DOI: 10.1148/radiol.2443060846
(Radiology 2007;244:875-882.)
© RSNA, 2007
Is It Possible to Recognize Pulmonary Infarction on Multisection CT Images?1
Marie-Pierre Revel, MD,
Rached Triki, MD,
Gilles Chatellier, MD,
Sophie Couchon, MD,
Nathalie Haddad, MD,
Anne Hernigou, MD,
Claire Danel, MD, and
Guy Frija, MD
1 From the Assistance Publique des Hôpitaux de Paris, Paris, France (M.P.R., R.T., G.C., S.C., N.H., A.H., C.D., G.F.); Department of Radiology (M.P.R., R.T., S.C., N.H., A.H., G.F.), Clinical Research Unit (G.C.), and Laboratory of Anatomy and Pathology (C.D.), Hôpital Européen Georges Pompidou, 20 rue Leblanc, F75015 Paris, France; and Université Paris Descartes, Paris, France (M.P.R., R.T., G.C., S.C., N.H., A.H., C.D., G.F.). Received May 16, 2006; revision requested July 12; revision received September 8; accepted October 12; final version accepted March 1, 2007.
Address correspondence to M.P.R. (e-mail: marie-pierre.revel{at}egp.aphp.fr).
 |
ABSTRACT
|
|---|
Purpose: To retrospectively determine sensitivity and specificity of four findings for distinguishing pulmonary infarction from other causes of peripheral pulmonary consolidations on multidetector computed tomographic (CT) images, with other CT and clinical findings as reference.
Materials and Methods: Institutional review board approved the study and waived informed consent. Three independent radiologists blindly analyzed selected multisection CT images of 50 pulmonary infarctions—not showing direct arterial signs of pulmonary embolism—and 100 peripheral consolidations of other origins. Readers analyzed four findings: triangular shape, vessel sign (defined as presence of an enlarged vessel at the apex of consolidation), central lucencies, and air bronchograms. Interobserver agreement; frequency on CT images with and without infarct; and sensitivity, specificity, and positive likelihood ratio (LR) for diagnosis of pulmonary infarction were assessed for each finding.
Results: One hundred fifty peripheral consolidations were analyzed in 134 (75 men, 59 women) patients (mean age, 55.9 years ± 17.4 [standard deviation] vs 54.7 ± 19.9; P = .71). Interobserver agreement was good for central lucencies and air bronchograms and poor to moderate for the other two findings (
< 0.61). Compared with CT images without infarct, CT images with infarct had a higher frequency of vessel sign (32% [16 of 50] vs 11% [11 of 100], P = .029) and central lucencies (46% [23 of 50] vs 2% [two of 100], P < .001) and a lower frequency of air bronchograms (8% [four of 50] vs 40% [40 of 100], P = .003). Frequency of triangular shape was similar in both groups (52% [26 of 50] vs 40% [40 of 100], P = .17). Positive LR was 23.0 for central lucencies, 2.9 for vessel sign, 1.3 for triangular shape, and 0.2 for air bronchograms. Presence of central lucencies had 98% specificity and 46% sensitivity for pulmonary infarction. When the vessel sign and negative air bronchogram were combined with central lucencies, specificity increased to 99% but sensitivity decreased to 14%.
Conclusion: Central lucencies in peripheral consolidations are highly suggestive of pulmonary infarction.
© RSNA, 2007
|
INTRODUCTION
|
|---|
Numerous studies have shown high diagnostic performance of helical computed tomographic (CT) angiography for pulmonary embolism (PE), with pooled reported sensitivity and specificity values of 86.0% and 93.7%, respectively (1). Recent multisection CT technology has resulted in a gain in sensitivity (2). Some authors have described CT parenchymal abnormalities associated with PE (3–8). In a study in 88 patients with suspected PE, Coche et al found that wedge-shaped consolidations were more likely to be observed in patients with PE than in patients without PE (62% vs 27%, P < .05) (3). However, the sensitivity and specificity of this finding were only 62% and 73%, respectively.
