Evaluation of AI-based nerve segmentation on ultrasound: relevance of standard metrics in the clinical setting
Authors' conclusion
Variable objective segmentation metric scores correspond to excellent performance on clinically oriented assessment and lack the context provided by subjective expert evaluations.
Study context
- Population
- Ultrasound cine loops of interscalene-supraclavicular brachial plexus, femoral and popliteal-level sciatic block regions rated excellent (5) by expert anaesthesiologists
- Participants
- 173 still image frames
- Comparison
- Objective pixel-overlap metrics (IoU, Dice) vs subjective expert rating of AI nerve segmentation (cNerve™)
- Outcome
- Intersection over Union and Dice similarity coefficient of nerve segmentation on frames from sequences rated 5
Performance
As reported in the abstract
Objective segmentation metrics in expert-rated excellent sequences
Median Intersection over Union for nerves
0.49
Median Dice similarity coefficient
0.65
Original publication
Evaluation of AI-based nerve segmentation on ultrasound: relevance of standard metrics in the clinical setting. Br J Anaesth. Published 26 February 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →