| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Department of Radiology, University of Vienna and Ludwig Boltzmann Institute for Clinical and Experimental Radiology, Vienna, Austria (C.B.H., S.G., K.F.L., M.M.T., C.C., C.J.H.); and Department of Radiology, Otto Wagner Hospital, Vienna, Austria (G.H.M.). From the 2000 RSNA scientific assembly. Received January 25, 2002; revision requested March 30; revision received June 17; accepted July 25. Address correspondence to C.B.H., Department of Radiology, University of California San Francisco, 505 Parnassus Ave L308, San Francisco, CA 94143-0628 (e-mail: christine.henk@univie.ac.at).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Two hundred forty-four spiral CT angiographic images of the pulmonary arteries obtained during deep inspiration in patients suspected of having pulmonary embolism (PE) were reviewed for evidence of abnormal early enhancement of the aorta. In 45 patients, enhancement of the ascending aorta was equal to or more than that of the pulmonary arteries. Nonenhanced or contrast material–enhanced echocardiography was performed in 39 of these cases. All CT images with abnormal enhancement patterns were graded for contrast quality with respect to sufficient enhancement of pulmonary arteries (four grades) at three anatomic levels: right and left main and lobar and segmental branches. In addition, all spiral CT angiographic images were evaluated concerning the diagnosis of PE and the grouping of central (main pulmonary artery to proximal lobar arteries) and peripheral (beyond proximal lobar branches) locations of emboli. Mean attenuation values of ascending aortas and main pulmonary arteries in group 1 (n = 244) were compared with those in groups 2 and 3 (n = 45) by means of the two-tailed Student t test for unpaired data (P < .05).
RESULTS: Attenuation values for ascending aortas in group 1 were significantly lower than those in groups 2 and 3 (P < .001). Attenuation values in main pulmonary arteries were significantly higher in group 1 than in groups 2 and 3 (P < .001). Echocardiographic images showed an intracardiac right-to-left shunt in all 39 cases with abnormal contrast dynamics in the CT study (16% of the whole study population). Three patients had an atrial-septal defect, and 36 had a PFO. Images with a shunt had good (9%), intermediate (37%), fair (33%), and poor (23%) contrast of the pulmonary arteries. Sufficient vessel contrast for the diagnosis of PE could not be achieved in 27 of 45 patients with a shunt, but severe central PE could be ruled out. PE could be diagnosed in 31% of the 244 images, 58% were negative, and 11% were indeterminate.
CONCLUSION: A PFO may frequently lead to insufficient attenuation of the pulmonary arteries, which potentially limits the diagnosis of PE if the examination is performed during deep inspiration.
© RSNA, 2003
Index terms: Atrial septal defect, 514.141 • Computed tomography (CT), helical, 60.12115, 60.12116 • Embolism, pulmonary, 60.72 • Foramen ovale
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Currently, spiral computed tomographic (CT) angiography of the pulmonary arteries has evolved as an established procedure in the work-up of suspected pulmonary thromboembolic disease (15–21). Contrast enhancement and image quality depend to a large extent on the amount of contrast material injected, the flow rate, and, most important, the scan delay. Appropriate bolus timing is mandatory for sufficient attenuation of the pulmonary arterial tree to the segmental and subsegmental level (15,16). To avoid motion artifacts as a result of breathing, the examination is usually performed during deep inspiration, which causes hemodynamic effects that are similar to those that occur with a Valsalva maneuver.
During our investigation of a large cohort of patients suspected of having pulmonary embolism (PE), we saw cases with a peculiar contrast enhancement pattern: early and strong attenuation of the thoracic aorta in combination with unsatisfactory pulmonary artery contrast. We hypothesized that such contrast dynamics despite use of a correct scan protocol could potentially be caused by an intracardiac right-to-left shunt that was supported by deep inspiration.
The purpose of this study was to investigate if abnormal early contrast enhancement of the aorta and decreased attenuation of pulmonary arteries might be caused by a PFO when spiral CT angiography is performed during deep inspiration.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Retrospective Component
Two hundred forty-four consecutive spiral CT angiographic images of the pulmonary arteries in 244 patients suspected of having PE were reviewed for evidence of abnormal contrast enhancement of the thoracic vessels. The study group comprised 127 (52%) female and 117 (48%) male subjects (age range, 17–89 years; mean age, 53.5 years ± 17.4 [SD]) who were seen during the 16 months from October 1997 to January 1999. The number of inpatients and outpatients was almost equal (127 [52%] inpatients, 117 [48%] outpatients). In 230 of the 244 patients, the clinical diagnosis was suspected acute PE; 14 patients were referred for CT evaluation of chronic thromboembolic pulmonary hypertension.
