Following viral transmission and infection of the lower respiratory tract, SARS-CoV-2 generates a radiographic lung injury pattern that is generally similar to that caused by other viral respiratory pathogens. In the early stages of COVID-19 pneumonia, the most common parenchymal abnormality is ground glass opacity, followed by consolidation [2]. Autopsy studies performed on patients with severe COVID pneumonia suggest these radiographic findings most likely represent alveolar hyaline membranes and reactive type II pneumocytes seen in the exudative diffuse alveolar damage phase [3]. Beyond the acute stage of infection, there is typically more radiographic heterogeneity including complete radiographic resolution, the persistence of ground glass opacities, or the development of other interstitial lung abnormalities (ILAs) such as reticulation, architectural distortion, or traction bronchiectasis (Figs. 1, 2, 3, and 4). Notably, ILAs must involve at least 5% of a lung zone and do not include radiographic features suggestive of alternative etiologies such as dependent atelectasis, focal fibrosis adjacent to spinal osteophytes, isolated nodularity, or interstitial edema [4]. This variability in radiographic features was demonstrated in a study of 52 hospitalized patients with COVID-19 pneumonia who underwent follow-up CT imaging approximately 4 months after discharge. The most common radiographic abnormalities identified on this follow-up scan were mosaic attenuation (30 of 52), reticulations (27 of 52), and architectural distortion (24 of 52). The authors did not report what proportion of patients had a normal 4-month follow-up CT scan. Patients in this cohort with severe or critical COVID-19 pneumonia—defined by the presence of hypoxia, tachypnea, or multiorgan failure—had a higher prevalence of mosaic attenuation pattern with hypoattenuated areas (66% versus 13%) and reticulations (59% versus 13%) on the follow-up CT scan when compared to patients with mild or moderate infection. There was also a non-significant trend towards more architectural distortion (52% versus 13%) in patients with severe or critical disease suggesting that increased disease severity is associated with a greater risk of post-COVID ILAs [5]. Similarly, a large single-center cohort study of 353 subjects from Wuhan, China, found that ground glass opacities (159 of 353) and irregular lines (56 of 353) were the most common radiographic features identified on CT scans performed six months after hospitalization for COVID-19 pneumonia. This study also identified a positive association between the increasing severity of COVID-19 pneumonia and the likelihood of persistent ILAs at 6 months [6]. This heterogeneity in radiographic features seen on initial follow-up CT imaging makes it challenging to predict which patients go on to develop post-COVID fibrosis, though recently published studies with longer follow-up periods offer slightly more clarity.

Representative axial CT scan slices of a patient with post-COVID interstitial lung abnormalities. A CT chest has done 2 months after discharge. Prominent radiographic features include ground glass opacities with mosaic attenuation, reticulations, and irregular lines. B CT chest has done 1 year later with improved ground glass opacities and scattered irregular lines

Representative axial CT scan slices of a common pattern of radiographic progression after COVID-19 pneumonia. A Baseline CT chest has done before the COVID-19 infection showing normal lung parenchyma. B CT chest done on admission for COVID-19 pneumonia showing peri-bronchovascular ground glass opacities and consolidations. This patient was treated with a course of systemic corticosteroids. C CT chest was done 10 months after hospitalization with minimal ground glass opacities but the architectural distortion was not present before COVID. Subpleural reticulations (arrow) and traction bronchiectasis (arrowhead) identified

Representative axial CT scan slices of a patient with baseline SSc-ILD exacerbated by COVID-19 pneumonia. A Baseline CT chest has done before the COVID-19 infection showing minimal ground glass opacities and reticulations in the left lung base. B CT chest has done 2 months after COVID-19 hospitalization showing a significant increase in diffuse ground glass opacities and traction bronchiectasis (arrow) not present before COVID. C CT chest has done 1 year later showing no significant improvement

Representative axial CT scan slices of a patient with post-COVID interstitial lung abnormalities. A CT chest has done during hospitalization for COVID-19 pneumonia demonstrating bilateral ground glass opacities, bibasilar consolidation (arrows), and pneumomediastinum. B CT chest has done 1 year after discharge shows improved ground glass opacities and residual architectural distortion with both peri-bronchovascular and subpleural reticulations (arrows). New traction bronchiectasis (arrowhead)

