Data annotation and its evaluation in artificial intelligence-based anatomy recognition for ultrasound-guided regional anesthesia: a clinical perspective - Summary - MDSpire

Data annotation and its evaluation in artificial intelligence-based anatomy recognition for ultrasound-guided regional anesthesia: a clinical perspective

  • By

  • Bernard V. Delvaux

  • Yoann Elmaleh

  • Alwin Chuan

  • Alex T. Sia

  • Rajnish K. Gupta

  • Kristopher M. Schroeder

  • Karim Guessous

  • James S. Bowness

  • May 29, 2026

  • 0 min

Share

Objective:

To address the influence of clinical endpoints on AI-based anatomy recognition in ultrasound-guided regional anesthesia (UGRA).

Key Findings:
  • Current AI-based guidance for UGRA primarily improves procedural orientation but has limited safety gains due to imperfect contour accuracy.
  • Data quality is crucial for robust model development, and strategies exist to manage poor-quality images.
  • Defining precise segmentation rules and managing uncertainty in annotations are essential for improving model training.
  • Combining annotations from multiple experts can reduce subjective biases and improve the reliability of ground-truth data.
Interpretation:

Limitations:
  • Insufficient standardization of training data and evaluation methods.
  • Lack of consensus on annotation strategies and expert aggregation methods.
Conclusion:

Original Source(s)

Data annotation and its evaluation in artificial intelligence-based anatomy recognition for ultrasound-guided regional anesthesia: a clinical perspective

  • by Bernard V. Delvaux, Yoann Elmaleh, Alwin Chuan, Alex T. Sia, Rajnish K. Gupta, Kristopher M. Schroeder, Karim Guessous, James S. Bowness

  • May 29, 2026

Related Content

Abdominal Aortic Aneurysm: When Endovascular Repair Is the Right Option

In this procedural case review, vascular surgeon Dr. Samuel Steerman demonstrates endovascular aneurysm repair (EVAR) for a patient in their 70s with an abdominal aortic aneurysm.

  • May 27, 2026

  • 1 min
DPEA-Net: a clinically-oriented lightweight 3D CNN for glioma segmentation in multiparametric MRI

Images in Vascular Medicine: Multimodal imaging of adventitial cystic disease of the superficial femoral artery