Today, we are extremely proud to announce a new and significant update to the project: COVIDNet-S.
COVIDNet-S is a suite of deep learning models designed to assess the disease severity of COVID-19.
As viral procedures like RT-PCR and antibody tests cannot diagnose the severity of COVID-19 for a given patient, chest X-Rays have routinely been leveraged to assist health care professionals in tracking the progression of the disease.
To this end, COVIDNet-S can quantitatively score the geographic and opacity extent in a patient’s lungs by analyzing key visual indicators of their chest X-Ray. The system was developed using over 10,000 chest X-Rays, with hundreds of these being COVID-19 positive patients with comprehensive lung disease severity assessments in collaboration with the wonderful team at Stony Brook Medicine in New York.
Per the following testimony of Dr. Richa Mittal, a practicing radiologist at Guelph General Hospital, we have high hopes that COVIDNet-S will be a useful tool in supporting front-line workers throughout the world:
“It has been half a year of the COVID-19 pandemic, and we have learned so much. One of the most crucial insights from a medical point of view is the importance of early diagnosis and accurate prognosis in infected patients. Chest radiography is the fastest and most economic test available from an image standpoint.
Although chest X-rays can appear normal in infected patients early in the disease course, differentiating COVID-19 from other pathologies can be challenging. Also assessing the severity of the findings on chest radiography can help appropriately triage patients and potential medical/ICU resources. COVIDNET-S has the potential to improve diagnostic accuracy and standardize a grading system for disease severity. It will be a valuable tool for physicians in managing the next phase of this pandemic.”
As before, we have open sourced all aspects of the project: the models for COVIDNet-S are available at our Github repository here, along with COVIDx-S, a dataset of over 480 lung disease severity annotations from the collaborating team of radiologists at Stony Brook Medicine.
In addition, our academic study on the efficacy of the system can be found here.
Finally, if you’re a researcher or clinician and would like access to our explainability platform to assist with this project and gain transparency or have data to share, please email us at firstname.lastname@example.org.
A special thanks to both NVidia and HPE for providing computing resources to accelerate the development of COVIDNet-S.
The DarwinAI team