← All decks
Anaesthesiology
Artificial Intelligence for Ultrasound in Anaesthesia, Critical Care and Pain Medicine
Built from Practical Ultrasound in Anesthesia
What’s inside
5 sections · 58 slides
Overview
- Topics covered
Artificial Intelligence
Machine learning, deep learning and convolutional neural networks
- Definition and reach of artificial intelligence
- Ultrasound imaging and operator skill
- Machine learning and deep learning
- Inside artificial intelligence
- Key Definitions in AI
- Key Definitions in AI (continued)
- Key Definitions in AI (continued)
- Training a convolutional neural network
- Convolutional neural network learning and prediction
- Supervised versus unsupervised learning
- Two ways to train machine learning
AI Across Point-of-Care Ultrasound
Cardiac, respiratory, gastric, neuraxial and vascular applications
- AI applications in point-of-care ultrasound
- Point-of-care ultrasound themes applying AI
- Cardiac ultrasound in anaesthesia and critical care
- AI steering novices in cardiac ultrasound
- Automated ejection fraction measurement
- Single-click ejection fraction and its limits
- AI reading transoesophageal echo
- Respiratory ultrasound and AI
- Machine learning for COVID-19 lung ultrasound
- Gastric ultrasound for airway safety
- AI for neuraxial ultrasound
- The Accuro handheld device
- AI-assisted vascular access
AI in Regional Anaesthesia
The ScanNav colour overlay case study, its evidence and training
- Ultrasound-guided regional anaesthesia
- Skills required for ultrasound scanning
- AI support for regional anaesthesia
- The Plan A Blocks and the anatomy challenge
- The benefit of colour
- The ScanNav colour overlay
- AI colour overlay on ultrasound for nerve blocks
- Inside the ScanNav CNN
- Convolutional neural network used in ScanNav PNB
- Similar nerve-block AI tools
- 93.5%
- Expert judgement on the colour overlay
- Clinical impact of the colour overlay
- AI in training and assessment
- Beyond training: fatigue, distraction and error
Adopting AI in Clinical Practice
Evidence, workflow, limitations and accountability
- Sources of AI device evidence
- The workflow gap
- Evidence for patient benefit
- Toward a shared assessment framework
- Pitfalls in AI development
- NICE guidance on medical technologies
- Regulation and communication barriers
- Stakeholder teams required for AI in healthcare
- The five stakeholder groups
- Accountability with AI-assisted decisions
- Conclusion
- References
- References (continued)
- References (continued)
- References (continued)
- References (continued)
- References (continued)
- Practical Ultrasound in Anesthesia for Critical Care and Pain Management