| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Report |
1 Department of Radiology, Pamela Youde Nethersole Eastern Hospital, 3 Lok Man
Rd., Chai Wan, Hong Kong, SAR China.
2 Department of Medicine, Pamela Youde Nethersole Eastern Hospital, Hong Kong,
SAR China.
Received May 16, 2003;
accepted after revision July 8, 2003.
Address correspondence to M. S. M. Chan
(drmonicachan{at}hotmail.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. A defined pattern of HRCT findings is observed in different phases of SARS, which is characterized by focal ground-glass and crazy paving patterns in a scattered distribution at presentation, followed by development of interstitial thickening, consolidation, pleural reaction, and scarring. Spontaneous pneumomediastinum is a distinct complication during the course of the illness.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Subsequently, the disease has also been reported worldwide [6, 7] and has prompted a global investigation of the causal agent. The initial investigation by Peiris et al. [8] suggested a Coronavirus organism as the possible cause of SARS. With the collaboration among different laboratories from countries around the world, it is now confirmed that a novel virus, being a member of the Coronavirus genus [9, 10], which was named "SARS virus" by WHO on April 16, 2003 [4], is the cause of this new disease. Clinical data for SARS are continuously gathered and published in various periodicals [1–3, 11]. As of May, it was estimated that the SARS fatality rate in Hong Kong would reach 14–15% [12].
Serial chest radiography has been the main technique in the initial investigation of patients with suspected SARS [13, 14]. On the basis of the findings of chest radiography, early experience in Hong Kong [13] showed that unilateral or bilateral peripheral pleural-based opacities [1, 2, 13], which range from ground-glass opacities to consolidation with a lower zone predominance, are observed [1]. More widespread opacification in both lungs is seen at more advanced stages of the disease [1, 2]. No pleural effusion is found on radiography [1, 2]. On high-resolution CT (HRCT), ground-glass opacities with or without thickening of the intralobular or interlobular interstitium, consolidation, or a combination of both has been described [13, 14]. HRCT is more sensitive than chest radiography and provides detailed characterization of shadows shown on chest radiography. However, because of the highly infectious nature of the disease, which needs appropriate isolation and infection-control measures during the acute phase, HRCT has not yet been adopted routinely in Hong Kong as a screening tool for this disease. In suspected cases with a history of contact but with normal findings on chest radiography or with clinical signs strongly suggestive of SARS, HRCT is recommended [13].
In Hong Kong, SARS has been diagnosed by strict adherence to guidelines issued by the Hong Kong Hospital Authority [15], which were based on WHO criteria. These guidelines rely on both clinical and radiologic features [13, 15]. After diagnosis is confirmed, our patients are treated with a regime of a combination of steroids and ribavirin. The timing for initiation of therapy is important because prompt treatment is found to have a better outcome for patients on the basis of early experience in Hong Kong. During the treatment and recovery phases, persistent shadows on radiographs may not be informative enough to distinguish disease progression and healing. HRCT may play a role in detection and characterization of the disease, in monitoring of disease progress and response to treatment, and in identification of complications. There are as yet few detailed descriptions of HRCT findings in different stages of SARS [1, 2, 13, 14]. We share our experience of HRCT findings in patients with SARS. Our analysis focuses on a pattern-based approach.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
HRCT was performed using a helical CT scanner (Somaton Plus S, Siemens, Erlangen, German) and dynamic scanning, 1-mm collimation and 10-mm spacing, 120 kV and 165 mAs, lung window setting with window center at –700 H and window width at 2,000 H or using a multidetector 16-track CT scanner (Aquillion TSX-101A M16, Toshiba, TochigKen, Japan) and 1-mm collimation, at a pitch of 15, 0.5 sec/rotation, reconstruction to 1-mm thickness and at a 10-mm interval, 120 kV, 75 mAs, lung window setting with window center at –600 H and window width at 1,600 H. All examinations were performed with the patient in the supine position with breath-holding during inspiration without IV contrast administration.