Although the diagnosis of acute thromboembolism is based on direct arterial findings, as defined by Remy-Jardin et al (9), it may be important to recognize pulmonary infarction in situations in which these findings are missing. Such situations may occur when vascular opacification is inadequate and the CT examination results are therefore inconclusive—situations that were reported to have occurred in 3% of CT examinations (10)—or when the CT examination is performed without contrast medium administration because clinical signs of PE are lacking. The presence of parenchymal findings suggestive of pulmonary infarction indicates the need for performance of additional CT angiography. In addition, with distal PE or when CT angiography is delayed relative to clinical onset, direct arterial findings may be missing, even on high-quality enhanced CT images. In such situations, images suggestive of pulmonary infarction may indicate the need for additional investigations to diagnose or rule out thromboembolism.
Thus, the purpose of our study was to retrospectively determine the sensitivity and specificity of four findings for distinguishing pulmonary infarction from other causes of peripheral pulmonary consolidations on multidetector CT images, with other CT and clinical findings as the reference standard.
|
MATERIALS AND METHODS
|
|---|
The study was approved by the institutional review board of Hôpital Européen Georges Pompidou and Université Paris Descartes, Paris, France. Informed consent was waived.
Selection of CT Images
Images from CT examinations with infarct and without infarct were sought by using a keyword query in our picture archiving and communication system for the period from July 2000 to July 2005 (Fig 1). The selected keyword was "pulmonary infarction" for images from CT examinations with infarct. For images from CT examinations without infarct, we used keywords corresponding to diseases that usually cause chronic or acute pulmonary consolidations, such as "cryptogenic organizing pneumonia," "bronchiolo-alveolar cell carcinoma," "pulmonary lymphoma," "Wegener's granulomatosis," "sarcoidosis," "tuberculosis," and "pneumonia." Other selected terms were "lipoid pneumonia," "radiation pneumonitis," "hemoptysis and consolidation," and "lung contusion." The number of images from CT examinations without infarcts (n = 100) was chosen to be twice the number of images from examinations with infarcts (n = 50) to improve the specificity estimates.
Only images from CT examinations with peripheral consolidations abutting the pleura were included. Images from CT examinations were selected consecutively and retrospectively from July 2005 back to July 2000 until the target numbers of 50 images from CT examinations for the group with infarct and 100 images from CT examinations for the group without infarct were reached. Images from CT examinations were included only if the diagnosis was either histologically proved or based on the following criteria: (a) For the group with infarct, the criterion was peripheral consolidation with adjacent subsegmental PE or more central PE with pleuritic pain or hemoptysis. (b) For the group without infarct, the criterion was the presence of an alternative cause of pulmonary consolidation and no mention of thromboembolism in the hospital charts. For each image, patient identification was removed. Two window settings, one each at lung parenchymal and mediastinal window settings, were selected by two radiologists (M.P.R., R.T.) who did not participate in the blinded readings.
The two selected images were cropped to conceal the CT acquisition parameters, such as the reconstruction section thickness, and were pasted in a computer graphic presentation document (PowerPoint; Microsoft, Redmond, Wash). For the group with infarct, special attention was paid not to include in the selection any image of an arterial filling defect or of arterial occlusion, when present.
Reference Standard
Group with infarct.—The diagnosis of pulmonary infarction was histologically proved in one patient. In another 43 infarcts, the diagnosis was based on the presence of a wedge-shaped consolidation in a territory of subsegmental PE, with or without clinical signs suggesting pulmonary infarction. In six infarcts, the patients had arterial filling defects in arteries supplying segments other than those showing the wedge-shaped consolidations, associated with signs of pulmonary infarction, such as pleuritic chest pain and/or hemoptysis.
Group without infarct.—The diagnosis was histologically proved in 33 patients with tumors, cryptogenic organizing pneumonia, sarcoidosis, or Wegener granulomatosis. The diagnosis of pneumonia relied on bacteriologic data and/or short-term resolution following antibiotic therapy. Tuberculosis had been bacteriologically proved. The results of bronchoalveolar lavage were used for validating the diagnosis of lipoid pneumonia and of radiation pneumonitis, with lipid-laden macrophages in the former and lymphocytic alveolitis in the latter. For the diagnosis of radiation pneumonitis, we also relied on a history of radiation therapy that included the lungs in the 3 months prior to CT examination and on improvement after steroid therapy.
