Skip Navigation

Leveraging on Computational Tools to Facilitate the Severity Assessment of COVID-19 Pneumonia

April 6, 2023

Leveraging on Computational Tools to Facilitate the Severity Assessment of COVID-19 Pneumonia

Photo Caption: Ana Vargas Angles, Ph.D., Dery Gamero Tejeda, Ph.D., Javier Jo, Ph.D., Ana Maria Gutierrez Valdivia, Ph.D., Eveling Castro, Ph.D.
Photo Caption: Ana Vargas Angles, Ph.D., Dery Gamero Tejeda, Ph.D., Javier Jo, Ph.D., Ana Maria Gutierrez Valdivia, Ph.D., Eveling Castro, Ph.D.

Through an international collaborative research project, a multidisciplinary team of engineers and clinicians from the University of Oklahoma and the Universidad Nacional de San Agustin (UNSA) in Arequipa, Peru, is developing computer-aided diagnosis tools that could enable faster, more accurate pneumonia severity assessment.

The accurate assessment of pneumonia severity caused by COVID-19 and other infectious lung diseases is critical for developing an effective, personalized treatment for each patient. Pneumonia severity assessment is commonly performed by radiological evaluation of computed tomography (CT) scans, which involve analyzing large volumes of time-consuming data and requires the availability of trained radiologists.

The team of OU and UNSA engineering researchers have developed state-of-the-art deep learning models to analyze chest CT scans. To evaluate the performance of these deep learning models and the computer-aided diagnosis tools, UNSA engineering professors and students work closely with UNSA radiologists to identify CT scan datasets and assess the severity of pneumonia in 80 COVID-19 patients from Arequipa. Preliminary validation results, recently reported in a paper titled, "Automated Quantification of Pneumonia Infected Volume in Lung CT Images: A Comparison with Subjective Assessment of Radiologists" in the open access journal Bioengineering, indicate promising agreement between the radiologists and the computer-aided diagnosis tools and deep learning models severity assessment of COVID-19 pneumonia. 

This project ultimately seeks to disseminate a working computer-aided diagnosis system to medical centers throughout Arequipa that will enable fast and accurate pneumonia severity assessment for underserved populations lacking access to qualified radiologists.

About the project:

This research collaboration project titled, "Development of Machine Learning Models Based on Radiology Images for Stratification and Severity Assessment of COVID-19 Pneumonia" is led by Javier Jo, Ph.D., principal investigator at the University of Oklahoma, and Eveling Castro Gutierrez, Ph.D., principal investigator at the Universidad Nacional de San Agustin. The team includes postdoctoral researchers at OU and undergraduate and graduate students from UNSA. The project is supported by the OU-UNSA Global Change and Human Health Institute and administered through the Latin America Sustainability Initiative, and Initiative of the OU Institute for Resilient Environmental and Energy Systems.