All HRCT scans were reviewed by two radiologists, with the final impression and diagnosis reached by consensus. We sought HRCT signs with the following working definitions:
Ground-glass opacity was defined as hazy increased attenuation of the lung, with preservation of bronchial and vascular markings [16]. Reticular pattern was defined as innumerable, interlacing fine-line shadows that suggested a mesh, which may be thin or thick and coarse [16]. Crazy paving pattern was defined as a thin reticular shadow superimposed on ground-glass opacities that resembled cobblestones. The underlying vessels were not obliterated, and there was no architectural distortion. Thick reticular shadows superimposed on background parenchymal opacification were defined as thick interlacing shadows resembling wire mesh superimposed on a background of increased parenchymal attenuation of the lung associated with obliteration of vessels. Architectural distortion or volume loss or both might be seen.
Consolidation with or without air bronchogram was defined as homogeneous density with obliteration of underlying vessels [16]. Marble shadow was defined as an area of patchwork-like inhomogeneous increase in attenuation without obliteration of underlying vessels. Micronodule was defined as a discrete, small, round, focal opacity with a diameter no greater than 7 mm [16]. A subpleural line was defined as a thin curvilinear opacity a few millimeters or less in thickness usually less than 1 cm from the pleural surface and paralleling the pleura [16]. Septal thickening perpendicular to the pleural surface was defined as any linear opacity of irregular thickness of 1–3 mm; distinct from interlobular septa, bronchovascular bundles, and nodular opacities [16]; and perpendicular to the pleural surface. This may be intralobular or extend through several adjacent secondary lobules [16].
Stellate shadow was defined as an irregular stellate-shaped shadow. Masslike organizing density was defined as an area of irregular masslike dense shadow obliterating or distorting vessels and bronchi within it and around the lesion. Honeycombing was defined as clustered cystic air spaces usually of comparable diameters of 0.3- to 1-cm cysts and as large as 2.5 cm, usually subpleural and characterized by well-defined walls, which were often thick [16]. Subpleural emphysematous blebs were defined as focal thin-walled radiolucency contiguous with the pleura [17].
Localized pleural thickening with or without pleural tethering and puckering was also sought. A parenchymal band was defined as an elongated band-shaped opacity, usually several millimeters wide and up to approximately 5 cm long, often extending to the pleura, which may be thickened and retracted at the site of contact [16]. Traction bronchiolectasis was defined as bronchiole dilatation, which was commonly irregular, in association with juxtabronchial opacification that was interpreted as representing retractile pulmonary fibrosis [16].
Pleural effusion, other complications such as pneumothorax and pneumomediastinum, and other coexisting disease (e.g., tuberculous granuloma, bronchiectasis, emphysema bullae) if any, were also recorded. HRCT findings were correlated with the onset of clinical symptoms. The lobar distribution of the lesions was also recorded.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
As noted in other studies, our findings concur with the observation that there is a time lag from the onset of clinical symptoms to the appearance of radiologic changes. We had eight patients with abnormal findings on HRCT but with normal findings on chest radiographs, suggesting that HRCT could depict early changes not detectable on radiography.
Although none of the signs alone is diagnostic of the disease, a general pattern with a combination of signs has been observed. Our observation suggested that among the earliest signs of the disease, ground-glass density and crazy paving pattern are key features. These signs occurred in the first week after onset of symptoms.
Ground-glass opacity was observed in nine (33.3%) of 27 patients. Of these nine patients, three showed evidence of clinical relapse, and HRCT showed a combination of ground-glass shadow with other late changes (e.g., stellate shadow or scarring) in other parts of lung. Except for patients with relapse, this sign was found at a mean time of 7 days from onset of symptoms, and no patients with this sign showed concurrent abnormalities on chest radiographs. We have observed ground-glass shadows as small as 5 mm in axial diameter. They could be in clusters, round, or wedge-shaped. Distribution can be perihilar-, central-, or pleural-based (Figs. 2 and 3). Three patients who had follow-up examinations performed 2 weeks after the first HRCT examination showed complete resolution of the ground-glass opacities.
|
|
Although it is not pathognomonic, we have observed the crazy paving pattern to be highly characteristic of early active SARS in the appropriate clinical setting. This sign has been observed in 10 (37%) of 27 patients, including four patients showing evidence of clinical relapse. Similar to ground-glass opacity, no specific pattern of distribution was observed (Fig. 4), and the crazy paving pattern followed the same time of appearance as that of ground-glass opacity. The crazy paving pattern could be found alone or coexisting with ground-glass opacity in the same patient (Fig. 5A).
