Radiologic Pattern of Disease in Patients with Severe Acute Respiratory Syndrome: The Toronto Experience 1

SPECIAL REPORT 553 Radiologic Pattern of Disease in Patients with Severe Acute Respiratory Syndrome: The Toronto Experience 1 Narinder S. Paul, MD Heidi Roberts, MD Jagdish Butany, MD TaeBong Chung, MD Wayne Gold, MD Sangeeta Mehta, MD Eli Konen, MD Anuradha Rao, MD Yves Provost, MD Harry H. Hong, MD Leon Zelovitsky, MD Gordon L. Weisbrod, MD Severe acute respiratory syndrome (SARS) is a transmissible febrile respiratory illness caused by a recently discovered coronavirus. Various patterns of disease progression may be observed that have different implications for the prognosis in those affected by SARS. The appearance of the lungs on chest radiographs of patients with this condition may be normal or may include focal airspace opacity or multifocal or diffuse opacities. Thoracic computed tomography (CT) is more sensitive in depicting SARS than is conventional chest radiography, and CT images obtained in patients with normal chest radiographs may show extensive disease and airspace consolidation. However, because the radiologic appearance of SARS is not distinct from that of other diseases that cause lower respiratory tract infection, early identification of SARS will depend in part on the prompt recognition of clusters of cases of febrile respiratory tract illness. To aid in the differential diagnosis and management of SARS, radiologists must be familiar with the typical clinical and histopathologic findings, as well as the radiologic features of the disease. RSNA, 2004 Abbreviations: CoV coronavirus, SARS severe acute respiratory syndrome, WHO World Health Organization Index terms: Lung, infection, 60.2199, 60.4134 Respiratory distress syndrome, adult (ARDS), 60.4134 Severe acute respiratory syndrome (SARS), 60.2199 Lung, CT, 60.1211 2004; 24:553 563 Published online 10.1148/rg.242035193 1 From the Thoracic Division, Department of Medical Imaging, University Health Network and Mount Sinai Hospital, University of Toronto, Canada (N.S.P., H.R., T.B.C., E.K., A.R., Y.P., L.Z., G.L.W.); Departments of Pathology (J.B.) and Medicine (W.G.), University Health Network, University of Toronto, Canada; Department of Medicine, Divisions of Respirology and Critical Care, Mount Sinai Hospital, University of Toronto, Canada (S.M.); and Department of Pulmonary and Critical Care Medicine, York Central Hospital, York, Ontario, Canada (H.H.H.). Received September 18, 2003; revision requested October 16 and received November 7; accepted November 7. Address correspondence to N.S.P., Department of Medical Imaging, 3-956 Princess Margaret Hospital, 610 University Ave, Toronto, ON, Canada 5MG 2M9 (e-mail: narinder.paul@uhn.on.ca). RSNA, 2004

