We've updated our Privacy Policy to make it clearer how we use your personal data. We use cookies to provide you with a better experience. You can read our Cookie Policy here.


Combining AI and Point-of-Care Diagnostics for Rapid COVID-19 Screening

Combining AI and Point-of-Care Diagnostics for Rapid COVID-19 Screening content piece image
Credit: Pixabay
Listen with
Register for free to listen to this article
Thank you. Listen to this article using the player above.

Want to listen to this article for FREE?

Complete the form below to unlock access to ALL audio articles.

Read time: 3 minutes

Rapid and accurate screening can play a vital role in reducing the spread of COVID-19 and ensuring that patients receive timely and appropriate treatment upon arrival at hospital. In a study recently published in The Lancet, CURIAL, an AI-based screening system developed by researchers at Oxford University, was shown to effectively rule out COVID-19 within an hour of patients arriving at a hospital emergency department.

Technology Networks
interviewed Dr Andrew Soltan, NIHR Academic Clinical Fellow at the John Radcliffe Hospital, and Yossi Pollak, co-founder and CEO of Sight Diagnostics, to learn more about the algorithm and how it is being used alongside Sight OLO, Sight Diagnostics’ machine vision-based blood analyzer, to rapidly triage patients and support infection control within hospitals.

Anna MacDonald (AM): How are patients arriving at
hospital emergency departments currently screened for COVID-19? What are some of the limitations of this approach?

Dr Andrew Soltan (AS):
In acute care settings, such as hospital emergency departments, patients are currently being screened using the Polymerase Chain Reaction (PCR) test performed on a sample taken by a nose and throat swab. PCR results are typically available in around 24 hours, however, the workflow in an emergency department calls for rapid decision making, demanding quickly available and reliable test results. Recognizing the demand for a shorter time-to-result, lateral flow rapid antigen testing has been adopted by many emergency departments across the country. However, the ability of the lateral flow tests to confidently exclude infection in a symptomatic population is thought to be limited, with a recognized false negative rate posing challenges for infection control within hospitals.

To reduce spread of COVID-19, while avoiding care delays, there is a need for a test that is able to confidently rule-out the infection within the first minutes-to-hours of a patient coming to the emergency department. This reduces the probability of COVID-19 being unintentionally spread while waiting for results, helping to keep patients and staff safe.

AM: Can you tell us more about the CURIAL algorithm, how it was developed and how it performed in the study?

The CURIAL algorithm was developed at the onset of the pandemic by infectious disease and machine learning experts at Oxford University.

The initiative was led by myself, in conjunction with Professor David Clifton’s “AI for Healthcare” lab at Oxford’s Institute of Biomedical Engineering.

The CURIAL artificial-intelligence algorithm looks at data that is routinely collected within the first hour of a patient coming to hospital to predict the probability of a patient testing positive for COVID-19 much sooner. The routine data assessed includes a panel of routinely performed laboratory blood tests and vital signs.

During an evaluation over a two-week test period in Oxford University Hospitals’ emergency departments, CURIAL correctly predicted a patient’s COVID-19 status with an impressive 92.5% accuracy.

AM: Validation of the algorithm is being supported by Sight OLO, Sight Diagnostics’ blood analyzer. Can you tell us more about the platform, how it works and the advantages it offers over conventional blood testing?

Yossi Pollak (YP): Sight OLO is built with high-powered microscopes, computer vision, and artificial intelligence and relies on a patented method of “digitizing” blood samples to produce over 1,000 highly detailed images of blood from each blood sample. Once the samples have been digitized, Sight OLO then deploys machine learning to measure, classify, and count the different cells and identify abnormalities. As a result, Sight OLO measures 19 different parameters with 5-part differentials, the highest grade CBC test available.

Sight OLO is a more accessible CBC testing option in comparison to traditional CBC machines which are large, heavy, and expensive, making them extremely difficult to transport or set up outside a laboratory setting. Additionally, traditional CBC testing requires a trained healthcare professional to draw a significant amount of blood from a patient, which then needs to be transferred to a central lab – from there, it can take hours or days to get the results. Sight OLO, on the other hand, only requires two drops of blood and gives patients accurate test results in ten minutes, ultimately enabling doctors to make informed treatment decisions quickly.

There are a range of benefits Sight OLO provides for both patients and the physician. Sight OLO can be placed on any stable surface and uses disposable cartridges without the need for external reagents nor liquid waste management. Sight OLO is FDA 510(k) cleared to take blood directly from a finger prick or a venous sample, which can be performed with minimal training of an operator, and the CBC result is available directly on the screen within 10 minutes.

Sight OLO received regulatory approval for use in the UK and is CE Marked according to the IVD European directive at point-of-care settings. In the United States, Sight OLO is 510(k) cleared for use in moderate complexity laboratories.

AM: How important is rapid on-the-spot diagnostic testing, and what difference could the use of Sight OLO with CURIAL-Rapide make to clinicians and patients?

The pandemic has further highlighted the need for diagnostic testing that is performed closer to the patient and that provides fast, actionable results. By having CBC test results available in a matter of minutes, doctors are able to diagnose and treat patients more quickly, and this routine data is available sooner for AI approaches making a COVID-19 status prediction.

The University of Oxford researchers are validating a version of CURIAL, named CURIAL-Rapide, that can predict the COVID-19 status of patients using only CBC results and vital signs. With OLO, a CBC can routinely be performed at the patient bedside in minutes, allowing CURIAL-Rapide to make a COVID-19 status prediction within minutes of a patient arriving at the hospital. The CURIAL-Rapide algorithm is currently being validated at the John Radcliffe Hospital, part of Oxford University Hospitals NHS Foundation Trust, using Sight OLO to obtain CBC results within minutes. Results of this study are anticipated to be made available later this year.

Dr Andrew Soltan and Yossi Polak were speaking to Anna MacDonald, Science Writer for Technology Networks.