| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, University of Bonn, Bonn, Germany.
2
Philips Medical Systems, Eindhoven, The Netherlands.
3
Department of Internal Medicine, University of Bonn, 53127 Bonn,
Germany.
4
Department of Pathology, University of Bonn, 53127 Bonn, Germany.
Received August 9, 1999;
accepted after revision January 4, 2000.
Address correspondence to C. C. Leutner, Radiologische Klinik,
Universität Bonn, Sigmund-Freud-Str. 25, 53127
Bonn, Germany.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Sixteen patients with pneumonia shown on helical CT were examined using MR imaging within 2 days. MR examinations were performed on a 1.5-T system with a transversal T2-weighted ultrashort turbo spin-echo sequence using expiratory gating and diastolic triggering. Two radiologists reviewed the MR and CT images independently. The number, localization, and morphology of lesions were noted. MR image quality was rated using a 4-point scale.
RESULTS. The results of the CT and MR examinations concerning the number and morphology of pulmonary lesions caused by pneumonia were identical in 75% of the patients (n = 12). MR imaging was able to depict all typical features of pneumonia including different stages of parenchymal infiltration (ground-glass versus consolidation). MR imaging depicted early necrotizing pneumonia not shown on contrast-enhanced CT in 25% of the patients (n = 4); 82% of the MR examinations were rated as excellent (1 point) or good (2 points).
CONCLUSION. T2-weighted turbo spin-echo imaging is able to depict characteristic features of pneumonia and shows excellent results compared with CT. This MR technique offers advantages in patients with pneumonia because of its higher sensitivity for necrotizing pneumonia.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Despite the increasing interest in lung MR imaging, only a few studies compare MR imaging of the lung parenchyma with helical CT [8, 10, 11]. To our knowledge, no systematic study of MR imaging in patients with pneumonia has been published until now, although the evaluation of MR imaging as an alternative to CT in the diagnosis of pneumonia is of considerable clinical interest: an increasing number of CT examinations are performed to rule out pneumonia in immunocompromised patients because studies indicate that the early diagnosis of opportunistic pneumonia leads to a significant change in patient treatment [12, 13]. Therefore, patients at risk for opportunistic pneumonia may undergo repetitive CT examinations during the course of illness. The aim of our study was to find out how MR imaging compares with CT regarding the depiction of typical features of pneumonia and the detectability of lesions. For this reason MR imaging and CT were compared in immunocompromised patients with typical findings of pneumonia in the initially performed CT examination.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Unstable patients, minors, and patients with standard medical contraindications to MR examination (e.g., a cardiac pacemaker) were excluded. Patients in whom the CT and MR examination could not be performed within 2 days were excluded too. This time limit was chosen to avoid divergent results of MR and CT examinations caused by rapid changes of pulmonary infectious lesions during therapy. Because of these limitations 19 patients had to be excluded. Sixteen patients could be included in the study (six women and 10 men; age range, 18-71 years; mean age, 46 years). Informed consent was required from all patients. The study was approved by our internal review board.
The diagnosis of pneumonia was established by the typical clinical findings (e.g., fever, cough) and the pathologic findings of the CT examination. The initial chest radiograph in these patients showed normal findings in five patients and was suggestive of early pneumonia in 11 patients. In eight patients, a definite diagnosis was made using bronchoscopy or biopsy. Two patients died and the diagnosis was established at autopsy. In four patients, a bacterial infection was assumed because pneumonia resolved completely under antibacterial therapy. In two patients, the causal infectious organism remained unknown despite percutaneous or open lung biopsy. However, clinical and radiologic findings were highly suggestive of an invasive pulmonary aspergillosis. The following infectious pathogens were found: Aspergillus species in five patients, Mycobacterium tuberculosis in two patients, Pneumocystis carinii in two patients, Geotrichum capitatum in one patient. All patients were immunocompromised because of leukemia (n = 11), AIDS (n = 4), or malnutrition (n = 1).
All surviving patients had clinical and radiologic follow-up during antibacterial or antifungal treatment. All patients including those with normal findings on a chest radiograph showed typical findings of pneumonia on the chest radiograph during the follow-up interval. Both the symptoms of pneumonia and the pathologic findings on the CT scan or chest radiograph disappeared completely under antimicrobial treatment in all surviving patients. Therefore, the pathologic findings described in this article can definitely be assumed to have been caused by infectious pneumonia.