It remains unclear whether radiographic abnormalities which may persist for several months after COVID-19 pneumonia represent slowly resolving diffuse alveolar damage or the emergence of fibrotic lung disease from aberrant repair mechanisms. Several meta analyses have attempted to answer this question by characterizing the evolution of these radiographic abnormalities over a longer period of follow-up after hospitalization for COVID-19 pneumonia. The largest of these meta analyses done by Fabbri and colleagues compiled 46 studies examining the persistent radiographic sequelae up to twelve months after SARS-CoV-2 infection. They noted that patients with ground glass or consolidative opacities during the initial hospitalization had a higher rate of radiographic improvement on follow-up imaging when compared to patients with ILAs (0.92 to 0.44 and 0.32 to 0.26, respectively). The timing of the follow-up scan in this study was positively associated with the improvement or resolution of the ground glass or consolidative changes. In other words, taken collectively the ground glass and consolidative opacities became less prominent the farther out from COVID-19 hospitalization. The authors suggest this association reflects the typical course of resolving diffuse alveolar damage. In contrast, there was no such temporal association in patients with ILAs suggesting these findings represent an irreversible process [7•]. Similar findings were seen in a prospective cohort study of 62 participants from Wuhan, China, published in 2021. In this study, 56% of participants had ILAs on follow-up CT imaging done 6 months after hospitalization with COVID-19. All of these participants demonstrated persistent ILAs on another follow-up CT scan done 6 months later; however, approximately 25% (8 of 35) showed a slight reduction in severity suggesting that some patients may continue to show slow radiographic improvement for several years after hospitalization with COVID-19. Additionally, of the remaining participants with non-fibrotic abnormalities on 6-month follow-up imaging, 85% (23 of 27) showed partial or complete resolution on the 1-year CT scans [8]. This slow improvement was also seen in a study by Pan et al. of 209 subjects hospitalized with severe COVID-19. The CT abnormalities seen during the index hospitalization in this study resolved in 61% of participants (128 of 209) at 3 months and in 75% of participants (156 of 209) at 6 months [9]. These studies suggest that while a significant proportion of patients hospitalized with COVID-19 have persistent abnormal findings on follow-up imaging, many of these patients show partial or complete resolution over the following months. However, those with ILAs—non-dependent reticulation, architectural distortion, traction bronchiectasis, honeycombing—on initial follow-up imaging typically do not show significant improvement over time.

As residual ILAs are an emerging complication following COVID-19 infection, there is increasing interest in defining both the prevalence and risk factors associated with persistent ILAs after COVID-19 hospitalization. The largest study to date of residual lung abnormalities following COVID-19 is currently being performed by the PHOSP-COVID Study Collaborative Group in the UK. Interim analysis of this UKILD study released in December 2022 found residual lung abnormalities—defined as ground glass opacities or reticulations—in nearly 80% of CT scans performed at a median follow-up time of 140 days after COVID-19 hospitalization. Twenty-six out of the twenty-eight subjects that had a second follow-up CT scan done at least 3 months later still had residual abnormalities. Based on Bayesian modeling, the authors of the UKILD study estimate the prevalence of residual lung abnormalities to be up to 11.7% in patients hospitalized with COVID-19. The clinical indicators with the strongest association with residual lung abnormalities in this cohort were severe illness requiring NIPPV, IMV, or ECMO (RR 1.40, 1.23–1.63), abnormal CXR findings at follow-up (RR 1.40, 1.22–1.61), and post-hospitalization percent predicted DLco < 80% (RR 1.26, 1.02–1.58) [10••]. Other risk factors identified as being associated with residual lung abnormalities after COVID-19 hospitalization include older age, male sex, comorbid hypertension, longer length of stay, prior tobacco use, higher C-reactive protein, lower circulating interferon gamma, and lymphopenia [11,12,13]. Additional studies, including the secondary analysis of the UKILD study, will likely better define the risk factors associated with residual ILAs after COVID-19 hospitalization and may lead to the creation of scoring systems to facilitate more targeted surveillance and treatment of such a large patient population.

While longer-term studies are ongoing, the preponderance of the available evidence suggests that approximately 30% of patients have residual lung abnormalities 6 months after COVID-19 hospitalization. Most patients continue to have slow improvement in imaging over the following months or years; however, it is estimated that up to 12% of patients have persistent ILAs consistent with irreversible fibrotic lung disease.