Artificial Intelligence for Anatomical Structure Identification on Ultrasound in Regional Anaesthesia: A Scoping Review Protocol

Background Ultrasound is the most common form of guidance for regional anaesthesia. There is increasing interest in developing supporting technology, particularly in the form of artificial intelligence (AI), to aid in the acquisition and interpretation of optimal ultrasound views for these procedures. However, this is a broad field, with academia, clinical practice, and industry all providing disparate contributions. We will undertake a scoping review of publicly-available data, to assess methods of evaluation for accuracy and utility of such systems. Methods We will perform searches in multiple databases, including ACM Digital Library, CINAHL, EMBASE, IEEE Explore, and OVID MEDLINE. We will search the International Committee of Medical Journal Editors approved clinical trial registries and the World Health Organisation (WHO) clinical trials registry for studies registered in this field. Grey literature will be searched through the online library of doctoral theses ( http://ethos.bl.uk/Home.do ), regulatory authority registries and competent authority websites of North America and the UK, the websites of international learned societies in regional anaesthesia, and material from commercial organisations with products in the field. The primary goal is to summarise the approaches used to evaluate accuracy and utility of these devices. A secondary goal is to assess the standardisation of reporting in this field, with particular reference to whether reporting guidelines have been followed. Discussion To the best of our knowledge, this is the first scoping review of this type. Synthesis of the available evidence will enable us to make recommendations on standardised evaluation approaches of assessment, to allow robust and relevant evaluation which can be compared to similar evaluations of other devices.