MR Examinations
MR studies were performed with a 1.5-T MR imager (NT Powertrak 1000;
Philips Medical Systems, Eindhoven, The Netherlands) using a body coil or a
dedicated surface coil (Synergy Coil; Philips Medical Systems). The imaging
protocol consisted of a transversal T2-weighted ultrashort turbo spin-echo
sequence, which is the standard T2-weighted sequence for lung MR imaging in
our department. The ultrashort turbo spin-echo sequence is a modified turbo
spin-echo sequence with an increased bandwidth of the frequency-encoding
gradient. It allows the application of high turbofactors, thus leading to
short echo-spacing (7.5 msec). To compensate for respiratory and cardiac
motion, image acquisition was performed during expiration (expiratory gating)
and during the diastolic heart phase (peripheral pulse measurement on the
right hand of the patient; 200-msec trigger delay to the peak of the pulse
wave). Depending on the respiration frequency of the patient, the effective TR
ranged from 2000 to 4000 msec and the TE was 90 msec. Twentyfour slices with a
slice thickness of 6 mm and an interslice gap of 0.6 mm were acquired (field
of view, 350 mm; matrix size, 256 x 192; number of excitations, six).
Spatial presaturation of the chest wall tissue was used. The average scan time
for this sequence was 10 min (multislice acquisition).
CT Examinations
All CT examinations were helical (Somatom 4A; Siemens, Erlangen, Germany).
Twelve CT examinations were contrast-enhanced. In four patients, unenhanced CT
was performed because the patient had standard contraindications to contrast
agents (e.g., renal failure). The helical CT scans were acquired with a
collimation of 8 mm, a longitudinal table speed of 8 mm/sec, and image
reconstruction intervals of 8 mm. The images were viewed at window settings
for lung parenchyma and mediastinum.
Image Analysis
Two observers experienced in chest CT and MR imaging reviewed the CT and MR
images separately. A decision for each MR or CT examination was reached by
consensus. The reviewers were asked to assess the presence, number, and
location of pulmonary infiltrations or other pulmonary pathologic findings
including infiltrations of lung parenchyma-like consolidations or ground-glass
infiltrations, nodular infiltrations, reticular infiltrations, cystic disease
and cavitation, or bullae. According to the standardized nomenclature defined
for parenchymal findings in CT, consolidation was defined as a homogeneous
increase in lung parenchyma attenuation (CT) or signal intensity (MR imaging)
that obscures the margins of vessels and airway walls. Ground-glass
infiltration was defined as an opacity (CT) or hyperintensity (MR imaging) not
obscuring bronchovascular margins
[14]. A systematic study of
the signal intensity of the MR findings in comparison with that of other
anatomic structures, such as the fatty tissue of the chest wall, was not
performed. However, the reviewers were asked to note signal variations within
lesions—for example, lesions consisting of a hypointense rim in a
hyperintense infiltration of the lung parenchyma. This finding has been
described in MR studies of various anatomic regions and has been called the
"reverse-target" sign
[10,
15,16,17,18].
This term also will be used in our study. A reverse-target appearance is a
strong indicator for a necrotizing lesion or an abscess according to the
literature [10,
15,16,17,18].
For the contrast-enhanced CT scans, the following criteria for image analysis
were applied: solid enhancement; rimlike enhancement, indicating necrotizing
pneumonia; and no enhancement.
The consensus findings of the MR and CT examinations were compared concerning the morphology, the location, and the number of lesions. Findings were rated for each patient as identical or divergent. In the divergent cases, MR imaging and CT examinations as well as follow-up examinations and histologic proof were reviewed to determine which imaging method was correct.
The MR images were assessed for quality (1, excellent [no artifacts]; 2, good [few artifacts]; 3, moderate [of diagnostic value but impaired by artifacts]; 4, poor [of no diagnostic value]). The causes of impaired quality were attributed to ghosting artifacts, motion artifacts, patient movement artifacts, or a combination thereof.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
The image quality of the transversal T2-weighted ultrashort turbo spin-echo images was rated as excellent in six patients (37%), good in seven patients (44%), and moderate in three patients (19%). The causes for reduced image quality were artifacts resulting from cardiac arrhythmia (n = 1) or irregular respiration in a patient with claustrophobia (n = 2).
Correlation with CT Examinations
In all patients, both MR and CT examinations showed identical results
concerning the number, the location, and morphology of parenchymal
infiltrations caused by pneumonia. In one patient, CT showed additional
findings (bullae) not related to pneumonia. MR imaging indicated necrotizing
pneumonia in seven patients (44%) because of the reverse-target sign. Only one
case had typical findings of abscess formation seen on the corresponding CT
examination (Table 1). In six
patients (38%), CT did not reveal any sign of a necrotizing pneumonia
(contrast-enhanced CT in four patients and unenhanced CT in two patients).