The patient's history and the hospital charts were used for the remaining diagnoses as follows: (a) Peripheral consolidation due to contusion was described in trauma patients. (b) Posthemoptysis consolidations were seen in patients with large or massive hemoptysis. The site of the opacity corresponded to the bleeding site found at bronchoscopy. The pulmonary opacities had resolved, as evidenced on follow-up CT images obtained 4–6 weeks later to rule out a tumoral origin of the bleeding. (c) Multiplanar reformation was used to validate atelectasis as the cause of pulmonary opacities at initial interpretation (Table 1).
CT Examinations
All CT examinations had been performed with a four-detector CT scanner (LightSpeed; GE Healthcare, Milwaukee, Wis) with a 1.25-mm section thickness. Reconstructed 1.25-mm-thick images were available in mediastinal and lung window settings for all CT examinations for the group of images with infarct and for 31 CT examinations for the group of images without infarct. For the remaining 69 CT examinations in the group of images without infarct, the available reconstruction section thickness was 2.50 mm for lung window settings and 2.50 or 5.00 mm for mediastinal window settings. Contrast medium had been used for 46 of the 50 CT examinations for the group of images with infarct. In the four remaining CT examinations, CT images acquired without contrast medium administration were selected for the readings. These remaining CT examinations concerned three patients in whom wedge-shaped opacities were first seen on images obtained before administration of contrast medium and one patient with a case of chronic peripheral consolidation that was diagnosed as pulmonary infarction after percutaneous biopsy.
Forty-seven images of the group without infarct were enhanced and 53 images were unenhanced.
Reading Sessions
Three independent radiologists with 20 years of experience (A.H. [reader 1]), 6 years of experience (S.C. [reader 2]), and 1 year of experience (N.H. [reader 3]) in thoracic imaging were asked to independently analyze the selected CT images and were blinded to the underlying diagnosis (infarct vs no infarct). The 150 images from CT examinations were randomized in a different order for each of the three readers. Three reading sessions were then organized, each including 50 images from CT examinations.
The readers were asked to analyze the consolidations for the presence of four findings: triangular shape, vessel sign, central lucencies, and air bronchograms. They were aware of the purpose of the study, but they were not asked to distinguish between images with and without infarct from CT examinations. The readers had no recommendations concerning any association or presence of an individual finding that suggested infarction.
An initial training session was conducted by using images from five CT examinations that were not included in this study. The two radiologists (M.P.R., R.T.) who selected the images from the CT examinations explained how the data sheets were to be completed and defined each finding.
Triangular shape corresponded to a consolidation with a broad base abutting a pleural surface and an apex in a more central portion of the lung (Figs 2–6). The vessel sign was defined as the presence of an enlarged vessel that led to the apex of a wedge-shaped opacity (Fig 4). Central lucencies were defined as round foci of hypoattenuation in the central portion of wedge-shaped opacities. They had to be larger than the bronchial lumen, and they had to show no linear shape or bifurcation, especially on images obtained with mediastinal window settings (Figs 2–5). Air bronchograms were defined as linear bifurcated areas of hypoattenuation that corresponded to aerated small bronchi inside consolidations (Fig 6).
View larger version (107K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a: Transverse contrast-enhanced CT scans show pulmonary infarction in lower lobe of left lung in a 37-year-old man. (a) Mediastinal window setting. (b) Lung window setting. Wedge-shaped peripheral consolidation (arrow) shows central foci of hypoattenuation with no linear shape or bifurcation on a. These central lucencies are better seen on a than on b. Air bronchogram is absent. This patient had associated iliac vein thrombosis at CT venography, chest pain, and hemoptysis.
