Advancing Minimally Invasive Robotics through the Development of an AI-Driven Brain Surgery Platform with Usability Evaluation

The application of AI in the field of medicine, particularly the field of neurosurgery, has shown outstanding potential. In this paper, we propose an artificial intelligence system aimed at helping neurosurgeons properly detect and localize brain tumors using advanced techniques such as deep learning. The algorithms were taught using a comprehensive brain MRI dataset which included the segmentation and classification tasks of MRI obtained images. In testing the system on an independent set of brain MRI scans, the audio-visual feedback revealed a mean accuracy of 87% in the form of Dice coefficient. According to a User Experience assessment, there was also a significant improvement in accuracy, effectiveness, and overall workload. Also, the system was found to be flexible in many surgical situations where it was able to customize instructions for the users. This demonstrates the potential of AI, especially with regards to worsening the quality of the neurosurgical procedures and the effects on the patients’ welfare. Our next objectives will be to merge the system with robotic arms to make changes to the traditional approach of surgical procedures, making them less invasive.