Case G: Urgent Care Referral Using Lung Ultrasound Imaging of Lungs

Work Packages: WP2, WP3

Background: The urgent need in industrialised countries is to reduce the cost of healthcare by making commonly acquired imaging examinations possible in primary healthcare. However, the type of equipment and expertise available for image acquisition and interpretation at local healthcare facilities are frequently insufficient to accurately identify urgent conditions such as heart-failure, pneumonia or pneumothorax. This can lead to over or under-referral of patients to a hospital setting, neither of which are desirable. The problem is exacerbated in low-resource settings where a large portion of the population lives rurally and for whom hospital access is frequently very logistically difficult. In both settings, there is an urgent need for more effective triage at local healthcare facilities, but a severe shortage of medical doctors and limited resources for equipment.

Objective: In this project we will develop AI systems to detect urgent conditions using low-cost medical images acquired with X-ray and US. Specifically, we will develop AI technology to support novice users to acquire diagnostic lung US images and to interpret chest X-ray and/or US combined with clinical symptoms and vital signs. These systems will be designed to be operated by local staff with minimal training and to accurately differentiate between patients that need referral to hospital and those whose care can be managed in their home.