|
|
View larger version (189K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b: Transverse contrast-enhanced CT scans show pulmonary infarction in lower lobe of left lung in a 37-year-old man. (a) Mediastinal window setting. (b) Lung window setting. Wedge-shaped peripheral consolidation (arrow) shows central foci of hypoattenuation with no linear shape or bifurcation on a. These central lucencies are better seen on a than on b. Air bronchogram is absent. This patient had associated iliac vein thrombosis at CT venography, chest pain, and hemoptysis.
|
|
View larger version (107K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a: Transverse contrast-enhanced CT scans show consolidation with central lucencies in a 68-year-old woman with pulmonary infarction in the upper lobe of the left lung. (a) Mediastinal window setting. (b) Lung window setting. The central portion of hypoattenuation, better seen on a than on b, is surrounded by dense consolidation (arrow), which probably reflects a peripheral inflammatory reaction. Air bronchograms are absent. This patient had PE involving several segmental and subsegmental pulmonary arteries (not shown).
|
|
View larger version (161K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b: Transverse contrast-enhanced CT scans show consolidation with central lucencies in a 68-year-old woman with pulmonary infarction in the upper lobe of the left lung. (a) Mediastinal window setting. (b) Lung window setting. The central portion of hypoattenuation, better seen on a than on b, is surrounded by dense consolidation (arrow), which probably reflects a peripheral inflammatory reaction. Air bronchograms are absent. This patient had PE involving several segmental and subsegmental pulmonary arteries (not shown).
|
|
View larger version (71K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4: Transverse contrast-enhanced CT scan obtained with mediastinal window setting shows pulmonary infarction in upper lobe of the right lung in a 43-year-old man. Peripheral consolidation (arrowhead) is seen. Central lucencies and enlarged vessel (arrow) leading to the apex of the wedge-shaped consolidation are clearly visible. Air bronchogram is absent. Central embolism was present on adjacent CT sections.
|
|
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a: Pathologically proved pulmonary infarction in the lower lobe of the left lung in a 65-year-old man. (a) Transverse unenhanced CT scan obtained with mediastinal window setting shows central lucencies visible within the consolidation (arrow). (b) Photomicrograph of core biopsy material from the infarcted area showed that the central area (bottom of the figure) was necrotic. Only ghosts of alveolar walls can be seen, with acidophilic cells and nuclear loss, and alveolar lumen are empty. Biopsy was performed to find the cause of chronic unexplained consolidation in the lower lobe of the left lung. Malignancies such as pulmonary lymphoma or bronchiolo-alveolar cell carcinoma were suspected. (Hematoxylin-eosin stain; original magnification, x100.)
|
|
View larger version (69K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b: Pathologically proved pulmonary infarction in the lower lobe of the left lung in a 65-year-old man. (a) Transverse unenhanced CT scan obtained with mediastinal window setting shows central lucencies visible within the consolidation (arrow). (b) Photomicrograph of core biopsy material from the infarcted area showed that the central area (bottom of the figure) was necrotic. Only ghosts of alveolar walls can be seen, with acidophilic cells and nuclear loss, and alveolar lumen are empty. Biopsy was performed to find the cause of chronic unexplained consolidation in the lower lobe of the left lung. Malignancies such as pulmonary lymphoma or bronchiolo-alveolar cell carcinoma were suspected. (Hematoxylin-eosin stain; original magnification, x100.)
|
|
View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a: Transverse unenhanced CT images obtained with (a) mediastinal and (b) lung window settings show consolidation (arrow) in a 77-year-old man with pathologically proved cryptogenic organizing pneumonia. Air bronchograms are visible on both images.
|
|
View larger version (163K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b: Transverse unenhanced CT images obtained with (a) mediastinal and (b) lung window settings show consolidation (arrow) in a 77-year-old man with pathologically proved cryptogenic organizing pneumonia. Air bronchograms are visible on both images.
|
|
Statistical Analysis
Interobserver agreement was assessed by using the statistic with 95% confidence intervals (CIs). The values were interpreted on the following basis: less than 0.20, poor agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, good agreement; and 0.81–1.00, very good agreement. A finding was considered to be present if all three radiologists identified it and was considered to be absent in all other situations. Differences in the frequency of the findings between the groups with and without infarct were assessed by using the Fisher exact test. The sensitivity, specificity, and 95% CIs were assessed for each finding. The positive likelihood ratio (LR), which was calculated by dividing sensitivity by specificity subtracted from one, and its 95% CI were calculated to determine the increase in the probability of pulmonary infarction when that particular finding was present.
