Prediction of temperature elevation in rotary ultrasonic bone drilling using machine learning models: An in-vitro experimental study.

Bone drilling is frequently used during orthopaedic surgeries to treat the fractured part of the bone. A major concern for surgeons is the increase in temperature during real-time orthopaedic bone drilling. The temperature elevation at the bone-tool interface may cause permanent death of regenerative soft tissues and cause thermal osteonecrosis. A robust predictive machine-learning model is suggested in this in-vitro research for monitoring temperature rise during surgery. The objective of the present work is to introduce different machine learning algorithms for predicting temperature elevations in rotary ultrasonic bone drilling. Different machine-learning models were compared with the standard response surface methodology. The performance and accuracy of different predictive models were compared at different error metrics. It was witnessed that support vector machines performed the best for predicting the change in temperature in comparison to other predictive models. Moreover, the error metrics for statistical response surface methodology analysis were comparatively higher than the machine learning algorithms. By using machine learning models, it is possible to predict temperature rise during bone drilling.