554 March-April 2004 RG f Volume 24 Number 2 Introduction Severe acute respiratory syndrome (SARS) is a transmissible respiratory tract infection caused by a recently identified coronavirus (CoV) (1). On March 12, 2003, the World Health Organization (WHO) issued a global alert in response to the rapid spread of atypical pneumonia among previously healthy adults in Guangdong Province, in southern China (2). Initial reports of the disease indicate that from November 16, 2002, to August 7, 2003, there were 8,422 probable cases of SARS and 916 deaths, with a global fatality of approximately 11% (3). To familiarize radiologists with the possible manifestations of SARS and to enable more effective surveillance and response, this article presents an overview of the clinical, histopathologic, and radiologic features of SARS in patients in Toronto, Ontario, with emphasis on the radiologic pattern of disease. Infectious Agent The coronavirus known as SARS-CoV has been isolated from respiratory secretions and stool and blood specimens from living patients and detected in postmortem specimens from lungs and other organs including bowel, liver, spleen, lymph nodes, and kidney (1,4). SARS-CoV is transmitted mainly through respiratory tract secretions or by direct contact. Although there are genetic similarities between SARS-CoV and previously known coronaviruses, SARS-CoV is a new virus strain. It is believed to be the first coronavirus to cause severe respiratory illness in immunocompetent adults (1). Clinical Features In Toronto and Hong Kong, most SARS exposures have occurred in the hospital setting (5,6). The WHO definitions of possible and probable SARS cases are based on clinical and laboratory findings (7). The clinical findings that determine a case of suspected SARS, according to WHO criteria, include the presence of fever (temperature 38 C) and symptoms of lower respiratory tract illness, a history of recent contact with a patient known to have SARS or of travel within 10 days to a WHO-identified SARS-affected area, and the absence of an alternative diagnosis that would fully explain the illness. A case of probable SARS, in addition to these findings, includes a chest radiograph that shows infiltrate in the lungs Table 1 WHO Clinical Definitions of SARS Suspected SARS History of high fever ( 38 C) and cough or breathing difficulty and SARS exposure during the 10 days prior to onset of symptoms Probable SARS Suspected SARS and radiographic evidence of pneumonia or respiratory distress syndrome (RDS) or autopsy findings consistent with RDS without an identifiable cause (Table 1). The laboratory definition of SARS includes all of the symptoms and signs in the clinical definition of probable SARS, as well as positive laboratory findings of SARS-CoV. The incubation period of the virus has been reported to be 2 10 days. Most SARS patients report a history of prodromal illness (headache, malaise, myalgias) with onset at a mean of 6 days after exposure. Nearly all patients develop a high fever at a mean of 7 days after exposure, although the time delay can be as long as 16 days (8). Other common symptoms include dry cough, dyspnea, and arthralgia (9). Approximately 50% have mild hypoxemia (a blood oxygen saturation level 95%) during the course of the illness (9), and 10% 20% require intubation and mechanical ventilation (6,8). Asymptomatic cases, which are infrequent, have occurred among exposed health-care workers (10). Common clinical symptoms (6,8,11) reported at the time of admission are presented in Table 2. Laboratory Indices Frequently observed hematologic abnormalities in SARS patients include thrombocytopenia and lymphopenia (6,8). Hemolytic anemia has been observed, but it was probably caused by ribavirin treatment (6,9). Observed biochemical abnormalities in blood serum have included elevated levels of aminotransferases, lactate dehydrogenase, and creatine kinase, as well as reduced levels of calcium, magnesium, phosphate, and potassium (Table 3). However, no single hematologic or biochemical marker can be used to distinguish SARS from community-acquired pneumonia caused by other agents (12).

RG f Volume 24 Number 2 Paul et al 555 Table 2 Symptoms of SARS at the Time of Admission Symptom Booth et al (6) (n 144) Lee et al (8) (n 138) Peiris et al (11) (n 50) Fever 99.3 100 100 Nonproductive cough 69.4 57.3 62 Myalgia 49.3 60.9 54 Dyspnea 41.7 NR 20 Dizziness 4.2 42.8 12 Rhinorrhea 2.1 22.5 24 Note. Data are percentages of patients. NR not reported. Table 3 Laboratory Indices in SARS Patients at Admission Finding Booth et al (6) (n 144) Lee et al (8) (n 138) Peiris et al (11) (n 50) Hemoglobin (g/dl) NR 13.5 12.9 White cells (10 9 /L) 5.2 5.1 5.17 Neutrophils (10 9 /L) 3.6 3.9 NR Lymphocytes (10 9 /L) 0.9 0.9 0.78 Lactate dehydrogenase (U/L) 396 NR* NR Creatine kinase (U/L) 222 in men, 95 in women NR 244 Alanine aminotransferase (U/L) 29 60 63 Note. Data are mean values. NR not reported. *Level was elevated in 71% of patients. Level was elevated in 32% of patients. Criteria for the confirmation of SARS at laboratory analysis include detection of antibodies in a convalescent-phase blood serum sample, detection of SARS-CoV in a clinical specimen with reverse transcriptase polymerase chain reaction, or isolation of SARS-CoV in a cultured clinical specimen (13). SARS may be excluded if a convalescent-phase blood serum sample, obtained more than 28 days after symptom onset, is found negative for antibodies to SARS-CoV (14). Treatment There are no scientifically proven therapies for the treatment of SARS. On admission to the hospital, patients have been treated with broad-spectrum antibiotics to cover the common pathogens that cause community-acquired pneumonia (6,8). During the initial outbreak of SARS in Hong Kong, patients were given antiviral therapy (oseltamivir) for possible influenza virus infection. If fever persisted for more than 48 hours, ribavirin and corticosteroids were administered (8). Almost half (49%) of the patients who received ribavirin experienced hemolysis associated with a drop in hemoglobin of at least 2 g/dl (6). Corticosteroids were administered on the basis of similarity in computed tomographic (CT) findings between patients with SARS and patients with bronchiolitis obliterans organizing pneumonia, or BOOP (8). Preliminary results indicate that interferon also may be of benefit in the treatment of SARS (15). Histopathologic Analysis The following description of the histopathologic features of SARS is based on findings in tissue specimens obtained at autopsy in four patients. Pulmonary tissue samples taken from the upper lobes of both lungs in all four patients tested positive at reverse transcriptase polymerase chain reaction analysis for SARS-CoV. The course of the disease in all patients was similar although the