Positive LRs were classified according to Jaeschke et al (11): Findings with LRs greater than 10 or less than 0.1 were considered to have a very good diagnostic value. Findings with LRs of 5–10 and 0.1–0.2 were considered to have a good diagnostic value. Findings with LRs of 2–5 and 0.5–0.2 were considered to have a moderate diagnostic value. Findings with LRs of 1–2 and 0.5–1 were considered to have a poor diagnostic value.
Mean ages between men and women were compared by using the Student t test. Differences with a P value less than .05 were considered significant. Software (SPSS for Windows, version 12.0; SPSS, Chicago, Ill) was used for all statistical tests.
|
RESULTS
|
|---|
Clinical Data
The 150 peripheral pulmonary consolidations were observed in 134 patients who were 18–92 years old: There were 75 men and 59 women of similar age (mean age, 55.9 years ± 17.4 [standard deviation] vs 54.7 ± 19.9, P = .71).
CT Features
The size range of the peripheral consolidations was 2.6–11.2 cm (mean, 5.2 cm). The size range was 2.0–7.8 cm for infarctions (mean, 4.5 cm) and 2.3–11.2 cm for other causes of consolidations (mean, 5.5 cm). The difference was significant (P = .01, Mann-Whitney test).
Seventy-four of the 150 peripheral consolidations were located in the lower lobes; of the 74, 24 were of 50 infarctions (Figs 2 and 5) and 50 were of the other 100 consolidations.
Twenty-six of the 50 infarctions were not located in the lower lobes; 11 were in the middle lobe of the right lung, five were in the upper lobe of the right lung, and the remaining 10 were in the upper lobe of the left lung. Fifty of the lesions without infarct were not located in the lower lobes; eight were in the middle lobe of the right lung, 19 were in the upper lobe of the right lung, and the remaining 23 were in the upper lobe of the left lung.
Reading Results
Overall, agreement was good for central lucencies and air bronchograms, with values ranging from 0.64 to 0.74, although agreement between reader 2 and reader 3 was only moderate for central lucencies ( = 0.46). Agreement was poor to moderate for the vessel sign and triangular shape (Table 2). Compared with CT examinations without infarct, CT examinations with infarct had a higher frequency of the vessel sign (32% [16 of 50] vs 11% [11 of 100], P = .029) and central lucencies (46% [23 of 50] vs 2% [two of 100], P < .001) and a lower frequency of air bronchograms (8% [four of 50] vs 40% [40 of 100], P = .003). The frequency of triangular shape was similar in the two groups (52% [26 of 50] vs 40% [40 of 100], P = .17) (Table 3).
The positive LRs were 23.0 for central lucencies, 2.9 for the vessel sign, 0.2 for air bronchograms, and 1.3 for triangular shape (Table 3).
Sensitivity and specificity of each finding (Table 3) showed that central lucencies had the highest specificity (98%) for the diagnosis of pulmonary infarction.
Central lucencies were present and air bronchograms were absent in 24 patients. This combination had better specificity (98% [95% CI: 93.0%, 99.4%]) than sensitivity (44% [95% CI: 31.2%, 57.7%]) and a positive LR of 22.0 (95% CI: 5.39, 89.9).
Eight patients had central lucencies and the vessel sign but no air bronchograms. This combination had high specificity (99% [95% CI: 94.6%, 99.9%]), very poor sensitivity (14% [95% CI: 7.0%, 26.2%]), and a positive LR of 14.0 (95% CI: 1.8, 110.7).
|
DISCUSSION
|
|---|
Parenchymal CT abnormalities are frequent in patients with PE. In the Prospective Investigation of Pulmonary Embolism Diagnosis study, findings at chest radiography were normal in only 12% of patients with PE (12). Atelectasis and parenchymal areas of increased opacity are the most common findings but are nonspecific. Atelectasis can be an accompanying finding in PE but may also be an alternative diagnosis in patients with suspected PE (13).