Overall, more than 16 lesions with a reversetarget appearance on MR images did
not show typical signs of a necrotizing pneumonia on the corresponding CT
scans (including one patient with multiple lesions).
To answer the question of which imaging technique enabled correct diagnosis of necrotizing pneumonia, we took a look at follow-up examinations with CT (four cases) or at autopsy results (two cases). We considered the MR diagnosis correct in all cases in which the typical signs of necrotizing pneumonia in the questionable lesions were depicted on follow-up CT (e.g., rimlike enhancement, cavitation, and air crescents) or in which histologic proof was shown at autopsy. On the basis of these conditions, the MR diagnosis was confirmed as correct in four cases (25%). Two cases were rated as indeterminate because the initial CT examination was unenhanced, thus not allowing early diagnosis of necrotizing pneumonia. However, in one of these cases the diagnosis of necrotizing pneumonia was confirmed in three of four lesions on follow-up CT performed after 2 weeks of antifungal treatment. In the other case, a definite confirmation could not be established because the follow-up CT examination was unenhanced. None of the patients without the reverse-target sign developed necrotizing pneumonia within the follow-up time interval until complete resolvement of pneumonia.
In conclusion, 75% of the MR and CT examinations were rated as showing identical results regarding lesions caused by pneumonia. This included the two cases with suspected necrotizing pneumonia that were rated as indeterminate because definite proof of necrotizing pneumonia was not available. Twenty-five percent of the MR examinations were rated as showing more accurate results than CT because of a higher sensitivity for revealing necrotizing lesions.
Comparison of CT and MR Images with Chest Radiographs
Although it was not the major aim of our study to compare chest radiographs
with CT or MR images, the following results underline findings of previously
published studies. In five patients with normal findings on chest radiographs,
CT and MR examinations showed ground-glass hyperintensities, nodular
infiltrations, or both (three cases each); consolidations (two cases); or a
reticular pattern (two cases). In the remaining 11 patients with chest
radiographs suggestive of early pneumonia, both CT and MR imaging showed more
lesions than conventional radiographs in nine patients, as shown in Figure
2A,2B,2C.
In three patients, chest radiographs showed nonspecific findings, whereas CT
and MR imaging clearly showed nodular infiltrations highly suggestive of
invasive pulmonary aspergillosis, as shown in Figure
3A,3B,3C,3D.
|
|
|
|
|
|
|
Pathohistologic Correlation
In two patients, a correlation of reverse-target lesions with the autopsy
findings was possible. The diagnosis of necrotizing pneumonia was confirmed in
all lesions (Table 1). In three
lesions, a complete pathologic workup was performed. In the center of the
lesions, a necrosis was found consisting of cell detritus and some hyphae. In
two lesions, a pulmonary artery running through the lesions was occluded by
hyphae. The surrounding infiltration of the lung parenchyma consisted of
inflammatory cells, a few hyphae, and blood remnants including iron-laden
macrophages. The rim between central necrosis and surrounding infiltration
showed a high content of fibrin. There was no considerable accumulation of
hyphae or hemosiderin- or iron-laden cells in this rim.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Although there are some well-known advantages of MR imaging (e.g., differentiation of tumor and atelectasis), it is generally accepted that CT is the gold standard for the assessment of pulmonary parenchyma. In fact, only a few studies compare MR imaging with CT—in particular, helical CT—in an acceptable large patient group [9,10,11]. None of these studies showed any striking advantage of using MR imaging over CT. Therefore, CT remains the imaging method of choice to assess the lung parenchyma.
The results of our study indicate that MR imaging is able to show a variety of features of opportunistic pneumonia already well-known from CT including cavitation, air crescents, cysts, and reticular infiltrations. The same number of lesions was detected by MR imaging and CT in all patients of our study. In a number of cases, typical findings were revealed—for example, bronchogenic spread in pulmonary tuberculosis (Fig. 1A,1B).
Most of the CT and MR examinations (75%) were rated as showing identical results concerning not only the number but also the morphology of different lesions that were due to opportunistic pneumonia. We found it particularly interesting that MR imaging—just like CT—was able to depict different stages of infiltrations of the pulmonary parenchyma, which are known as groundglass infiltration and consolidation. These two terms represent precisely defined stages of lung parenchyma infiltration based on the visibility of the bronchovascular bundle, although ground-glass opacities detected on helical CT with an 8-mm collimation probably represent less subtle parenchymal findings than ground-glass opacities shown on high-resolution CT.
