[Evaluation of a training system for middle ear surgery with optoelectric detection]

This work presents a new training concept for surgery of the temporal bone. It is based on a model of gypsum plastic with optoelectric detection of risk structures. A prototypical evaluation is given.

The training models are based on high-resolution computed tomographic data of a human skull. The resulting data set was printed by a three-dimensional (3D) printer. A 3D phantom is created from gypsum powder and a bonding agent. Risks structures are the facial nerve, semicircular canal, cochlea, ossicular chain, sigmoid sinus, dura, and internal carotid artery. An electrically conductive metal (Wood's metal) and a fiber-optic cable were used as detection materials for the risk structures. For evaluating the training system, a study was done with eight inexperienced and eight experienced ear surgeons. They were asked to perform temporal bone surgery using two identical training models (group A). In group B, the same surgeons underwent surgical training with human cadavers. In the case of injuries, the number, point in time, degree (facial nerve), and injured structure were documented during the training on the model. In addition, the total time needed was noted.

The training systems could be used in all cases. Evaluation of the anatomic accuracy of the models showed results that were between 49.5% and 90% agreement with the anatomic origin. Error detection was evaluated with values between 79% and 100% agreement with the perception of an experienced surgeon. The operating setting was estimated to be better than the previous"gold standard." The possibility of completely replacing the previous training method, which uses cadavers, with the examined training model was affirmed.

This study shows that the examined system fulfills the conditions for a new training concept for temporal bone surgery. The system connects the preliminary work with printed and sintered models with the possibilities of microsystem engineering. In addition, the model's digital database permits a complete virtual representation of the model with appropriate further applications ("look behind the wall," virtual endoscopy).