US20260031221
2026-01-29
Physics
G16H40/20
The disclosed invention pertains to a system and method for assessing surgical skills using computer systems and machine learning. The system processes raw data from surgical procedures, extracting features to train a machine learning classifier. This classifier is then used to evaluate the performance of other surgeons. The system can handle data from both robotic and non-robotic surgical environments, offering a comprehensive approach to skill assessment without relying heavily on human reviewers.
Traditional methods of assessing surgical skills face numerous challenges, such as the difficulty of isolating specific skills from overall surgical outcomes and the subjective nature of human evaluations. Expert surgeons are often too busy to review videos, and there is an imbalance between the number of expert and novice surgeons, complicating data analysis. The invention addresses these issues by providing an automated, scalable system that reduces reliance on expert reviews and accounts for data asymmetries.
In a non-robotic surgical theater, the system involves the use of various tools like visualization devices and mechanical instruments. Visualization tools provide interior views of the patient, which can be displayed and recorded for analysis. The surgical process consists of multiple tasks, each requiring specific actions and possibly different tools. The system captures and analyzes data from these tasks to assess surgical skills, even accommodating depth data for enhanced analysis.
Robotic surgical theaters use systems like the da Vinci™ surgical system, where operations are performed remotely via a console. This setup allows for the collection of extensive data, including tool movements and operator inputs. The data from these robotic systems are recorded and analyzed to assess surgical skills, providing a richer dataset compared to non-robotic environments. The system facilitates seamless task transitions and records various metrics for comprehensive performance evaluation.
The invention integrates machine learning to analyze surgical data, utilizing various model architectures and methods. While the document does not exhaustively cover all machine learning models, it highlights the flexibility and adaptability of the system to employ different architectures. The machine learning models are trained to recognize patterns in surgical data, providing objective, data-driven assessments of surgical skills, thus enhancing the feedback provided to surgeons.