Real-time artificial intelligence-based needle tracking for ultrasound-guided regional anesthesia training: a pilot...

Area of Science:

• Medical Education
• Anesthesiology
• Medical Simulation

Background:

• Ultrasound-guided regional anesthesia (UGRA) training presents challenges in needle visualization for novices.
• A novel system using YOLOv5 technology was developed to provide real-time feedback on needle trajectory.
• This system aims to enhance UGRA training by offering quantitative data and auditory/visual cues.

Purpose of the Study:

• To evaluate if short-term use of a real-time feedback system improves puncture safety in UGRA.
• To determine if needle oscillation amplitude can serve as an objective measure of technical proficiency in UGRA.
• To assess the effectiveness of a YOLOv5-based feedback system in novice UGRA training.

Main Methods:

• Twenty-three novice medical students were randomized into control and YOLOv5 feedback groups.
• Participants performed three in-plane punctures of increasing difficulty on phantoms.
• Outcomes measured included needle tip disappearance frequency and needle oscillation amplitude.

Main Results:

• Needle tip disappearance frequency did not significantly differ between groups.
• Needle oscillation amplitude was significantly smaller in safe punctures compared to unsafe punctures (P=0.003).
• This metric demonstrated preliminary ability to differentiate between safe and unsafe UGRA procedures.

Conclusions:

• Needle oscillation amplitude is a potential objective metric for assessing UGRA proficiency.
• The current feedback system prototype did not show measurable improvements in puncture safety over a short term.
• Usability insights gained will inform the development of improved UGRA trajectory-based feedback systems.