|
|
We noted the frequent presence of a sharply defined line of demarcation between normal and affected areas in the same lobe and the presence of focal areas of subpleural sparing (Fig. 6A). Marble shadow, pleural effusion, and consolidation appeared within a mean time of 14–21 days from the onset of symptoms. Marble shadow was observed in five (18.5%) of 27 patients (Fig. 7). This sign is thought to represent an intermediate stage of evolving changes from the ground-glass or crazy paving pattern.
|
|
Contrary to descriptions by other investigators indicating the absence of pleural effusion [1, 2, 14], we detected pleural effusions in seven (25.9%) of 27 patients. Pleural effusions tended to be small, localized, and associated with extensive disease involvement of the underlying lung and have been observed in minor and major fissures (Fig. 6B).
|
Small areas of consolidation with or without air bronchogram were observed in 13 (48%) of 27 patients. This sign appeared to occur in areas of extensive involvement with mixed changes (Fig. 6C). Complete lobar involvement or collapse was not observed. Consolidation appeared to be associated with subsequent changes such as stellate shadow or masslike organizing density.
|
Other signs were found with a mean time between 21 and 28 days from onset of the symptoms.
One striking observation was the finding of spontaneous pneumomediastinum in patients with SARS. Seven (25.9%) of 27 patients developed pneumomediastinum, usually during or just after completion of drug treatment. All these patients had coexisting subcutaneous emphysema. Three of them had coexisting small, localized pneumothoraces, detectable only on HRCT. One patient had severe vomiting and retching, but none had been treated with ventilation immediately before the development of pneumomediastinum and pneumothorax or had a history of preexisting chronic lung disease.
Pneumomediastinum has been associated with findings in patients showing extensive involvement and presence of masslike organizing densities (Fig. 8). Subpleural blebs (Fig. 8), observed particularly in areas of subpleural sparing, could be related to the development of pneumomediastinum and pneumothorax.
|
Interstitial changes in various forms, including a thickened septal line, a linear shadow parallel or perpendicular to the pleural surface (Fig. 9), a thin or thick reticular shadow (with or without background parenchymal opacity), and a honeycomb shadow, were dominant features of the disease at this stage (Fig. 10).
|
|
Architectural distortion, localized pleural thickening or tethering or both, and traction bronchiolectasis were seen related to consolidation, thick reticular shadow, stellate shadow, and masslike organizing density (Fig. 5B). Masslike organizing density showed compaction with persistent architectural distortion on follow-up, suggestive of scar formation.
|
Honeycomb shadow could be seen in seven (25.9%) of 27 patients and often in association with masslike organizing density that was found in 14 (51.8%) of 27 patients. Stellate shadow was found in 19 (70.4%) of 27 patients.
Parenchymal band (Fig. 11) was observed in early or late stages of the disease (range, 7–38 days from onset of symptoms) in 13 (48.1%) of 27 patients, mostly in the lower lobes. It may represent different pathologic changes in various stages (e.g., atelectasis in early disease and fibrous band in the later course of disease).
|
Associated findings were observed. We had one patient with a large calcified tuberculous granuloma and bronchiectasis in the right upper lobe and one patient with bronchiectasis in the left lower lobe unaffected by other changes. Peribronchial thickening was also noted in some patients. Small clusters of micronodular shadows were observed in two patients, but we are not certain whether this finding represented a coexisting disease such as tuberculosis.
We did not observe any patient with lymphadenopathy or cavitation lesions. Pulmonary thromboembolism could not be documented because of a lack of contrast enhancement.
Bilateral involvement occurred in 24 (89%) of 27 patients. All lobes were involved in one stage or another in 17 (63%) of 27 patients. We noted that scattered distribution was a feature of SARS, and our observation suggested no lobar predilection, contrary to reports by other researchers [1, 2, 13, 14].
