Ryan Armstrong, Dayna Noltie, Roy Eaglesona, and Sandrine de Ribaupierre
Journal of Surgical Simulation 2018; 5: 74 – 86 Published: 28 December 2018
Background: The placement of an external ventricular drain is one of the most commonly performed neurosurgical procedures, and consequently, is an essential skill to be mastered for neurosurgical trainees. In this paper, we describe the development of a simulation environment to train residents on the acquisition of these skills. In addition, the environment was explored as a safe test-bed for the evaluation of novel techniques and factors that influence performance, particularly in regard to anatomic variations that occur clinically. Methods: Patient-specific simulation of freehand ventriculostomies was provided with an integrated mobile augmented reality (AR) image-guidance system. Patient-specific cases represented a progression of burr-hole selection and ventricle targeting tasks. Seven residents and one expert neurosurgeon completed a number of targeting tasks with and without AR guidance. Novel performance metrics are presented and examined. Results: A strong correlation was found between expert-scored accuracy and subject experience (r = 0.93 with 95% confidence interval [0.90–0.95]), but this effect was not present with AR guidance. There was a significant difference in performance between cases classified by experts as simple compared with complex (P 5 0.05). Expert subjective classification of difficulty was a stronger predictor of the challenge of a case than the chosen anatomic measurements. Conclusion: AR guidance showed slight task time improvement, but this was not significant. Objective measures of geometric accuracy show promise, but require further development.