AI diagnostic tool helps doctors visualize lung damage after COVID infection

A new computer-aided diagnostic tool developed by scientists at King Abdullah University of Science and Technology (KAUST) could help overcome some of the challenges of monitoring lung health after a viral infection. Like other respiratory infections, COVID-19 can cause lasting damage to the lungs, but doctors have struggled to visualize that damage. Traditional chest scans cannot reliably detect signs of lung scarring and other lung abnormalities, making it difficult to track the health and recovery of people with ongoing breathing problems and other post-COVID complications.

The new method developed at KAUST — called deep lung parenchymal enhancement (DLPE) — overlays an artificial intelligence algorithm on top of standard chest imaging data to reveal otherwise indiscernible visual features that indicate lung dysfunction.

With DLPE enhancement, "radiologists can discover and analyze new sub-visual lung lesions," said computer scientist and computational biologist Xin Gao. Analysis of these lesions could then help explain patients' respiratory symptoms, "allowing for better disease management and treatment," he added.

Xin Gao and members of his structural and functional bioinformatics groups and the Center for Computational Bioscience Research created the tool with artificial intelligence researcher and current KAUST Provost Lawrence Carin and clinical collaborators from Harbin Medical University in China.

The method first eliminates any anatomical features that aren’t related to the lung parenchyma, the primary site of COVID-19-induced damage. That means removing airways and blood vessels, then enhancing the remaining images to expose lesions that might be missed without computer help.

The researchers used computed tomography (CT) chest scans from thousands of hospitalized COVID-19 patients in China to train and validate their algorithm. They refined the method with input from expert radiologists and then applied DLPE in a prospective fashion to dozens of COVID-19 survivors with lung problems who had experienced severe disease requiring intensive care treatment.

In this way, Gao and his colleagues demonstrated that the tool could reveal signs of lung fibrosis in patients with long-term COVID symptoms, helping to explain shortness of breath, coughing and other lung problems. Such a diagnosis, he suggested, would not be possible with standard CT image analysis.

"Using DLPE, we have demonstrated for the first time that long-standing CT lesions can explain this symptom. Therefore, treatment of fibrosis may be very effective in addressing the long-term respiratory complications of COVID-19," he said.

Although the KAUST team developed DLPE primarily with post-COVID recovery in mind, they also tested the platform on chest scans collected from people with a variety of other lung problems, including pneumonia, tuberculosis, and lung cancer. The researchers demonstrated how their tool could serve as a broad diagnostic aid for all lung diseases, enabling radiologists to "see the invisible," as Gao Xin said.