Research Spotlight: An Explainable Artificial Intelligence Approach to Predicting Oxygen Requirement in COVID-19 Patients from Chest X-Rays

Synho Do, PhD, an investigator in the Department of Radiology and Director of Laboratory of Medical Imaging and Computation at Massachusetts General Hospital, was the senior author of a new paper in Nature Scientific Reports, Prediction of oxygen requirement in patients with COVID-19 using a pre-trained chest radiograph xAI model: efficient development of auditable risk prediction models via a fine-tuning approach.

What Were You Investigating with this Study?

We have presented a method that can be easily repurposed for new problems with explainable artificial intelligence (xAI). xAI is a unique and fundamentally different approach to black-box artificial intelligence that my team is working on.

What Methods or Approach Did You Use?

We present an upgraded chest radiograph (CXR) explainable artificial intelligence (xAI) model, which was trained on 241,723 well-annotated CXRs obtained prior to the onset of the COVID-19 pandemic.

To demonstrate the feasibility of a fine-tuning approach for efficient and scalable development of xAI risk prediction models, we applied our CXR xAI model, in combination with clinical information, to predict oxygen requirements in COVID-19 patients.

What Were the Results?

The results demonstrated that our CXR xAI model considers and validates the relevance of cardiorespiratory comorbidities, such as decreased lung volume, pleural effusion, pulmonary edema, atelectasis and cardiomegaly, to COVID-19 disease severity.

What’s Next?

Future studies should validate the performance of our CXR xAI model on other clinical important risk prediction tasks.

This could include, for example, fully automated prediction of unplanned return to the ED with or without hospitalization or intensive care unit (ICU) admission in patients with acute decompensated heart failure, acute exacerbation of COPD and elderly patients with community acquired pneumonia.

Paper Cited:

Chung, J., Kim, D., Choi, J., Yune, S., Song, K., Kim, S., Chua, M., Succi, M. D., Conklin, J., Longo, M. G. F., Ackman, J. B., Petranovic, M., Lev, M. H., & Do, S. (2022). Prediction of oxygen requirement in patients with COVID-19 using a pre-trained chest radiograph xAI model: efficient development of auditable risk prediction models via a fine-tuning approach. Scientific reports, 12(1), 21164. https://doi.org/10.1038/s41598-022-24721-5

About the Massachusetts General Hospital

Massachusetts General Hospital, founded in 1811, is the original and largest teaching hospital of Harvard Medical School. The Mass General Research Institute conducts the largest hospital-based research program in the nation, with annual research operations of more than $1 billion and comprises more than 9,500 researchers working across more than 30 institutes, centers and departments. In July 2022, Mass General was named #8 in the U.S. News & World Report list of "America’s Best Hospitals." MGH is a founding member of the Mass General Brigham healthcare system.