556 March-April 2004 RG f Volume 24 Number 2 Figure 1. (a) Photomicrograph (original magnification, 100; hematoxylin-eosin stain) of a lung tissue specimen from a SARS patient shows early histopathologic changes, including hyaline membrane disease (solid arrows) and hemorrhage (open arrow). (b) Photomicrograph (original magnification, 250; hematoxylin-eosin stain) of a lung tissue specimen shows late histopathologic changes, including organizing alveolar exudate (arrows) and diffuse alveolar damage. Table 4 Radiologic Pattern on Initial Chest Radiographs Findings at Initial Chest Radiography Wong et al (16) (n 138) Grinblat et al (17) (n 40) Paul et al (18) (n 51) Normal 30 (21.7) 17 (42.5) 10 (19.6) Focal opacity 59 (42.8) 20 (50.0) 20 (39.2) Multifocal or diffuse opacity 49 (35.5) 3 (7.5) 21 (41.2) Note. Data are numbers of patients, with percentages in parentheses. severity of disease varied. At gross examination, the lungs appeared congested, showed areas of hemorrhage, and felt more solid than is normal. At histologic analysis, the contents of respiratory secretions showed evidence of hyaline membrane disease, congestion, and focal alveolar hemorrhage (Fig 1). Tissue from patients who had survived 10 days or more with SARS (n 3) showed pneumocytes and macrophages in the alveoli. No viral cytopathic effects were seen. These changes are consistent with diffuse alveolar damage and were seen bilaterally. Clinical Outcome The global fatality rate for SARS is 11% (3). One fifth of all SARS patients require admission to the intensive care unit, and most in this group require mechanical ventilation (6,8). A poor clinical outcome is associated with increased age ( 60 years), male sex, and medical comorbidity. Laboratory parameters predictive of a poor outcome include elevated blood serum levels of neutrophils, creatine kinase, and lactate dehydrogenase (6,8). Radiologic Procedures The radiologic images presented in this article were obtained with computed radiography, conventional CT, or low-dose CT. The CT examinations were performed with a four-row multidetector scanner (LightSpeed; GE Medical Systems, Milwaukee, Wis). The following parameters were used for low-dose CT: 140 kvp, 50 ma, 2.5-mm collimation, 0.8-second gantry rotation, and a table speed of 15 mm/sec. Conventional CT was performed with 120 kvp, 160 200 ma, 5-mm collimation, 0.8-second gantry rotation, and table speed of 11.25 mm/sec. The technologists who performed radiography or CT of patients with suspected SARS were required to wear protective garments in accordance