Ground-glass infiltrations are known to be a less severe form of parenchymal infiltration and a precursor of consolidation. Recent studies comparing high-resolution CT and chest radiography in immunocompromised patients have shown that ground-glass infiltrations may be an early sign of pneumonia and are easily missed on chest radiographs [12]. Furthermore, ground-glass opacities in a specific distribution allow confident diagnosis of P. carinii pneumonia in HIV-positive patients [26].
Our study shows that MR imaging is able to depict and to show ground-glass infiltrations and consolidations as well as helical CT in patients with opportunistic pneumonia (Fig. 1A,1B). However, there are some limitations that are due to the design of our study. Our major aim was to evaluate the potential of MR imaging in depicting acute inflammatory lesions. We did not establish the sensitivity and specificity of MR imaging compared with CT in detecting early pneumonia. Therefore, the results of our study need to be confirmed by studies with a larger number of patients. In particular, the false-negative rate of MR imaging has to be established because there is a bias toward more extensive and severe lung lesions caused by pneumonia in this study. This bias results from the small number of patients with ground-glass infiltrations and the lack of comparison with high-resolution CT, which is known to be more sensitive to discrete lesions caused by early viral or P. carinii pneumonia. In particular, the sensitivity of MR imaging regarding ground-glass infiltrations detected on high-resolution CT must be established. Concerning the superiority of CT—as well as MR imaging—to conventional chest radiography, our study confirms the results of recently published studies [12, 13] because a considerable number of CT and MR examinations showed pulmonary infiltrations in patients with normal findings on chest radiography (five of 16 patients in our study).
Furthermore, ground-glass infiltrations are known to represent a number of pathohistologic findings located in the alveolar space, interstitial space, or a combination thereof. Further studies need to determine whether MR imaging is able to show subtle ground-glass infiltrations in diseases like fibrosis and sarcoidosis, which are mainly caused by an infiltration of the pulmonary interstitium by inflammatory cells. It is obvious that in most cases of pneumonia the alveolar disease with a high content of protons predominates, which may be the major reason for the good results in our study. Regarding other definitely noninflammatory findings like bullae, CT was superior to MR imaging (one case in our study).
An interesting finding of this comparative study is the possible impact on diagnostic strategies in immunocompromised patients. Until now, CT—in particular, contrast-enhanced CT—has been the method of choice to establish a definite and early diagnosis of necrotizing pneumonia and abscess formation. In our study, T2-weighted MR imaging was able to depict necrotizing pneumonia earlier than CT in 25% of all examinations. Regarding the patients with suspected necrotizing lesions, MR imaging detected abscess formation earlier in 57% (four of seven patients suspected to have necrotizing lesions). Although there is no definite proof of abscess formation in these patients for the time of the MR examination (because of severe bleeding risks preventing an invasive procedure), we still believe that the MR diagnosis was sufficiently confirmed by findings of follow-up CT examinations or histologic proof from a subsequently performed biopsy. However, it is possible that the MR imaging depicted only a precursor of abscess formation and not a fully developed abscess.
As a characteristic feature of necrotizing pneumonia we observed the reverse-target sign, which consists of a hypointense rim localized at the border between the central necrosis and surrounding infiltration (both showing high signal intensity). This sign has already been described in the lung, brain, and liver by other authors [15,16,17,18]. To the best of our knowledge, there is no report in the literature until now connecting this sign with any other infectious disease except necrotizing pneumonia or abscess formation. Our study indicates that regarding this issue unenhanced T2-weighted MR imaging is superior to contrast-enhanced CT (Figs. 2A,2B,2C and 3A,3B,3C,3D). This is probably because of the excellent soft-tissue contrast of MR imaging.
Prior studies have proposed that the characteristic features of the reverse-target sign result from magnetic susceptibility effects of free radicals in the abscess capsule or iron-laden macrophages [17, 27]. In our study, we had the ability to correlate the imaging findings with pathohistologic findings in two cases. The striking feature in these two cases was a rim of fibrin located at the border between the central necrosis and the surrounding inflammatory cell infiltration and hemorrhage. From the MR appearance of other lesions with a considerable content of fibrin (e.g., fibrinous pleuritis), it is known that this substance is characterized by a low signal intensity on T2-weighted imaging sequences. Any other explanation for the dark rim of the reverse-target sign could not be found—in particular, no rim of fungal hyphae or iron-laden cells, both of which could also appear as areas of low signal intensity on T2-weighted images. We believe that a rim of fibrin is a plausible explanation for the reverse-target sign because fibrin is a ubiquitously found substance, whereas neither iron-laden macrophages nor fungal hyphae occur in all the diseases connected with the reverse-target sign. Furthermore, it is unlikely that a turbo spin-echo sequence would show susceptibility artifacts that were due to production of free radicals by macrophages because this MR technique is otherwise known to be insensitive to this kind of artifact.
