The exponential increase in COVID-19 patients is overwhelming healthcare systems across the world.
With limited testing kits, it is impossible for every patient with respiratory illness to be tested using conventional techniques (RT-PCR). The tests also have long turn-around time, and limited sensitivity. Detecting possible COVID-19 infections on chest X-Ray and chest CT scan may help quarantine high risk patients while test results are awaited. X-Ray machines are already available in most healthcare systems, and with most modern X-Ray systems already digitized, there is no transportation time involved for the samples either.
In this work we propose the use of chest X-Ray to prioritize the selection of patients for further RT-PCR testing. This may be useful in an inpatient setting where the present systems are struggling to decide whether to keep the patient in the ward along with other patients or isolate them in COVID-19 areas. It would also help in identifying patients with high likelihood of COVID with a false negative RT-PCR who would need repeat testing. Further, we propose the use of modern AI techniques to detect the COVID-19 patients using X-Ray and CT scan images in an automated manner, particularly in settings where radiologists are not available, and help make the proposed testing technology scalable.
We present CovidAID: COVID-19 AI Detector, a novel deep neural network based model to triage patients using X-Ray and CT scan images, for appropriate testing. On the publicly available covid-chestxray-dataset, our model gives 90.5% accuracy with 100% sensitivity (recall) for the COVID-19 infection. We significantly improve upon the results of Covid-Net on the same dataset.
We are open to collaborate with partners from clinical and industry sectors, who could be data owners, domain experts or medical/technology device owners. We are also looking at opportunities for prototyping and testbedding of our technology solution.
We are able to provide a demo. So please feel free to write-in.
AI - X-Ray CT Scan Dataset Model
There is currently no consensus for any COVID-19 imaging protocols formed by the radiology community. The situation is changing rapidly, and new cases of the disease are being documented and reported on by medical experts on a daily basis. As such, industry is not in a position to provide guidance on this issue. Any promotion of a CT protocol and its benefits must be proposed and substantiated by a medical professional.
What we can offer, however, is a complete set of standard protocols that are the default on every medical CT system. These have been rigorously tested over many years to demonstrate a wide range of pathologies for each body region, for example, the chest CT protocol which can show early inflammatory changes within the lungs, in addition to severe pneumonia.
This technology solution could be utilised by Health Institutions, Clinics and labs to detect the pressence of COVID-19 as well as severity of the illness through our AI-based autonomous models.
Existing COVID-19 detection is conducted through rapid test kit and PSR test. The severity of the cases are not easily identified due to asymptomatic and symptomatic cases. But through our AI-based learning model we could predict the severity output. The accuracy of the output could be increased by training of the AI model with more diverse and region-specific datasets.
Using the severity prediction for cases detection, we can identity the probable patients and their treatment methods as soon as possoble. This can decrease turn around time for treatment and assist in reducing health cost by making the workflow more efficient.