standardized temporal bone over the same period of time. At the end of the training period participants’ performance on two novel temporal bones was assessed. It was observed that ENT residents who were exposed to a range of anatomical variations performed significantly better than those trained on a single anatomical model