In the CT study by Shah et al (4), the frequencies of atelectasis and pleural effusion were not different between patients with and without PE. The only parenchymal abnormality significantly associated with PE was peripheral wedge-shaped opacity, which was seen in 25% of patients with PE and 5% of patients without PE, and was considered to represent pulmonary infarction.
However, wedge-shaped opacity is not specific for pulmonary infarction, as it can also be observed in acute and organized pneumonia, tumors, Wegener granulomatosis, and numerous other diseases with various underlying pathologic changes. In pulmonary infarction, consolidation is mainly caused by alveolar filling with blood, with a peripheral inflammatory reaction around central necrosis (14).
In this study, we examined whether peripheral opacities due to pulmonary infarction could be differentiated from other causes on the basis of findings that also included internal morphologic characteristics such as the presence of central lucencies and the absence of air bronchograms. The latter had been reported to be a reliable finding for pulmonary infarction on chest radiographs (15), but the finding has not been previously evaluated on CT images, to our knowledge. The other two findings we evaluated were the triangular shape of the consolidation and the vessel sign.
Readings were performed independently by readers with different years of experience, with no attempt to reach a consensus between readers. A finding was considered to be present only if all three radiologists identified it. The vessel sign had a positive LR of 2.9 for the diagnosis of pulmonary infarction. Ren et al (7) reported that a vessel sign associated with a subpleural density was more common in patients with infarcts than in patients with other causes of subpleural opacities. In a CT–pathologic correlation study, this finding corresponded to pulmonary branches containing thrombi or dilated patent channels proximal to an area of vascular obstruction (8).
The positive LR of air bronchograms was only 0.2, meaning that pulmonary infarction is unlikely if this finding is present. Conversely, the positive LR for central lucencies was high, meaning that pulmonary infarction is very likely if this finding is present. Air bronchograms and central lucencies, respectively, have good and very good diagnostic value according to the classification of Jaeschke et al (11) of positive LR values, whereas the diagnostic value of the vessel sign is only moderate.
Low-attenuation areas within peripheral consolidations were observed in seven (58%) of 12 patients with pulmonary infarcts in the previously mentioned CT–pathologic correlation study (8). Microscopically, they corresponded to viable lung tissue amid infarcted secondary pulmonary lobules and were better seen on a soft-tissue window setting. However, these areas were not described as being central. Internal air lucencies were observed in 32% of pulmonary infarcts on images from multidetector CT examinations in a recent review of 37 cases (16). The authors suggested two possible explanations for these images: survival of some of the pulmonary lobules within the region of the infarct (with reference to the study of Balakrishnan et al [8]) or cavitation of pulmonary infarcts. In the case studied pathologically in our series, in which centrally located lucencies within the consolidation were observed on the CT image, the central portion of the lesion showed necrosis surrounded by peripheral inflammation. This is the reason that we believe central lucencies tend to represent necrosis, whereas more eccentric lucencies may represent viable lung tissue.
A central location and the absence of an associated air bronchogram are noteworthy features for distinguishing central lucencies of pulmonary infarction from bubblelike lucencies seen in some subtypes of peripheral adenocarcinoma (17). We could not determine whether this distinction is possible because the keywords used for the selection of images from CT examinations focused on causes of consolidation. Thus, our series only included a few tumors (mainly bronchiolo-alveolar carcinomas and only seven tumors of other types). Our results show that combining other findings with central lucencies, such as the vessel sign and the absence of air bronchograms, helps to slightly increase the specificity (from 98% to 99%) but leads to marked reduction in sensitivity (from 46% to 14%).
