Artificial Intelligence Assistance Does Not Provide Immediate Benefit for Right Upper Quadrant Window Acquisition.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Emergency Medicine Ultrasound

Background:

• Point-of-care ultrasound (POCUS) increasingly features AI-assisted acquisition.
• The utility of AI for the Focused Assessment with Sonography for Trauma (FAST) exam is not well-established.
• This study investigates AI's impact on FAST exam acquisition for emergency medicine trainees.

Purpose of the Study:

• To evaluate the immediate effect of AI auto-labeling and auto-grading on FAST exam acquisition time and image quality.
• To compare AI's impact on novice versus experienced emergency medicine (EM) physician trainees.
• To determine if AI assistance aids in acquiring the right upper quadrant (RUQ) FAST exam window.

Main Methods:

• A repeated-measures study involving 14 novice and 10 experienced EM trainees.
• Trainees acquired RUQ FAST exam windows with and without AI assistance on standardized patients.
• Acquisition time was measured in seconds, and image quality was assessed based on structure visibility, plane correctness, and probe orientation.

Main Results:

• AI assistance significantly increased acquisition time for both novice (85 vs 53 seconds) and experienced trainees (44 vs 28 seconds).
• Experienced trainees were more likely to meet image quality criteria than novices.
• No significant difference in image quality was observed with or without AI assistance within either group.

Conclusions:

• AI auto-labeling and auto-grading increased RUQ FAST exam acquisition time for physician trainees.
• AI assistance did not immediately improve image quality for either novice or experienced trainees.
• Current AI features are unlikely to assist trainees in acquiring RUQ FAST exam windows during single attempts.