RG f Volume 24 Number 2 Paul et al 557 Figure 2. SARS in a 25-year-old female healthcare worker with mild symptoms and normal findings on the chest radiograph obtained at admission. (a) Axial low-dose CT image obtained at admission depicts a small focus of consolidation in the left lower lobe. (b) Axial image obtained 7 days after admission shows that the area of consolidation has become less dense but larger, with an increase in ground-glass attenuation. (c, d) Axial images obtained 4 days later show central clearing but increased peripheral ground-glass attenuation (c) and a new focus of consolidation in the left upper lobe (d). with guidelines from infection control specialists. Each technologist wore double gowns, double gloves, a disposable head covering, a face mask with a rating of at least N95 (Moldex-Metric, Culver City, Calif), disposable booties, and protective eyewear and face shields. The patient also was required to wear a mask with an N95 rating. All surfaces in contact with or within 1 meter of the patient were cleaned with an antiviral agent (Virox 5; Virox Technologies, Ontario, Canada). The cleaning procedure was performed twice, and the CT suites were not used for 30 45 minutes after the second cleaning to allow for several air exchanges prior to the entry of the next patient. Radiologic Pattern of Disease This section provides a pictorial review of our experience with 54 patients for whom the diagnosis was probable SARS. The pattern of disease seen in our patient population is similar to that reported in other studies (16,17). Radiologic images from patients with probable SARS may be normal in appearance or may show focal, multifocal, or diffuse consolidation (Table 4). Normal Chest Radiographs In approximately 20% of patients, the lungs appear normal on the initial chest radiograph at presentation (Figs 2, 3), while high-resolution CT, conventional CT, and low-dose CT images depict airspace disease (19,20). Most of these patients (66.7% 70.6%) develop bilateral airspace consolidation, as demonstrated by multifocal opacities or multiple areas of attenuation on subsequent chest images. The radiologic features of the disease are most evident at a mean interval of 1 week after presentation (range, 1 22 days) (17,18). In two thirds of patients, chest radiographs obtained at a mean interval of 12 days after presentation appear normal.

558 March-April 2004 RG f Volume 24 Number 2 Figure 3. SARS in a 38- year-old male healthcare worker in whom initial chest radiography at admission depicted no abnormality. (a) Axial low-dose CT image obtained on the 2nd day after admission shows extensive ground-glass attenuation with interstitial thickening ( crazy paving ) in the left upper lobe. (b, c) Axial images obtained 4 days after admission show interval central clearing (b) and multifocal extension of crazy paving to the periphery and to other lobes (c). Focal Opacity Focal opacity is the most common abnormality seen on initial chest radiographs and is observed in 40% of patients. Focal opacity is detected predominantly in the periphery and the middle and lower zones of the lung. The difference in attenuation between a focal opacity and the surrounding lung tissue is often subtle on chest radiographs, particularly on portable radiographs, and it may be difficult to distinguish focal opacities from breast shadows in women (Fig 4). These areas may appear as regions of extensive consolidation (Fig 5), but more often they are small and relatively dense (Fig 6). Consolidation peaks in severity at a mean of 6 days after presentation (range, 1 25 days). After this, the opacity becomes larger and less dense, and there is resolution to a normal appearance on chest radiographs in 15 (75%) of 20 patients at a mean interval of 16.3 days after presentation. Figures 5, 6. (5) SARS in a 42-year-old woman who presented 9 days after exposure. Posteroanterior radiograph shows extensive consolidation in the left lower lobe. (6) SARS in a 24-year-old woman who presented 10 days after exposure. (a) Radiograph obtained at admission shows focal airspace consolidation in the right lower lobe. (b) Radiograph obtained 1 day later shows that the consolidation has become less dense centrally but more extensive. (c) Radiograph obtained 8 days after admission shows that the consolidation has decreased in size and density. Findings were normal on the chest radiograph obtained 10 days after admission.

RG f Volume 24 Number 2 Paul et al 559 Figure 4. SARS in a 29-year-old woman who presented 7 days after exposure. (a) Posteroanterior radiograph depicts a subtle focus of consolidation in the right lower zone, partly obscured by breast tissue. (b) Posteroanterior radiograph obtained 5 days later shows that the consolidation has expanded and become more dense. The chest radiograph obtained 13 days after admission was normal.

560 March-April 2004 RG f Volume 24 Number 2 Figure 7. SARS in a 46-year-old woman who presented 5 days after developing symptoms. (a) Anteroposterior radiograph shows bilateral multifocal opacities, which are more extensive in the left lung. (b) Anteroposterior radiograph obtained 12 days after admission shows a resolution of central airspace consolidation and residual peripheral consolidation. The patient was asymptomatic. (c e) Axial low-dose CT images obtained on the same day as b show air bronchograms (c), multiple high-attenuation foci (d), and peripheral subpleural areas of high attenuation in the lower lobes (e). The extent of disease was underestimated on b.