In summary, in the first week of illness, focal areas of ground-glass opacity and crazy paving pattern or a combination of both in a scattered distribution on HRCT were highly characteristic of SARS. From the second to third week onward, thin or thick reticular lines developed in the opacities, producing a lattice effect. With disease progression, consolidation and atelectasis were found, which might appear as masslike organizing densities. Other areas of the lungs might become involved. Pneumothorax and pneumomediastinum were distinctive complications at this subacute stage. Later, some HRCT findings resolved, leaving minimal residual changes, whereas some persisted with compaction to form scars and other associated changes, including blebs and traction bronchiectasis. Honeycombing might also develop.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
From our initial experience, we attempted to classify the HRCT findings of lung involvement into different radiologic patterns and to analyze the results against the mean time from onset of symptoms.
Although none of the signs is specific for SARS, a broad but distinctive pattern has been observed on HRCT. HRCT findings of focal areas of ground-glass opacity, crazy paving pattern, or a combination of both in a scattered distribution are highly characteristic of SARS in the first week of illness, given the correct clinical setting such as contact history. Subsequently, in the subacute phase from second to third week onward, thin or thick reticular lines develop in the opacities, producing a lattice effect. When the disease progresses, other areas of the lungs may become involved, and ground-glass opacity develops into consolidation and atelectasis, which may appear as a masslike organizing density. Vascular and bronchial anatomy become distorted, resulting in the formation of blebs and traction bronchiolectasis. Pneumothorax and pneumomediastinum are distinctive complications observed in a considerable number of our patients at this stage. In the recovery phase, most of the changes resolve, leaving minimal residual changes, whereas a masslike organizing shadow persists with compaction to fibrosis and scarring, with associated blebs and traction bronchiectasis. Irreversible changes such as honeycombing may develop with time.
Given the limitations of our study, we confirm that there is a delay in onset of radiologic changes after the onset of clinical symptoms such as fever or myalgia. It has been postulated that a hyperactive immune system against the causative Coronavirus organism after the phase of viraemia contributes to lung damage and could account for the delay [18]. Given the limited mode of response of the lung to physical, chemical, or biologic insults, an admixture of signs on HRCT shows significant overlap with other lung diseases.
Our observations suggest that ground-glass opacity and crazy paving pattern
are the earliest signs of the disease detectable on HRCT and can precede
detectable changes on chest radiography. In our review, up to 20–30% of
our patients had abnormal findings on HRCT with normal findings on chest
radiography. All of them met the WHO criteria for suspected SARS (fever 
38°C and history of contact with patients with SARS) and were later
confirmed to have SARS. Because of the highly contagious nature of the disease
and the need for isolation in controlling the spread, early diagnosis is
prudent. HRCT of the thorax allows early confirmation of diagnosis in
suspected cases of SARS. The use of HRCT is advocated and is valuable in this
particular group of patients in whom clinicoradiologic discrepancy exists.
Ground-glass opacity on HRCT is highly specific, representing air-space disease, and could be a reflection of underlying pathologic processes such as alveolar exudation. The term "ground-glass shadow" carries a somewhat different meaning in relation to chest radiography and could represent different diseases, including alveolar exudate, consolidation, or pleural effusion. Confusion should be avoided in correlating the findings of ground-glass opacity between HRCT and chest radiography. Crazy paving pattern could be a reflection of alveolar exudation and early interstitial edema or cellular infiltration. We suggest that the finding of scattered areas of ground-glass opacity or crazy paving pattern is a highly characteristic feature of the active phase of SARS in the appropriate clinical setting. In patients with clinical relapse, ground-glass opacity and a crazy paving pattern may coexist with other patterns of change.
A heterogeneous admixture of signs is seen in the next stage of the disease, suggesting active inflammatory changes, which include the development of thicker interstitial shadows representing cellular proliferation and areas of consolidation and pleural thickening suggestive of bronchiolitis obliterans organizing pneumonia. The marble shadow is thought to be a transitory change after the initial stage of ground-glass or crazy paving pattern, but we are not certain whether this sign signifies disease resolution or progression.
The reparative phase, which is characterized by disease resolution, architectural distortion, interstitial fibrosis, pleural thickening, and scar formation, follows. Stellate shadows are likely to be transitory before disease resolution. Architectural distortion is shown by traction bronchiolectasis, retraction scars, fibrous bands, and subpleural emphysematous changes.