Pathologic validation was only available for one infarct and 33 consolidations of other causes. However, we used precise diagnostic criteria for the remaining 116 cases, especially for the group with infarct. The diagnosis of pulmonary infarction was based on the presence of wedge-shaped opacities in the territory of a thrombosed subsegmental artery in 43 of 49 infarcts without pathologic documentation. In the remaining six cases, the diagnosis of pulmonary infarction was based only on clinical manifestations of the pulmonary infarction syndrome defined as pleuritic pain or hemoptysis (18) combined with the presence of clots in territories other than those where the consolidation was observed. Such symptoms in patients with wedge-shaped consolidation and arterial thrombi on CT images are nevertheless highly suggestive of infarction, although subsegmental emboli are not detected adjacent to the consolidation.
Pathologic correlation was not available for the four findings evaluated here. However, in one case we were able to match central lucencies with areas of central necrosis surrounded by an inflammatory reaction. Other cases were not studied pathologically.
Another limitation is that the readers were asked to analyze the findings on only two images. This small number of images could explain the low values for some of the findings, especially for the vessel sign and triangular shape, that may be difficult to assess on a single section. However, this restriction was necessary to perform the reading sessions in a reasonable time and to avoid biased infarct recognition that was based on direct arterial findings. Despite the restriction of this analysis to two images per case, we observed that the finding of central lucencies, which is the main result of our study, was useful for diagnosis of infarction because of both good interobserver agreement and high positive LR.
Finally, we did not evaluate the interval between symptom onset and performance of CT in the group with infarct. For the lung, as well as for other organs, detectability of ischemic injury is not immediate after arterial occlusion even though there are no specific reports in the literature that allowed us to assess this delay precisely. However, our aim was not to determine the optimal timing of CT for detection of pulmonary infarction but to examine whether pulmonary infarction, when present, can be recognized by using multisection CT.
In conclusion, we found good interobserver agreement for two findings of pulmonary infarction, namely central lucencies and air bronchograms, despite differences in the level of experience. Interobserver agreement for the vessel sign was only fair. The presence of central lucencies within a peripheral consolidation strongly suggests pulmonary infarction. We believe the lucencies correspond to central necrosis, as shown by results of pathologic studies in one of our patients. There is no benefit to be drawn from combining this finding with other findings, as combination leads to marked reduction in sensitivity without enhancement of specificity or positive LR. Findings of pulmonary infarction may be especially useful when direct findings of thromboembolism are missing, especially when CT examination is delayed relative to symptom onset.
|
ADVANCES IN KNOWLEDGE
|
|---|
- Pulmonary infarction is very likely if a peripheral consolidation contains central lucencies (positive likelihood ratio [LR] of 23).
- Pulmonary infarction is unlikely if a peripheral consolidation contains air bronchograms (positive LR of 0.2).
- The presence of central lucencies within a peripheral consolidation has 98% specificity and 46% sensitivity for pulmonary infarction.
|
IMPLICATION FOR PATIENT CARE
|
|---|
- Contrast-enhanced CT should be performed if not initially performed in the setting of findings consistent with pulmonary infarction at unenhanced CT.
|
ACKNOWLEDGMENTS
|
|---|
We thank David Young, BS, and Chadi Khalil, MD, Centre Hospitalier Regional Universitaire de Lille, Hôpital Calmette, Lille, France, for editorial assistance.
|
FOOTNOTES
|
|---|
Abbreviations: CI = confidence interval • LR = likelihood ratio • PE = pulmonary embolism
Authors stated no financial relationship to disclose.
Author contributions:Guarantors of integrity of entire study, M.P.R., R.T.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, M.P.R., R.T.; clinical studies, S.C., N.H., A.H.; statistical analysis, G.C.; and manuscript editing, M.P.R., R.T., C.D., G.F.