RG f Volume 24 Number 2 Paul et al 561 Figure 8. SARS in a 74-year-old man who developed symptoms 4 days after exposure. (a) Initial anteroposterior chest radiograph shows bilateral airspace disease that is more extensive in the left lung. (b) Anteroposterior radiograph obtained 5 days later shows a resolution of consolidation in the left lung but increased consolidation in the right lung. (c) Anteroposterior radiograph obtained 1 day later shows diffuse persistent bilateral airspace disease. The patient died 13 days after exposure to SARS. which the opacities fluctuate in appearance and anatomic location over time. Chest radiographs from these patients show a slow resolution of consolidation; the appearance of the lungs remains abnormal in most (69%, 9 of 13) at a mean follow-up interval of 9.8 days. Multifocal Opacities This is the second most common pattern on initial radiographs and is seen in 27% of patients. The multiple areas of airspace consolidation are predominantly peripheral and confined to the middle and lower zones of the lung. The extent of disease is underestimated with chest radiography (Fig 7). A sinusoid pattern is observed in one fourth (23%) of patients, in Diffuse Opacity Diffuse consolidation of the pulmonary airspace and interstitium is seen on chest images obtained in 14% of patients with SARS (Fig 8). Patients in this group typically are older than those in other groups and have significant medical comorbidity and a high fatality rate. All of the documented deaths from SARS have occurred in patients who developed diffuse airspace consolidation.

562 March-April 2004 RG f Volume 24 Number 2 Figure 9. SARS in a 73-year-old woman who presented 17 days after exposure. The chest radiograph obtained on admission (not shown) was normal. (a) Anteroposterior radiograph obtained 2 days after admission shows focal consolidation in the right lung. (b) Anteroposterior radiograph obtained 8 days later shows bilateral multifocal consolidation. (c) Anteroposterior radiograph obtained 8 days later shows diffuse airspace disease. The patient died 36 days after exposure to SARS. In these patients, a progressive deterioration in the radiologic pattern of disease despite medical treatment (Fig 9) is associated with a poor outcome (16). Conclusions A diagnosis of suspected SARS is based on the presence of fever and respiratory symptoms in an individual with a history of exposure to SARS- CoV. A patient who presents with clinical symptoms compatible with SARS and has a history of recent exposure may be suspected of having SARS. Radiographic evidence of airspace consolidation is required for a diagnosis of probable SARS. Although the initial chest radiograph will be normal in 20% of patients, thoracic CT in many of these will depict airspace consolidation. CT can be performed either with a low-dose technique or with a high-resolution protocol to minimize the radiation dose to the patient. Consolidation in many patients initially is small in volume and patchy in distribution, and low-dose helical CT is advantageous for depicting such abnormalities. Chest radiographs obtained in most SARS patients at initial presentation are abnormal, and the disease progression in these patients can be monitored with serial chest radiography. The pattern of disease at presentation has prognostic significance. In young patients, normal findings or focal opacities are most common on initial chest radiographs. Although in 31% of patients in this group a progression to multifocal opacities is observed, the clinical outcome appears favorable (16). Older patients present with more extensive disease, and opacities seen on

RG f Volume 24 Number 2 Paul et al 563 their initial chest radiographs are predominantly multifocal. The clinical course is protracted, but most patients in this group also recover. Many patients in whom initial chest radiographs show a diffuse pattern of disease are older men with an underlying medical condition. This group has the highest fatality (16). Outlook On July 5, 2003, the last country remaining on the WHO list of areas of active local transmission of SARS Taiwan was removed from the list. At the time of this writing, the WHO has not reported any SARS-affected areas. Our goal in this article was to prepare radiologists for future encounters with SARS by describing the radiologic pattern of disease at presentation and throughout the course of illness. However, it is important to understand that the radiologic appearance of SARS is not distinct from that of other diseases that cause lower respiratory tract infection. Therefore, early identification of new cases of SARS will depend partly on the prompt recognition of clusters of febrile respiratory tract illness. The authors hope that this illustrative presentation of the radiologic appearance of SARS will help to familiarize radiologists with the radiologic manifestations of SARS and enable them to become part of a better surveillance and response system. Acknowledgments: We thank our physicians, nursing staff, and medical technologists, whose dedication to patient care during the SARS crises was exemplary. References 1. Ksiazek TG, Erdman D, Goldsmith CS, et al. A novel coronavirus associated with severe acute respiratory syndrome. N Engl J Med 2003; 348: 1953 1966. 2. World Health Organization. Media release. Available at: www.who.int/mediacentre/releases. Accessed August 30, 2003. 3. World Health Organization. Cumulative number of reported probable cases of severe acute respiratory syndrome (SARS). Available at: www.who.int /csr/sarscountry. Accessed June 27, 2003. 4. Farcas GA, Mazzulli T, Willey BM, et al. Fatal severe acute respiratory syndrome is associated with multi-organ involvement by coronavirus (SARS-CoV) [late-breaker abstract K-1315b]. Presented at the 43rd Intersciences Conference on Antimicrobial Agents and Chemotherapy, Chicago, Illinois, September 15, 2003. 5. Poutanen SM, Low DE, Henry B, et al. Identification of severe acute respiratory syndrome in Canada. N Engl J Med 2003; 348:1995 2005. 6. Booth CM, Matukas LM, Tomlinson GA, et al. Clinical features and short-term outcomes of 144 patients with SARS in the greater Toronto area. JAMA 2003; 289:2801 2809. 7. World Health Organization. Case definitions for surveillance of severe acute respiratory syndrome (SARS). Available at: www.who.int/csr/sars /casedefinition. Accessed June 27, 2003. 8. Lee N, Hui D, Wu A, et al. A major outbreak of severe acute respiratory syndrome in Hong Kong. N Engl J Med 2003; 348:1986 1994. 9. Avendano M, Derkach P, Swan S. Clinical course and management of SARS in health care workers in Toronto: a case series. CMAJ 2003; 168:1649 1660. 10. Gold WL, Mederski BM, Rose D, et al. Prevalence of asymptomatic infection with severe acute respiratory syndrome coronavirus (SARS-CoV) in exposed healthcare workers [late-breaker abstract K-1315c]. Presented at the 43rd Intersciences Conference on Antimicrobial Agents and Chemotherapy, Chicago, Illinois, September 15, 2003. 11. Peiris JS, Lai ST, Poon LL, et al. Coronavirus as a possible cause of severe acute respiratory syndrome. Lancet 2003; 361:1319 1325. 12. Muller MP, Tomlinson G, Marrie T, et al. Discriminative ability of laboratory parameters in severe acute respiratory syndrome (SARS) [latebreaker abstract V-796b]. Presented at the 43rd Intersciences Conference on Antimicrobial Agents and Chemotherapy, Chicago, Illinois, September 14, 2003. 13. Centers for Disease Control and Prevention (CDC). Guidelines for laboratory diagnosis of SARS-CoV infection. Available at: www.cdc.gov /ncidod/sars/lab.htm. Accessed July 1, 2003. 14. Centers for Disease Control and Prevention (CDC). Update: severe acute respiratory syndrome worldwide and United States, 2003. MMWR Morb Mortal Wkly Rep 2003; 52:664 665. 15. Loutfy MR, Blatt L, Ward S, et al. Preliminary results on the potential therapeutic benefitofinterferon alfacon-1 plus steroids in severe acute respiratory syndrome [late-breaker abstract- K1315-e]. Presented at the 43rd Intersciences Conference on Antimicrobial Agents and Chemotherapy, Chicago, Illinois, September 15, 2003. 16. Wong KT, Antonia GE, Hui DS, et al. Severe acute respiratory syndrome: radiographic appearances and pattern of progression in 138 patients. Radiology 2003; 228:401 406. 17. Grinblat L, Shulman H, Glickman A, et al. The radiology of SARS: a review of 40 probable cases in Toronto, Canada. Radiology 2003; 228:802 809. 18. Paul N, Chung T, Konen E, et al. The prognostic significance of the radiographic pattern of disease in patients with severe acute respiratory syndrome. AJR Am J Roentgenol (in press). 19. Rao TN, Paul N, Chung T, et al. Value of CT in assessing probable severe acute respiratory syndrome. AJR Am J Roentgenol 2003; 181:317 319. 20. Wong KT, Antonio GE, Hui DS, et al. Thin-section CT of severe acute respiratory syndrome: evaluation of 73 patients exposed to or with the disease. Radiology 2003; 228:810 815.