Another complication is the development of spontaneous pneumomediastinum (25.9%), sometimes associated with small pneumothorax. Fibrosis in interlobular septa, contraction of dense fibrous scars, loss of lung compliance, and the presence of focal areas of subpleural sparing very likely all predispose to development of pneumomediastinum. Masslike organizing density, changes similar to massive pulmonary fibrosis with peripheral emphysema in pneumoconiosis, has been observed in patients with SARS.
Permanent lung damage is present when fibrous bands, interstitial fibrosis, honeycomb changes and pleural thickening and tethering, or progression of masslike organizing shadows to scars is noted. Consolidation tends to be associated with eventual formation of masslike organizing density and scars. We have also observed subtle increases in parenchymal background density in affected areas of the lung on follow-up studies in some patients, despite resolution of most other changes (Fig. 12). We are not sure whether this represents diffuse fibrosis at a microscopic level and hence permanent residual lung damage. Volume loss has also been observed.
|
Scattered distribution of involvement is a feature of this disease. A discrete, sometimes well-demarcated area of sparing is a prominent feature. Contrary to Wong et al. [14], we saw no specific lobar distribution of involvement on HRCT at all stages of illness. Pleural effusion has been observed in 25.9% of our patients, which was not described by Tsang et al. [1] and Lee et al. [2].
There are limitations to our study. Selection bias exists, partially because of the overall small number of patients in our group, and is also related to the clinical referrals. Our study group was biased toward patients in the subacute and later phases of illness. Few HRCT examinations are performed for diagnosis in the early stages of disease. Because it is highly contagious, in our early encounters with the disease, we adhered strictly to the radiologic diagnosis guidelines [13, 14], which greatly rely on chest radiography. HRCT was indicated in those suspected cases with normal findings on chest radiography. Also, at the early stage of the epidemic, most of our patients were referred from other hospitals and were confirmed SARS cases in whom consolidative changes were present on chest radiographs.
Our retrospective review of HRCT findings in a limited number of patients can at best provide a cross-sectional study with a snapshot of a segment of an evolving disease process. A longitudinal study with a longer follow-up will be more informative. Radiologic–pathologic correlation has not been available because lung biopsies are rarely performed, and autopsy is risky because of the highly infectious nature of SARS. Many questions remain unanswered, including the role of HRCT in screening suspected cases, timing of treatment, prognostic planning, and quantification of long-term lung damage. Further studies will shed more light on this new disease.
SARS is a highly contagious disease. Special infection-control measures must be instituted for HRCT examination. We scheduled suspected SARS patients before confirmed SARS patients toward the end of our routine session followed by a rest period for cleansing and decontamination. Patients were thoroughly bathed and changed into clean clothes and masks in wards before being escorted to the radiology department. Special transport arrangement was made via a dedicated route to minimize contact with people and the time of stay in the public areas. Strict droplet precautions were taken, including the use of personal protective equipment such as goggles, face shields, masks, gloves, and disposable gowns and caps. In the CT scanning room, the number of staff in close contact with the patient was kept to a minimum. More intense protective measures, including a positive-pressure breathing apparatus, were used for highly infectious cases. A viral filter is mandatory on mechanical ventilators. Modification work will be undertaken to enhance air exchange using negative-pressure ventilation and to minimize the spread of droplets.
In conclusion, on the basis of our initial experience, a fairly defined but broad pattern of change has been observed on HRCT in different phases of SARS. Ground-glass opacity and crazy paving pattern in a scattered distribution are typical early changes of the disease. Further progression of disease leads to development of interstitial changes and consolidation, which are followed by a reparative phase with disease resolution or fibrosis and scarring. Spontaneous pneumomediastinum is a distinctive complication of SARS in the early reparative phase. Pleural effusion and other forms of pleural reaction have been observed in our patients.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Roberts, C. Paddock, L. Vogel, E. Butler, S. Zaki, and K. Subbarao Aged BALB/c Mice as a Model for Increased Severity of Severe Acute Respiratory Syndrome in Elderly Humans J. Virol., May 1, 2005; 79(9): 5833 - 5838. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