All examinations were performed with a single–detector row spiral CT scanner (Tomoscan SR 8000; Philips, Best, the Netherlands) with the patient supine. Examinations were performed with the following scan parameters: section thickness of 3–5 mm, table increment of 5 mm/sec, and reconstruction index of 2–3 mm. Coverage along the z axis was defined on the anteroposterior localization topogram and spanned from 1 cm above the aortic arch to the lower lateral costophrenic sinus. The field of view covered the entire thorax in the x and y axes, and a 5122 matrix was used.
Between 80 and 150 mL of iodinated contrast media (iopamidol, Iopamiro [300 mg of iodine per milliliter], Bracco, Milan, Italy; iopromide, Ultravist [300 mg of iodine per milliliter], Schering Diagnostics, Berlin, Germany; or iodixanol, Visipaque [270 mg of iodine per milliliter], Amersham Health, Oslo, Norway) was administered (mean, 112 mL ± 17). The amount of contrast agent was chosen according to the body weight of the subjects: 150 mL in subjects who weighed more than 110 kg, 120 mL with weight more than 75 kg, 100 mL with weight between 55 and 75 kg, and 80 mL with weight less than 55 kg. One (0.4%) patient received 150 mL of contrast material intravenously, 165 (67%) patients received 120 mL, 63 (26%) received 100 mL, and 15 (6%) received 80 mL.
A flow rate between 2 and 3 mL/sec via a cubital or antecubital venous access (18-gauge needle) was used: 2 mL/sec in 16 (6%) patients, 2.5 mL/sec in 25 (10%), 2.7 mL/sec in 156 (64%), and 3 mL/sec in 37 (15%) (mean, 2.67 mL/sec ± 0.22). As a default parameter, a scan delay of 15 seconds was used in 237 of 244 cases. The delay was 20 seconds in two patients with more distally located venous access and was 10 seconds in five patents with central venous access. Since the amount of contrast material to be administered was related to patient body weight, flow rates were adapted to deliver a uniform rate of iodine.
To avoid motion artifacts as a result of breathing, the standard procedure was to perform the examination during breath holding, which is best tolerated at end inspiration (15). For that reason, patients were instructed to breathe in and out two to three times and then finally take a deep breath and hold it as long as possible. Also, they were told that when they were no longer able to maintain the breath hold, they should breathe out slowly to minimize respiration artifacts. During the scan delay, this breathing command was given to the patient by the technician who performed the examination. The spiral scan direction was cephalocaudal in 177 (72.5%) patients and caudocephalic in 67 (27.5%).
Region-of-Interest Measurements
Images were evaluated at the working console of the CT unit. For contrast evaluation, a circular region of interest (ROI) with 344 pixels (231 mm2) was created and placed in the ascending aorta at the level of the pulmonary bifurcation and in the main pulmonary artery to measure the mean attenuation in the vessel (C.B.H., S.G., and K.F.L., with 8, 10, and 5 years of clinical practice, respectively) (Fig 1). One of the radiologists placed the ROI, the second read the attenuation value, and the third recorded it; the three radiologists alternated these responsibilities every 20–30 images. In cases (53 images) with streak artifacts caused by highly concentrated contrast material in the superior vena cava (15–17), the descending aorta was imaged instead of the ascending aorta (Fig 2a). On the basis of these measurements, patients were divided into three groups. In group 1 (n = 244), attenuation in the main pulmonary artery was higher than that in the ascending or descending aorta. In this group, no further contrast evaluation was performed. In group 2 (n = 4), mean attenuation values in the main pulmonary artery and ascending or descending aorta were equal (±5). In group 3 (n = 41), attenuation in the ascending or descending aorta was higher than that in the main pulmonary artery.
|
|
|
|
|
Diagnosis of PE
In the second assessment, the images of all 244 patients were evaluated with respect to the diagnosis of PE. Image interpretation was performed with both soft-tissue and pulmonary parenchymal window settings (center, 50 or -450 HU, respectively; window, 350 or 1,500 HU, respectively). The diagnostic criteria of PE by Remy-Jardin et al (16) were used: depiction of partial and complete filling defects, and floating, and mural thrombi. In cases of PE, central and peripheral embolisms were differentiated. PE was regarded as central when the main, right, and left pulmonary arteries were involved to the level of the proximal lobar arteries. Embolism beyond that level (distal lobar, segmental, or subsegmental) was categorized as peripheral.