|
References
|
|---|
- Hayashino Y, Goto M, Noguchi Y, Fukui T. Ventilation-perfusion scanning and helical CT in suspected pulmonary embolism: meta-analysis of diagnostic performance. Radiology 2005;234:740–748.[Abstract/Free Full Text]
- Perrier A, Roy PM, Sanchez O, et al. Multidetector-row computed tomography in suspected pulmonary embolism. N Engl J Med 2005;352:1760–1768.[Abstract/Free Full Text]
- Coche EE, Muller NL, Kim KI, Wiggs BR, Mayo JR. Acute pulmonary embolism: ancillary findings at spiral CT. Radiology 1998;207:753–758.[Abstract/Free Full Text]
- Shah AA, Davis SD, Gamsu G, Intriere L. Parenchymal and pleural findings in patients with and patients without acute pulmonary embolism detected at spiral CT. Radiology 1999;211:147–153.[Abstract/Free Full Text]
- Johnson PT, Wechsler RJ, Salazar AM, Fisher AM, Nazarian LN, Steiner RM. Spiral CT of acute pulmonary thromboembolism: evaluation of pleuroparenchymal abnormalities. J Comput Assist Tomogr 1999;23:369–373.[CrossRef][Medline]
- Reissig A, Heyne JP, Kroegel C. Ancillary lung parenchymal findings at spiral CT scanning in pulmonary embolism: relationship to chest sonography. Eur J Radiol 2004;49:250–257.[CrossRef][Medline]
- Ren H, Kuhlman JE, Hruban RH, Fishman EK, Wheeler PS, Hutchins GM. CT of inflation-fixed lungs: wedge-shaped density and vascular sign in the diagnosis of infarction. J Comput Assist Tomogr 1990;14:82–86.[Medline]
- Balakrishnan J, Meziane MA, Siegelman SS, Fishman EK. Pulmonary infarction: CT appearance with pathologic correlation. J Comput Assist Tomogr 1989;13:941–945.[Medline]
- Remy-Jardin M, Remy J, Deschildre F, et al. Diagnosis of pulmonary embolism with spiral CT: comparison with pulmonary angiography and scintigraphy. Radiology 1996;200:699–706.[Abstract/Free Full Text]
- Revel MP, Petrover D, Hernigou A, Lefort C, Meyer G, Frija G. Diagnosing pulmonary embolism with four-detector row helical CT: prospective evaluation of 216 outpatients and inpatients. Radiology 2005;234:265–273.[Abstract/Free Full Text]
- Jaeschke R, Guyatt GH, Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. B. What are the results and will they help me in caring for my patients? The Evidence-Based Medicine Working Group. JAMA 1994;271:703–707.
- Worsley DF, Alavi A, Aronchick JM, Chen JT, Greenspan RH, Ravin CE. Chest radiographic findings in patients with acute pulmonary embolism: observations from the PIOPED Study. Radiology 1993;189:133–136.[Abstract/Free Full Text]
- Tsai KL, Gupta E, Haramati LB. Pulmonary atelectasis: a frequent alternative diagnosis in patients undergoing CT-PA for suspected pulmonary embolism. Emerg Radiol 2004;10:282–286.[CrossRef][Medline]
- Frazier AA, Galvin JR, Franks TJ, Rosado-de-Christenson ML. Pulmonary vasculature: hypertension and infarction. RadioGraphics 2000;20:491–524.[Abstract/Free Full Text]
- Bachynski JE. Absence of the air bronchogram sign: a reliable finding in pulmonary embolism with infarction or hemorrhage. Radiology 1971;100:547–552.[Medline]
- He H, Stein MW, Zalta B, Haramati LB. Pulmonary infarction: spectrum of findings on multidetector helical CT. J Thorac Imaging 2006;21:1–7.[CrossRef][Medline]
- Nakazono T, Sakao Y, Yamaguchi K, Imai S, Kumazoe H, Kudo S. Subtypes of peripheral adenocarcinoma of the lung: differentiation by thin-section CT. Eur Radiol 2005;15:1563–1568.[CrossRef][Medline]
- Stein PD, Henry JW. Clinical characteristics of patients with acute pulmonary embolism stratified according to their presenting syndromes. Chest 1997;112:974–979.[Medline]
This article has been cited by other articles:
|
|
|
|
 
S. N. J. Pipavath and J. D. Godwin
Acute Pulmonary Thromboembolism: A Historical Perspective
Am. J. Roentgenol.,
September 1, 2008;
191(3):
639 - 641.
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
