Omnia Health is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

coronavirus.jpg

MENA's first free diagnostic tool for COVID-19 using AI in medical images

The AI-powered assistant to radiologists supports the interpretation and diagnosis of COVID-19 in chest CT scans of patients.

Radiologists and clinicians can now freely use AiRay, the Artificial Intelligence (AI) radiology solution, to assist in the detection of the novel coronavirus disease, COVID-19, signs at any Chest Computed Tomography (CT) scan. This is considered the region’s first implementation of the Convolutional Neural Network (CNN) in the field of radiology and medical imaging.

TachyHealth, the leading Artificial Intelligence healthcare startup, has been working on the research and development of this deep learning model for 2 months using 805 images of laboratory confirmed COVID-19 chest CT images to achieve a validated real-life accuracy performance of 88%. The computer vision model is based on a pre-trained very deep neural network with several hidden layers optimized based on Imagenet dataset which contains millions of images with several hundred categories.

AiRay for COVID-19 is part of TachyHealth’s efforts to utilise the power of cutting-edge technology, cloud computing, and machine learning to enhance the diagnosis of patients amid the COVID-19 pandemic and boost the ability of the healthcare systems to respond better to the surge increase in the need for the radiology services.

Artificial Intelligence plays a significant role in the current and future radiology practice by leveraging the machines capabilities to meet the needs of the radiologists, technologists, and the wider healthcare professionals. This free release of AiRay will be especially beneficial in geographical areas where the radiologists are in shortage and being overwhelmed by the interpretation of the CT scans.

This initial cloud release of AiRay was designed not to store any confidential information to ensure patient safety and confidentiality. The model has achieved a satisfactory performance for the current emigrant situation with a very strong capability for future improvement when more data is available.

AiRay for COVID19 Screen - TachyHealth.png

On the benefits of this initiative, Dr. Osama AbouElkhir, the CEO, shared: “We wanted to help in this crisis in areas where we have our unique capabilities. This is the first Artificial Intelligence solution in the Middle East and North Africa that is provided for free to support the identification of the COVID-19 findings in the chest CT images.

We followed a rigorous method to build our AI model following the latest studies and guidelines including those of the Radiological Society of North America (RSNA), Royal College of Radiologists (RCR), European Society of Radiology (ESR), and the UK Surgical Royal Colleges, amongst others.

There is a growing body of evidence supporting the role of the CT in the assessment of patients with severe respiratory distress and for aiding the diagnosis of COVID-19 in persons under investigation (PUI). As we are moving further with this development, we’re opening the door for collaboration with researchers, hospitals, universities, and radiology centers to further improve the model from both radiological and machine learning perspectives.”

Highlighting the merits of the data collection and model performance, Ahmed Sahlol, PhD, the head of AI research in TachyHealth added: “We’ve collected the images from credible evidence, studies, and resources to create a heterogenous dataset that encompasses wide array of images from different hospitals, axial cuts, CT machine, patient positions, patient demographics (age, gender), and clinical conditions. We minimized the bias in the algorithm by training the model on raw (real world) images instead of carrying out a heavy cleaning and pre-processing on the dataset. So, our model achieved both, reality and acceptable efficiency.

This was performed because we wanted the model to be used in real world scenarios instead of performing exceptionally well in a laboratory-controlled setting and very poor afterwards. Thus, the performance of our model reached 88% in accuracy, 75% in recall, 81% for accuracy precision, and 78% for the F1 score.

We believe that the model should be viewed as an assistant (computer-aided) rather than replacement or substitution to professional radiological practice. We keep improving the model performance with more data collected on daily basis as well as fine tuning the architecture of the model.”

In 2016, The World Health Organization (WHO) estimated 3.6 billion diagnostic medical imaging examinations are performed worldwide every year. This number continues to grow as more people access medical imaging facilities while the complexity of the healthcare diseases increases.

On the other side, radiologists are in shortage globally, especially in developing countries and rural areas. In 2015, Liberia for example, only had 2 radiologists, whilst Ghana had 34 radiologists and Kenya had 200 radiologists. Even in developed countries, in the UK for example, three quarters of radiology clinical directors say they do not have enough radiology consultants to deliver safe and effective patient care whilst their workload of reading and interpreting medical images has increased by 30% between 2012 and 2017.

From both the supply and demand data, there is a growing need for scalable and accurate cognitive radiological systems. Advanced Artificial Intelligence systems are gaining trust for their abilities to accurately analyse, interpret, and report on medical images.

Dr. Amr Fawzy, MD, the Chief Medical Officer said, “We see AI as an opportunity to better serve the patients around the world, support the research in deep technology, advance the collaborations between healthcare professionals across the region, and improve the overall value-based healthcare.

For hospitals, medical centers, and radiology centers who are facing issues in availability of the radiologists, the timeline of reporting, or the burn-out of the radiology team, AiRay is a boon as it brings AI easily into their everyday practice.

In the event of a pandemic like COVID-19, this solution would help manage the accelerating demand from those communicable diseases. Radiologists can also achieve productivity gains working with AI from anywhere with features-rich cloud-based secure system, enables them to see more patients, interpret more images who in turn, the patients, receive more timely and accurate care.

In the future, AiRay will be integrated with both the radiology information system (RIS) as well as the picture archiving and communication system (PACS) to be part of the routine workflow of the radiologists. AiRay is developed as a common platform for high scalability across the health ecosystem with little upfront costs”

Mo Khadra, the Chief Technology Officer said “We’re working on an end-to-end solution to increase AiRay capabilities by adding multiple features to augment access to medical services such as teleradiology services, prioritization of the radiology list, multi-party video conferencing for consultation with multi-disciplinary care teams, file sharing, visualizing medical reports for reference during the radiologist confirmation, amongst others.

The AI system communicates with both the RIS and PACS using the widely known health data exchange protocols (HL7) and using the global technical image standards (DICOM) to provide a platform that is interoperable with other IT systems at the care delivery point. We are redesigning the machine-human interaction with a view about interoperability between the rigorous legacy healthcare systems and keeping process simple and accessible using end-to- end encryption and protection with highest security measures.

We want to build a system that help physicians, clinicians, and radiologists, do what they are good at .... Taking care of our families and loved ones”

Hide comments
account-default-image

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish