| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Thoracic Imaging |
1 From the Department of Radiology (S.F.K., T.Y.L., C.C.H., Y.F.C., S.H.N., Y.L.K.), Department of Internal Medicine, Divisions of Pulmonary and Critical Care Medicine (M.C.L.) and Infectious Diseases (J.W.L.), Department of Medical Research (K.D.Y.), Department of Public Health and Biostatistics (M.C.C.), and Department of Surgery (C.L.C.), Chang Gung University, Chang Gung Memorial Hospital at Kaouhsiung, 123 Ta-Pei Rd, Niao-Sung Hsiang, Kaohsiung Hsien 833, Taiwan. Received September 24, 2003; revision requested December 4; revision received December 11; accepted January 30, 2004. Address correspondence to S.F.K. (e-mail: sfatko@adm.cgmh.org.tw).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Radiographic findings were reviewed by two radiologists for 52 patients with SARS. On each radiograph, each lung was separated into upper, middle, and lower zones. A four-point scale was used to score extent of SARS-related lesions in each zone; points from all zones were added for a cumulative score. Patient sex, age, comorbidities, duration of developing lesions, lesion score for each radiograph, need for mechanical ventilation, and percentage of lung affected were compared between patients who died (n = 20) and survivors (n = 32). Continuous and categorical variables were analyzed with Mann-Whitney test and Fisher exact or 
2 test, respectively.
RESULTS: Survival and mortality groups showed no significant differences with respect to patient sex, duration of SARS-related lesions, development of lesion shifting, and acute respiratory distress syndrome. Patients who died were significantly older (mean ± standard deviation, 56.9 years ± 17.2 vs 40.4 years ± 16.6; P = .002) and had higher frequency of comorbid lung illnesses (nine of 20 vs two of 32, P = .001), maximal lesion extent score of 7 or higher (20 of 20 vs five of 32, P < .001), involvement of four or more lung zones (17 of 20 vs four of 32, P < .001), bilateral lung involvement (19 of 20 vs 14 of 32, P < .001), need for mechanical ventilation (18 of 20 vs two of 32, P < .001), and higher percentage of affected areas (41.5% ± 8.6 vs 16.4% ± 10.0, P < .001) than those of survivors.
CONCLUSION: On chest radiographs, maximal SARS-related lesion extent score of 7 or higher is a strong predictor of mortality, especially in patients with comorbid lung illnesses and involvement of four or more lung zones.
© RSNA, 2004
Index terms: Lung, diseases, 60.2069 • Severe acute respiratory syndrome, 60.2069
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
During the worldwide outbreak, imaging studies provided important information for diagnosis, management, and control of SARS (6–22). On chest radiographs or computed tomographic (CT) scans, SARS typically manifests as unilateral or bilateral peripheral air-space disease with rapid progression (14–21). A small percentage of patients has evidence of fibrosis-like patterns on follow-up CT scans and radiographs after discharge from the hospital (22).
Although investigators in two recent studies found that large areas of ground-glass opacity with air bronchogram in both lungs indicated a poor prognosis (20) and that SARS with bilateral consolidation had a more protracted clinical course (21), data on the prognostic implications of the radiographic features of SARS remain limited. Thus, the purpose of our study was to retrospectively assess the prognostic implications of radiographic findings of SARS.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Individuals who had been isolated without fever for 10 days were sent home directly for further isolation at home for at least 10 days, and this isolation was monitored by means of governmental surveillance. Patients who were discharged from the hospital or visitors who had stayed in the hospital within 10 days of the outbreak duration and patients referred from other hospitals who were suspected of having SARS were screened at an outdoor fever triage station in the emergency department. Patients with clinical and laboratory manifestations compatible with SARS were transferred to negative-pressure isolation rooms directly.
Eventually, a total of 52 patients who fulfilled the World Health Organization clinical and epidemiologic criteria for the probable case definition of SARS (13) were identified during the outbreak. Diagnostic criteria included fever with temperature of at least 38°C (except for one immunocompromised patient with a highest temperature of 37.7°C), one or more clinical findings of severe respiratory illness (including cough, shortness of breath, difficulty breathing, and hypoxia), radiographic evidence of new pulmonary abnormality occurring within 10 days of the outbreak period, and (a) a history of exposure to patients known to have SARS or (b) a stay in an area with recent transmission of SARS, especially the chest ward of our hospital.
Of the 52 patients who met the clinical diagnostic criteria, 20 were men (age range, 23–84 years; mean age, 48.1 years), and 32 were women (age range, 21–83 years; mean age, 45.3 years). All 52 were treated by a team of specialists in infectious disease and respiratory and critical care medicine who were assigned to the outbreak. A standard treatment protocol was administered, including combined therapy with antibacterial, antiviral, and corticosteroid agents, as well as ventilation support when necessary (23). The past history, clinical data, disease course, and outcome of the patients were retrospectively reviewed together by a radiologist (S.F.K.) and two physicians (M.C.L., J.W.L.) after chest radiographs were obtained and interpreted. This retrospective study was approved by the institutional review board of our hospital, and no patient informed consent was required.
Acquisition and Interpretation of Chest Radiographs
At least one frontal chest radiograph (all were anteroposterior views except for four posteroanterior and five lateral views) was obtained each day after the onset of fever until marked regressive changes of the SARS-related lung opacities or marked improvement of the clinical conditions were noted. Lateral radiographs were obtained when the patient was suspected of having lesions that were not clearly seen on frontal chest radiographs because of overlapping structures, especially adjacent to the hila, heart, and mediastinum. Subsequently, at least one frontal chest radiograph was obtained every 2–3 days (depending on the clinical condition of a patient) until total resolution of the lung lesions or total recovery of clinical condition.
All radiographic images were obtained and processed with a computed radiographic system (FCR5000Plus; Fuji Photograph Film, Minato-Ku, Tokyo, Japan) with a standardized method (70–75 kV, 3.2 mAs, 100-cm film-focus distance for anteroposterior view; 80–85 kV, 4 mAs, 180-cm film-focus distance for posteroanterior view). Images were archived via a picture archiving and communication system, or PACS (Centricity Workstation, version 1.0; GE Medical Systems, Milwaukee, Wis), and hard copies were provided to the clinicians as requested. None of the patients underwent CT during this outbreak; however, two patients underwent CT 1 month and 1 week prior to this outbreak for the investigation of pleural mesothelioma and lung abscess, respectively.
All hard copies of chest radiographs were retrospectively reviewed by two radiologists (S.F.K., T.Y.L.), each with more than 15 years of experience in reading chest radiographs. The two radiologists worked together in consensus. They were blinded to the clinical information or clinical progress of the patients, except for the knowledge that these were cases of SARS. For each case, the radiographs were filed in chronologic order by filing clerks, and the identities on the chest radiographs were masked.
The radiologists were required to review the images sequentially, starting from the initial radiograph. Each lung was separated into upper, middle, and lower zones, with each zone spanning one-third of the craniocaudal length of the lung as measured on the chest radiograph. The distribution of SARS lesions on each radiograph was then assessed by means of classification into six categories on the basis of how many lung zones were affected (categories I, II, III, IV, V, and VI for one, two, three, four, five, and six lung zones involved, respectively).
The severity of the SARS-related lesions within each lung zone was evaluated by scoring the radiographs with a four-point scale based on visual assessment, as follows: 0 = normal, 1 = up to one-third of lung zone involved, 2 = between one-third and two-thirds of lung zone involved, and 3 = more than two-thirds of lung zone involved. The scores for all lung zones on each radiograph were added to provide a cumulative score, which had a possible range from zero to 18. Serial chest radiographs for each patient were categorized and scored accordingly. Acute respiratory distress syndrome (ARDS) was defined as acute onset of hypoxemia with partial arterial pressure of oxygen to fraction of inspired oxygen (PaO2/FIO2) of less than 200 mm Hg with characteristic rapid opacification of the whole lung and clinical absence of elevated left atrial pressure (24).
Maximal lung involvement was defined as the highest lesion extent score attained before the occurrence of ARDS. In addition, data collected from all available chest radiographs included whether the SARS-related lesions were unilateral or bilateral, had perihilar (lesions mainly distributed in an area approximately 3 cm from the hila) or peripheral (lesions mainly distributed in an area lateral or exterior to the perihilar region) locations, were associated with radiographically identifiable pleural effusion (defined as increased pleural density with obscuration of the costophrenic sinuses and the hemidiaphragm with meniscus-shaped or horizontal upper border), and included the presence of mediastinal or hilar lymphadenopathy (defined as widening or increased opacity of the mediastinum and pulmonary hila) and cavitary lung lesions.
After the scoring and categorization of chest radiographs, the areas of the lung and the SARS-related opacities were measured together by the two radiologists (S.F.K., T.Y.L.) with a high-resolution monitor (MGD 521MK II; Barco View, Kortrijk, Belgium) via the PACS system. By using an electronic cursor, the apical and lateral margins of the right and left lungs were traced along interfaces between the inner margin of the ribs and the outer shadow of the interspersed lung.
The inferior margins of the right and left lung were arbitrarily defined as a line drawn between the right costophrenic angle and 1 cm below the right pericardiophrenic junction and as a line drawn between the left costophrenic angle to the left diaphragmaticoaortic junction, respectively. The right medial margin was arbitrarily defined as a line drawn along the right paraspinal interface and right paratracheal stripe. The left medial margin was arbitrarily defined as a line drawn along the lateral margin of the descending thoracic aorta, aortic knob, and left paratracheal area. The total areas of the right and left lung were automatically calculated by the PACS system. Subsequently, the margins of the SARS-related lung abnormalities were outlined one by one, and the areas of regions of interest were calculated. The percentage of SARS-related lung involvement on each radiograph was calculated by dividing the areas of the regions of interest by the total areas of the right and left lungs and then multiplying the result by 100%.
Data and Statistical Analysis
Clinical data of all patients, including age, sex, presence of comorbid lung illnesses, duration from the onset of fever to the emergence of pneumonic lesions, duration from the onset of fever to the development of a maximal SARS-related lesion extent score of 7 or higher, maximal lung involvement, need for mechanical ventilation, and outcome, were recorded. Associated changes, including shifting of radiologic lesions (defined as evident improvement of the original lesion followed by the appearance of new lesions) and development of ARDS during the treatment course, were also documented.
Patients were separated into mortality (20 patients) and survival (32 patients) groups. Radiologic findings were analyzed and compared between these two groups by using the Mann-Whitney test for continuous variables and the Fisher exact or 2 test for categorical variables. Statistical analysis was performed with SYSTAT software version 11.0 (SPSS, Chicago, Ill). A P value less than .05 was considered to indicate a statistically significant difference. The correlation between lesion scores and the percentages of affected lung areas on chest radiographs at maximal involvement by SARS was determined by using Spearman rank correlation.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Eleven of 52 patients had various documented lung diseases, including chronic obstructive pulmonary disease (n = 5), bronchogenic carcinoma (n = 2), lung abscess after treatment with nearly total recovery (n = 1), thymic carcinoma (n = 1), bronchogenic carcinoma after resection (n = 1), and pleural mesothelioma (n = 1). Clinical symptoms included fever (n = 52) (range, 37.7°–40.5°C), cough (n = 36), dyspnea (n = 25), headache (n = 24), chill (n = 22), myalgia (n = 19), and diarrhea (n = 6). During the outbreak, 20 patients underwent mechanical ventilation because of respiratory failure and PaO2/FIO2 of less than 200 mm Hg. Among them, 18 patients died, and two survived. Two patients who met the criteria for mechanical ventilation died before endotracheal intubations could be performed.
The mortality rate in this study was 38% (20 of 52 patients). The mean age of the mortality group (56.9 years ± 17.2 [standard deviation]) was significantly older than that of the survival group (40.4 years ± 16.6) (P = .002).
Radiographic Findings
On the first chest radiograph obtained after the onset of fever, 30 of 52 patients showed pulmonary parenchymal abnormalities or SARS-related lesions (including hazy opacities with a ground-glass appearance in six patients and air-space disease in 24 patients). All patients showed development and progression of SARS-related lesions as single or multiple areas of air-space disease on follow-up chest radiographs (Figs 1–3). None of the lesions showed cavitary changes.
|
|
|
|
|
|
The initial and maximal distributions and categories of the SARS-related lesions noted on chest radiographs are shown in Tables 1 and 2. The lower lung zones were affected most frequently (particularly the lower right zone), followed by the upper, right middle, and left middle lung zones. A ratio of peripheral to perihilar lesions of 60:29 was noted on the first radiograph obtained after the onset of fever. Initially, 41 of 52 (79%) patients were classified in categories I and II, with a lesion extent score ranging from 1 to 6. Two of fifty-two (3.8%) patients had an initial lesion extent score of 7 or higher, and both died. Fourteen of fifty-two (27%) patients showed initial bilateral lung involvement, and nine of them died.
|
|
The duration from maximal lung involvement by SARS-related lesions to mortality ranged from 1 to 14 days (mean, 3.4 days ± 3.3). Among the 32 patients who recovered, six showed shifting of SARS-related lesions on chest radiographs (Fig 4). Six patients developed ARDS during treatment, and among them, one developed spontaneous pneumothorax, two had pneumomediastinum, and four eventually died. No apparent pleural effusion or enlarged mediastinal lymph nodes were found in 50 patients, while two patients had coexistent malignant pleural mesothelioma and empyema.
|
|
|
|
|
= 0.964, P < .001) at the maximal stage. Table 3 shows the comparison of patient age, sex, comorbid lung illnesses, and various radiographic features between mortality and survival groups. There were no significant differences between the two groups with respect to patient sex, presence of abnormalities on the first chest radiograph obtained, duration from onset of fever to identification of initial lesions on chest radiographs, duration from onset of fever to development of a lesion extent score of 7 or higher and to maximal lung involvement, the development of shifting of SARS-related lesions, and the development of ARDS.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The histologic characterization of SARS is that of an acute lung injury (26–28). If the injury has an accelerated course and involves both lungs in a rapid sequence that leads to opacification of the lungs at chest radiography, which is associated with severe respiratory distress, the clinical picture is indicative of ARDS. In addition, SARS may have systemic manifestations, including vasculitis, venous thrombosis, and necrosis of the spleen and lymph nodes, liver, heart, and adrenal glands (27). Although bronchitis was present in only one autopsy reported by Nicholls et al (26), it has been reported in other cases (27).
Shifting or fluctuating radiologic changes and an inverted V pattern in viral-load study of nasopharyngeal aspirates (increased viral load from day 5 to peak at day 10, followed by decreased viral load from day 10 to day 15) have also been reported, which suggests that worsening of the disease in the 2nd week may be related to immunopathologic damage (29).
Data obtained from investigations on the clinical and epidemiologic characteristics of SARS facilitated efforts to control the outbreak through an emphasis on immediate isolation of new cases, thereby reducing the risk of further transmission (6–12,25,30,31). The initial clinical features of our 52 cases, including fever, nonproductive cough, myalgia, dyspnea, and hematologic findings, were similar to those reported in several recent large-cohort studies from Hong Kong and Canada (6–10). The mortality rate in our series, however, with 20 of 52 patients dying within 3 weeks of onset, is higher than in those reports, in which mortality rates ranged from 4% to 10% (6–12,23,25).
This high mortality rate might have been related to a variety of factors, including the high proportion of elderly patients in the chest ward, the presence of comorbid thoracic malignancies and chronic lung diseases, frequent use of nebulized bronchodilators with resultant dissemination of droplets of high viral load, close contact of caregivers with patients, and unlimited visiting hours and number of visitors, with frequent crowding of the ward. These factors have all been documented in association with extensive dissemination and poor outcomes (4–10,14,25,30,31).
Fortunately, after relocation of all patients suspected of having or classified as probably having SARS to the negative-pressure isolation rooms on the top floor of our hospital, the institution of strict isolation regulations, and application of all available protective measures, no new cases of in-hospital infection occurred. Studies from Canada and Hong Kong showed a poorer prognosis in men (9,10), but our study showed no significant difference with regard to patient sex and mortality rate.
Chest radiography offers important clues to the diagnosis of SARS and is one of the clinical criteria in the World Health Organization case definition of SARS (6–18). The radiographic findings in our study are consistent with those in other series (8–10,13–21). Initial SARS lesions typically manifested as lung opacities that mostly occurred in the lower lung and had peripheral involvement. Rapid worsening of the lesions and progression to multifocal involvement were noted frequently. Furthermore, cavitation of the lesions and marked adenopathy were absent in our patients, which is consistent with prior reports of CT findings of SARS.
Although small pleural effusion has been described in some reports (15–22), except for two patients with coexistent pleural mesothelioma and empyema in our study, no apparent pleural effusion was found in the other 50 patients. Wong et al (16) identified four patterns of radiographic progression of SARS lesions in a series of 138 cases. An initial radiographic deterioration to peak level, followed by radiographic improvement, was the most common pattern in their study (pattern I, 70.3%), followed by fluctuating changes (pattern II, 17.4%), static radiographic appearance (pattern III, 7.3%), and progressive radiographic deterioration (pattern IV, 5.1%). None of our patients had a static radiographic pattern, but 20 patients (38%) showed progressive radiographic deterioration with eventual mortality. These patients were exposed to the highly contaminated environment in a chest ward before strict isolation was undertaken, which might have led to a bias toward inclusion of patients with severe SARS in the present study.
An epidemiologic study of SARS in Hong Kong showed that the clinical outcome of patients was not associated with the duration between onset of clinical symptoms and admission (23). We found no relationship between outcome and the duration from onset of fever to initial radiographic identification of SARS-related lesions or between outcome and duration from onset of fever to maximal lung involvement, the development of shifting of SARS-related lesions, and the development of ARDS. Therefore, monitoring of the radiographic progressive pattern of SARS, as described by Wong et al (16), does not appear to provide any prognostic implication. On the other hand, Ooi et al (32,33) reported that the severity of lung abnormalities quantified on chest radiographs correlates with clinical and laboratory parameters, oxygen supplementation, and treatment response of SARS.
In the present study, 27 patients with a maximal lesion extent score lower than 7 survived. All of these patients belonged to lesion distribution categories I–III, and none of them received mechanical ventilation. This might suggest that most patients can tolerate the disease when fewer than three lung zones are affected. Prognosis has been reported to be correlated with the disease extent of ARDS. Also, patients who died tended to have more consolidation and asymmetric disease (34), while patients with SARS who had bilateral consolidations had a more protracted clinical course (21,32,33).
In the present study, once the lesion extent score reached a value of 7 or higher, mortality was very high, especially for patients with lesion distributions in categories IV–VI—that is, with involvement of four or more lung zones. Furthermore, bilateral lung involvement by SARS is also indicative of wider distribution of the viral-load droplets, and bilateral disease with a score of 7 or higher had an unfavorable prognosis. On the other hand, for the four patients in category III who had a lesion extent score of 7 or higher, three of them had comorbid thoracic malignancies that might have further affected residual lung capacity and the risk of resultant mortality.
There are several limitations inherent in a study of SARS based solely on chest radiographs. The clinical conditions of the patients in the present study varied widely, and radiographs were obtained with patients in various positions (upright, supine, and semierect). It may not be possible to quantitatively account for the effects of these differences. Furthermore, preexisting lung diseases, such as lung malignancies, pulmonary edema, and atelectasis, might also affect the visual estimation of the extent of SARS.
During an in-hospital SARS outbreak, however, chest radiography seems to be the most practical imaging tool in real life (16). In fact, to obtain just a chest radiograph, the radiographer had to spend 20–30 minutes to put on and remove the full protective measures before and after going into the negative-pressure isolation rooms. In addition, the radiographic cassette had to be protected by several layers of plastic bags and had to be unwrapped in a stepwise manner before being removed from the isolation room as a measure of preventing SARS contamination.
The measurement of areas of lung opacity from a two-dimensional image did not precisely represent the true volume of the lesions. Nevertheless, during an epidemic crisis in clinical practice, a volumetric appraisal of lung damage may not be possible (16). We tried to standardize the measurement method by using the ratio or percentage of involved areas measured on serial chest radiographs. A similar measurement method has been reported in monitoring the response to treatment in patients with SARS (23).
In the present study, radiographic evidence of involvement by SARS lesions in more than 40%–50% of the lung area was noted in the mortality group. Our results also showed a strong positive correlation ( = 0.964) of prognosis with the lesion extent score and the percentage area of lung involvement in both the mortality and survival groups. From a practical point of view, a detailed measurement of each lesion on every chest radiograph may not be necessary, and a technique such as the proposed simple scoring method might be more practical for assessment of prognosis.
For patients with SARS, mechanical ventilation is an important treatment when patients experience respiratory failure (23). In our study, mechanical ventilation is a strong predictor of mortality, which reflects the severity of respiratory failure as a surrogate for radiographic findings. Although the decision to intubate is solely clinical, the radiographic findings do, however, warn the clinician of approaching clinical deterioration, especially for patients in categories IV–VI with a lesion extent score of 7 or higher.
The CT manifestations of SARS include peripherally predominant ground-glass opacification, inter- and intralobular interstitial lesions, and unilateral and bilateral consolidations (6–10,14–22). When patients show high spiking fever and lymphopenia while chest radiographs appear normal, CT is a sensitive imaging modality for diagnosis (8,14,17–19). Because of the extremely contagious nature and high fatality rate of SARS, however, concern for cross-contamination and strict isolation measures have limited the use of CT for suspected cases (14,25,31,35). Under these circumstances, close radiographic follow-up is mandatory and probably adequate, while CT is useful in the demonstration of the parenchymal changes and findings suggestive of pulmonary fibrosis caused by SARS after discharge from the hospital (22).
In summary, in addition to the value of chest radiographs in monitoring of disease progression and treatment response, the prognostic implications of SARS can be derived from chest radiographs by means of a simple scoring method. A maximal SARS-related lesion extent score lower than 7 is indicative of good prognosis. A score of 7 or higher is a strong predictor of mortality, especially in patients with comorbid lung illnesses and involvement of more than four lung zones.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, S.F.K.; study concepts, S.F.K., C.C.H., Y.F.C., K.D.Y., C.L.C.; study design, S.F.K., C.C.H., Y.F.C., Y.L.K., T.Y.L.; literature research, S.F.K., C.C.H., Y.L.K.; clinical studies, S.F.K., M.C.L., J.W.L., T.Y.L.; data acquisition, S.F.K., C.C.H., Y.F.C., Y.L.K., T.Y.L.; data analysis/interpretation, S.F.K., C.C.H., S.H.N., T.Y.L., M.C.C., K.D.Y.; statistical analysis, S.F.K., C.C.H., M.C.C.; manuscript preparation, S.F.K., C.C.H.; manuscript definition of intellectual content, S.F.K., Y.F.C., T.Y.L., C.L.C.; manuscript editing, S.F.K., S.H.N., C.L.C.; manuscript revision/review, S.F.K., S.H.N., K.D.Y.; manuscript final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. J. Cowling, M. P. Muller, I. O. L. Wong, L.-M. Ho, S.-V. Lo, T. Tsang, T. H. Lam, M. Louie, and G. M. Leung Clinical prognostic rules for severe acute respiratory syndrome in low- and high-resource settings. Arch Intern Med, July 24, 2006; 166(14): 1505 - 1511. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
