Artificial Intelligence-assisted Diagnosis and Prognostication in COVID-19 Using Electrocardiograms

Coronavirus Clinical Trial

— AI-COV-19  

Official title:

Artificial Intelligence-assisted Diagnosis and Prognostication in COVID-19 Using Electrocardiograms and Imaging

Clinical Trial Details — Status: Recruiting

Administrative data

• NCT number: NCT04510441
• Other study ID #: 20HH5967
• Status: Recruiting
• Start date: May 26, 2020
• Est. completion date: May 1, 2022

Study information

• Verified date: August 2021
• Source: Imperial College London
• Contact: n/a
• Is FDA regulated: No
• Study type: Observational

Clinical Trial Summary

Coronavirus Disease 2019 (COVID-19) has been widespread worldwide since December 2019. It is highly contagious, and severe cases can lead to acute respiratory distress or multiple organ failure. On 11 March 2020, the WHO made the assessment that COVID-19 can be characterised as a pandemic. With the development of machine learning, deep learning based artificial intelligence (AI) technology has demonstrated tremendous success in the field of medical data analysis due to its capacity of extracting rich features from imaging and complex clinical datasets. In this study, we aim to use clinical data collected as part of routine clinical care (heart tracings, X-rays and CT scans) to train artificial intelligence and machine learning algorithms, to accurately predict the course of disease in patients with Covid-19 infection, using these datasets.

Description:

Coronavirus Disease 2019 (COVID-19) has been widespread worldwide since December 2019. It is highly contagious, and severe cases can lead to acute respiratory distress or multiple organ failure and ultimately death. The disease can be confirmed by using the reverse-transcription polymerase chain reaction (RT-PCR) test. ECGs, Chest x-rays and CT scans are rich sources of data that provide insight to disease that otherwise would not be available.

Knowing who to admit to the hospital or intensive care saves lives as it helps to mitigate resource shortages. Novel Artificial Intelligence tools such as Deep learning will allow a complex assessment of the Imaging and clinical data that could potentially help clinicians to make a faster and more accurate diagnosis, better triage patients and assess treatment response and ultimately better prediction of outcome. Our group has significant experience implementing machine learning algorithms on vast quantities of ECGs, such as from the UK Biobank, and propose to extend our techniques to data from patients with Covid-19.

This is a retrospective data study on patients with suspicious and confirmed COVID-19.

The study aims to recruit up to 2000 patients who will be referred to have ECGs, chest X-rays or CT scans at Chelsea and Westminster Hospital NHS Foundation Trust, Imperial College Healthcare NHS Trust and London North West London University Healthcare NHS Trust.

To be included in this study, the patient must:

• have ECGs, Chest x-ray and/or chest CT imaging (with or without contrast)

• laboratory Covid-19 virus nucleic acid test (RTPCR assay with throat swab samples) or clinical suspicion for Covid19 infection

• be aged >18 years Patients with suboptimal ECGs, chest radiograph and CT studies due to artefacts will be excluded. Patients will also be excluded if the time-interval between ECGs, chest CT and the RT-PCR assay was longer than 7 days.

This study received HRA and Health and Care Research Wales (HCRW) approval on 18 May 2020 following review by Research Ethics Committee at a meeting held on 13 May 2020(Protocol number: 20HH5967; REC reference: 20/HRA/2467).

Recruitment information / eligibility

• Status: Recruiting
• Enrollment: 2000
• Est. completion date: May 1, 2022
• Est. primary completion date: May 1, 2022
• Accepts healthy volunteers: No
• Gender: All
• Age group: 18 Years and older

Eligibility

Inclusion Criteria:

• have ECGs, Chest x-ray and/or chest CT imaging (with or without contrast)

• positive laboratory Covid-19 virus nucleic acid test (RTPCR assay with throat swab samples) or clinical suspicion for Covid-19 infection

• be aged >18 years

Exclusion Criteria:

• Suboptimal ECGs, chest radiographs or CT studies for deep learning methods due to artefacts including severe

• motion artefacts which causes blurring of the contours of or significant artefacts due to metallic prosthesis which causes image degradation

• Time-interval between ECGs, chest CT and the RT-PCR assay was longer than 7 days

Related Conditions & MeSH terms

• Coronavirus
• Coronavirus Infections

Intervention

Other:
• Nil intervention
Nil intervention; retrospective cohort study

Locations

Country	Name	City	State
United Kingdom	Chelsea and Westminster Hospital NHS Foundation Trust	London
United Kingdom	Imperial College London (Hammersmith campus)	London
United Kingdom	London North West University Healthcare NHS Trust	London
United Kingdom	St Mary's Hospital	London

Sponsors (4)

Lead Sponsor	Collaborator
Imperial College London	Chelsea and Westminster Hospital NHS Foundation Trust, Imperial College Healthcare NHS Trust, London North West Healthcare NHS Trust

Country where clinical trial is conducted

United Kingdom, 

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	Accuracy of machine learning to be able to predict outcome of coronavirus (COVID-19) infection	Accuracy with which computer based analysis (machine learning) can diagnose and/or prognosticate Covid-19 Number of Participants With COVID19 who died or survived following hospital admission	At the end of data analyses, approximately 1 year
Primary	Accuracy of machine learning to be able to predict prognosis of coronavirus (COVID-19) infection	Number of participants who required invasive vs non-invasive ventilation vs ward-based care vs died	At the end of data analyses, approximately 1 year
Secondary	Accuracy of machine learning to be able to predict cardiac involvement of coronavirus (COVID-19) infection	Number of participants who had COVID19-related heart problems.	At the end of data analyses, approximately 1 year
Secondary	Accuracy of machine learning vs human assessment to diagnose coronavirus (COVID-19) infection	Number of participants that can be identified as having COVID19 using machine learning vs human or other clinical test or assessment	At the end of data analyses, approximately 1 year