Background and Aims Ultrasound-guided regional anaesthesia (UGRA) involves the acquisition and interpretation of ultrasound images. This pilot study explores the utility of an artificial intelligence (AI) device, which highlights key sono-anatomical structures with a real-time colour overlay.

Methods With ethical approval, 30 anaesthetists collected 240 ultrasound scans over nine block regions. Half (n=15) of participants were experts in UGRA, 120 scans were performed with the AI device (60 expert & 60 non-expert – under supervision). Half were performed in a subject of BMI <30 kg/m² (half BMI ≥30 kg/m²). Participants completed structured questionnaires on benefits of the device and potential risks of unnoticed incorrect highlighting.

Results Data are summarised in tables 1–3. Feedback was positive in 165/360 (45.8%) instances for non-experts, and 102/300 (34.0%) for experts (Yates’s χ² = 9.03, p <0.01). Feedback was neutral in 186/360 (51.7%) and 178/300 (59.3%) for non-experts and experts respectively (Yates’s χ² = 3.58, p = 0.06). Negative feedback was noted in 9/360 (2.5%) and 20/300 (6.7%) for non-experts and experts (Yates’s χ² = 5.81, p = 0.02). Experts reported no increased risk in 394/406 (97.05%) and that unnoticed incorrect device highlighting carried potential increased risk in 12/406 (3.0%).

• Download figure
• Open in new tab
• Download powerpoint

Abstract 141 Figure 1

Conclusions Non-experts viewed the device most positively, particularly in support of learning, training and identifying sono-anatomical structures. Perceived increased risk was infrequent, but some potential complications considered may be clinically important (e.g., nerve injury, local anaesthetic systemic toxicity). This pilot study provides evidence that such technology may support (rather than replace) clinician judgement for UGRA procedures.