Radiologic Findings after COVID-19 and the Correlation with Lung Function

See also the article by Han and Chen et al in this issue.

Dr van Beek is the Scottish Imaging Network (SINAPSE) chair of clinical radiology at the University of Edinburgh, director of Edinburgh Imaging, and a consultant cardiothoracic radiologist at the Royal Infirmary of Edinburgh. His research interests are multimodality imaging of heart and lung diseases, validation and implementation of artificial intelligence tools, and developments in lung cancer screening. He is a member of the Fleischner Society and former chairman of the International Workshop on Pulmonary Functional Imaging.

Radiology has been at the forefront of the clinical evaluation of patients with COVID-19 and has helped improve knowledge about this viral disease by demonstrating pathophysiologic imaging patterns. From the earliest days of the pandemic, imaging findings were reported that helped clinicians all over the world recognize the disease and make appropriate treatment decisions (1). Given the huge number of patients who required medical attention, the medical system was at a breaking point, and guidelines were offered on how to best employ radiologic investigations. It soon became apparent the lungs were not the only organ affected; vascular complications were also an issue (2,3). In the majority of patients who survived the disease, healing was uncomplicated. However, a significant number of patients had demonstrable abnormalities over a prolonged period of time, with some of those abnormalities becoming chronic and leading to disability (4). The term “long COVID” became widely known in society, and dedicated studies were proposed. These studies are demonstrating the long-term findings in COVID-19 (5), as well as the changing pattern of disease after the effects of vaccine programs and new variants (6,7). Clearly, the imaging findings are evolving and must be evaluated in the context of clinical symptoms. MRI methods have been used to evaluate more in-depth pathophysiologic correlations (3,8,9).

In this issue of Radiology, Han and Chen et al (10) provide insight into the long-term consequences of COVID-19 in a consecutive series of participants discharged from the hospital following severe COVID-19 in Wuhan, China. Of the initial 1251 participants, 144 participants (aged 18–80 years) with baseline chest CT scans of adequate image quality and lung abnormalities at discharge were ultimately included. Follow-up investigations included a self-reported respiratory symptoms questionnaire and lung function testing at 6 months, 12 months, and 2 years after symptom onset. Of the 144 participants, 129 completed the 2 years of follow-up with complete investigations.

CT findings focused on the presence of ground-glass abnormalities and interstitial lung abnormalities (ILAs), which were divided into a fibrotic category (architectural distortion with traction bronchiectasis and/or honeycombing) and a nonfibrotic category (ground-glass opacities or reticular abnormalities). The findings were grouped according to a semiquantitative CT score based on the area involved in each of the five lung lobes. The CT findings and symptoms were correlated with the diffusing capacity of lung for carbon monoxide (Dlco), which was considered abnormal if less than 75% of the predicted value.

The authors found that participants discharged following severe COVID-19 showed gradual overall improvement in respiratory symptoms, with the proportion of individuals with at least one respiratory symptom decreasing from 30% at 6 months to 22% at 2 years. The most common prevailing symptom at the 2-year follow-up was exertional dyspnea, which persisted in 14% of participants. The Dlco remained abnormal in more than one-third of participants at 2 years, even as the proportion of participants showing ILAs on CT scans decreased from 54% at 6 months to 39% at 2 years.

The Han and Chen et al (10) study offers insight into the long-term follow-up after severe COVID-19 infection. It shows persistent abnormalities in more than 30% of the participants and demonstrates that there is an incomplete correlation between symptoms, CT findings, and Dlco. A few points are important to consider when reading their article.

First, the study population was focused on individuals with severe COVID-19 and, therefore, the results only apply to those patients (11% of all potential cases in this cohort; more than 80% of these participants had severe COVID-19 or were in critical condition). As we now know, the majority of patients with COVID-19 will have mild to moderate symptoms or recover well following severe COVID-19. Thus, the fact that so few of the many patients with COVID-19 have persistent abnormalities should provide us with confidence. In this study, 32 participants (or 2.5% of the entire cohort) remained symptomatic. Although a small percentage, the sheer number of patients affected still means that this will continue to remain a significant burden on the health care system and society at large.

Second, the CT findings of fibrotic and nonfibrotic ILAs are important to distinguish, as shown in this study. Once distortion and traction bronchiectasis and bronchiolectasis are present, the findings will likely become irreversible, as noted by the authors. Therefore, it is important during follow-up examinations in clinical practice to recognize this difference, as patients with nonfibrotic ILA findings are more likely to continue to recover, even as late as 2 years after the initial infection episode.

Third, there is a correlation between symptoms, Dlco, and CT findings, but this correlation is far from perfect. This is important to realize, as CT will always be an imperfect test to assess lung function. What is more, COVID-19 is known to also cause vascular abnormalities, which can lead to inhomogeneous ventilation-perfusion and mismatching. This would also result in dyspnea symptoms and Dlco abnormalities, but cannot be assessed with thin-section CT alone.

Fourth, although these imaging findings are highly relevant in the long-term assessment of symptomatic patients after COVID-19 infection, other modalities are also starting to offer insight into the pathophysiologic characteristics of this illness. For instance, the application of dual energy CT (6,7) will allow the study of perfusion abnormalities, while xenon 129 (¹²⁹Xe) MRI has been effective in demonstrating oxygen transfer limitations, with reduced red blood cell oxygenation and reduced diffusion capacity in the lung tissue (8,9).

Finally, with the changing virus and establishment of vaccination programs, it appears that this disease is less severe in symptoms, as well as imaging findings (6,7). Together with the fact that relatively few patients following the initial COVID-19 wave have persistent symptoms and imaging findings, this should fill us with hope to be able to live with COVID-19 in the future.