AI-driven image-based outcome prediction in COVID-19 version 1.3
Research type
Research Study
Full title
Short And long term outcome prediction using AI-Based image Evaluation in coRonavirus (SABER)
IRAS ID
289801
Contact name
Simon Walsh
Contact email
Sponsor organisation
Imperial College London
Duration of Study in the UK
2 years, 0 months, 1 days
Research summary
Since December 2019, the worldwide spread of COVID-19 has had a significant impact on health and economy since December 2019(1). COVID-19 patients manifest mild symptoms, but ~20% develop more severe disease with ~1-7% of patients requiring intensive care unit (ICU) admission and resource-intensive procedures including mechanical ventilation. Mortality for ICU-admitted patients is ~40% and those who survive are at risk of ongoing pulmonary disability. In addition to the immediate impact of the pandemic, repeated outbreaks in the future as well as long term COVID-19 related pulmonary complications will create a major burden for global public health for months if not years to come.
Successfully responding to the threat of pandemic requires the optimisation of resource allocation to prevent healthcare systems from being overwhelmed and predicting the resulting long-term comorbidity. This is of particular importance since it is believed, although not proven, that a proportion of patients surviving severe SARS-CoV-2 will develop pulmonary fibrosis. Recently, Artificial Intelligence (AI)-based methods for the diagnosis of COVID-19, have shown promise, however, there has been limited work on the role of AI for accurate short and long term prognostication in COVID-19.
We will generate prognostic AI-based algorithms for short/medium and long-term outcome prediction in hospitalised COVID-19 patients using imaging and clinical data. This research will be conducted at Imperial College, London, one of the world's top ranking data and computer science institutions, led by an international consortium of world-class imaging and computer science experts. To achieve this goal, we are building an extensive database of COVID-19 patients using the combined resources of UK, Italian, Spanish and Chinese hospitals working on the frontline.The result will be a rapid, reproducible AI-powered system for patient stratification and resource management in COVID-19 as well as the prediction of long term complications including pulmonary fibrosis and pulmonary arterial hypertension.REC name
N/A
REC reference
N/A