In a healthy individual with necrotizing lesions in the lung, a wide range of possible causes has to be considered including septic emboli, infection with pyogenic bacteria, tuberculosis, noninfectious causes like Wegener's granulomatosis, and even metastases. In the immunocompromised host—in particular, the patient with neutropenia—a necrotizing pneumonia often represents a fungal infection that is caused by Aspergillus species or other fungi species. The early and rapid development of necrosis is a specific finding of invasive aspergillosis in neutropenic patients, which is caused by the angio-invasive growth of Aspergillus species leading to pulmonary infarctions [28]. In our study, five of seven patients with a necrotizing pneumonia suffered from a fungal pneumonia mostly caused by Aspergillus species. We believe that the reverse-target sign in the lung is a characteristic sign for a necrotizing pneumonia and may be a strong indicator of a fungal pneumonia in neutropenic patients. Recent studies in immunocompromised patients with fever of unknown origin advise early high-resolution CT to exclude pulmonary infection in these patients [12, 13]. The diagnosis of a necrotizing pneumonia with high-resolution CT is difficult and confined to advanced stages of the disease during which cavitation or air crescents are shown. This study indicates that T2-weighted MR imaging could be an alternative or at least a valuable imaging method in addition to CT in patients with pneumonia. However, this statement applies only to the particular subset of immunocompromised neutropenic patients at whom our study was directed.
MR imaging of the lung has been the subject of a number of studies during the past years. A variety of imaging sequences and different methods of motion compensation have been used. The latest studies indicate that T2-weighted turbo spin-echo sequences are able to compensate for the drawbacks of conventional T2-weighted spin-echo sequences, which include a marked loss in signal intensity and extensive motion artifacts [3, 9, 11]. Our approach to the motion compensation for lung MR imaging is different from prior studies: we performed a diastolic-triggered and expiratory-gated sequence. The image acquisition in the diastolic heart phase substantially reduces pulsation artifacts. Furthermore, normal lung vessels up to the subsegmental level are depicted with high signal intensity that results from the marked reduction of the flow-void effect [8, 9]. This allows exact localization of lung lesions because the pulmonary vessels are the most important "landmark" in the lung parenchyma. The major factor contributing to the long scan time compared with breath-hold T2-weighted techniques described before is the expiratory gating, which may cause a significant prolongation of the scan depending on the respiratory frequency of the patient. However, an MR examination with repeated breath-holds might fail to show pulmonary lesions because of variable depths of different breath-holds. Another problem is the deterioration of image quality because of motion artifacts in patients who cannot hold their breath as long and as often as necessary—obviously, a frequent condition in patients with lung disease. For T2-weighted MR imaging, expiratory gating seems to be an acceptable resolution for this problem. This is confirmed by the fact that most MR examinations in this study were rated as excellent or good regarding image quality and artifact reduction.
In conclusion, this comparison of helical CT and T2-weighted turbo spin-echo imaging in patients with opportunistic pneumonia confirms the good results of this MR technique indicated by prior investigations [8, 9, 11]. In patients with opportunistic pneumonia shown on standard helical CT, MR imaging is able to depict lung lesions with almost identical results as compared with CT. Typical features of pneumonia including different stages of lung parenchyma infiltration could be shown on MR imaging. Furthermore, the early detection of a necrotizing lesion using the reverse-target sign offers advantages in the immunocompromised host because this sign is a strong indicator of a fungal pneumonia. However, further studies comparing MR imaging with high-resolution CT are needed to establish sensitivity and specificity of MR imaging for early detection of opportunistic pneumonia.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Eibel, P. Herzog, O. Dietrich, C. T. Rieger, H. Ostermann, M. F. Reiser, and S. O. Schoenberg Pulmonary Abnormalities in Immunocompromised Patients: Comparative Detection with Parallel Acquisition MR Imaging and Thin-Section Helical CT Radiology, December 1, 2006; 241(3): 880 - 891. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
