While endoscopic procedures remain a gold standard diagnostic test for many cancers, current endoscopes require a combination of elaborate
choreography of psychomotor movements. Due to the non intuitive user interface of standard flexible endoscopes, novices usually have a steep learning curve. Additionally, the success rate and quality of endoscopic procedures to ensure that all abnormalities are found still highly depends on the operator. The performance rate of these examination is very variable, with a high rate of failure (11,6%) to diagnose cancer. Ultimately, for a patient an endos-copy can be very uncomfortable due to the difficult maneuvering of the endoscope, which can also lead to perforations in the gastrointestinal tract.
Every year in the US, 51.1 million gastrointestinal endoscopies are performed. An increasing percentage of these procedures includes a surgical intervention to, for example, remove a tumor. These interventions increase the complexity of the procedure and require even more advanced psychomotor skills and therefore more experience.
The invention from the Medical University Vienna combines a mixed-reality headset (like the Holo-lens) that enables intuitive steering in combination with a flexible endoscope to automate the detection of abnormalities while providing optimal visualization. The operators head movements are carefully tracked by the mixed-reality headset and are then converted into precise movements of the tip of the endoscope. An (AI) algorithm will be trained to support the detection of abnormalities and provide real time information to the operator. The system also allows other medical experts to join virtually or to check the intervention therefore increasing patient outcomes and enable faster training.
The first prototype of the headset and its endoscopic controls are being developed. The next steps involve (pre-)clinical validation and training of the algorithm towards an AI solution. Two patents have been filed (both pending with positive examiners response).