Prospective Component
For the prospective part of the evaluation, the files for groups 2 and 3 were checked for preexisting echocardiographic images and the presence of a PFO. In our institution, echocardiography is performed routinely with the Valsalva maneuver to elicit the absence or presence of a PFO. Images were available in 31 of the 45 patients. The 14 patients for whom images were not available were invited for an interview; eight were interviewed. At the interview, the type and nature of the planned examination (transthoracic color Doppler echocardiography with the possible need for intravenous injection of contrast material) and possible consequences of a positive result were explained to the patients. After they gave informed consent, the patients underwent transthoracic color Doppler echocardiography (CMF 800 [2.5-MHz transducer]; Vingmed Sonotron, Horton, Norway). The examination was performed by an experienced cardiovascular radiologist (C.B.H.), who obtained subcostal four-chamber views before and during the Valsalva maneuver (14). Nonenhanced image findings were inconclusive in two patients; therefore, contrast material (Echovist; Schering Diagnostics) was administered intravenously, and image findings were conclusive. The diagnostic criterion for PFO by Dubourg et al (2) were applied: visualization of either contrast material or a color Doppler jet in the left atrium within three heartbeats of the Valsalva maneuver.
Statistical Evaluation
Mean attenuation values in the ascending or descending aorta in group 1 were compared with those in groups 2 and 3 by means of the two-tailed Student t test for unpaired data. With the same test, mean attenuation values in the main pulmonary arteries in group 1 were compared with those in groups 2 and 3. A P value of less than .05 was considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
In groups 2 and 3, the diagnosis of PE could be confirmed on the basis of its central location (right and left main pulmonary artery, or riding bifurcation embolus) in all images in 13 of the 45 patients. Overall contrast enhancement was rated as good to intermediate in the images in five patients, and findings were negative with regard to PE.
In 27 (60%) patients with abnormal contrast dynamics, image findings were indeterminate, which meant that no diagnosis could be established because of insufficient attenuation of the pulmonary arteries in more distal lobar, segmental, and subsegmental branches. In these patients, severe central PE to the proximal lobar level could also be excluded.
Of the 244 patients, PE could be diagnosed in 75 (31%) patients (44 [18%] with central PE and 31 [13%] with PE in peripheral locations). Image findings were negative for PE in 142 (58%) patients, and the diagnosis of PE was not sufficiently assessable in 27 (11%).
Prospective Investigation
In groups 2 and 3, 39 of the 45 patients underwent echocardiography (31 preexisting images were found in patient files, and eight underwent echocardiography at the interview). In our institution, echocardiography is a part of the diagnostic work-up of patients suspected of having PE to look for intracardiac thrombi and to evaluate the performance of the right side of the heart. Of the six patients lost to follow-up, two had died and four did not keep the appointment for the interview. An intracardiac right-to-left shunt was found in the 39 (16% of the whole study population) patients with early contrast enhancement of the aorta in the CT study. Of the 39 patients, three (8%) had an atrial-septal defect, and 36 (92%) had a PFO. The atrial-septal defect in one of the three patients was not detected until the time of the interview. In 10 (26%) patients, there was evidence of failure of the right side of the heart, with mild (n = 4), moderate (n = 5), or severe (n = 1) tricuspid regurgitation. These 10 patients had central PE, and the preexisting echocardiographic examination was performed at the time of spiral CT angiography in the emergency department.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
Apart from contrast material administration, another factor that may influence scan quality in a major way is respiration. It is widely agreed that patients should undergo scanning during strict apnea to avoid motion artifacts. For an average scan time of 15–25 seconds, apnea is usually best tolerated at deep inspiration. However, patients with dyspnea might not be able to maintain respiratory arrest throughout the whole examination. In such cases, scanning during shallow respiration or near end expiration can be an alternative (16,28,29). Disadvantages such as unsatisfactory quality of multiplanar reconstructions have to be considered (15,16,18).
Our study was performed to test the hypothesis that unsatisfactory pulmonary artery contrast in combination with abnormal early and strong enhancement of the thoracic aorta might be caused by PFO supported by deep inspiration. In our results, 16% of the study population had abnormal contrast dynamics on their scans, and attenuation values in the ascending aorta were significantly higher than those in the pulmonary artery. We believe that the physiologic basis for this finding is the fact that deep inspiration, similar to the Valsalva maneuver, provokes a sudden increase in right atrial pressure, which exceeds left atrial pressure. In addition, power injection of contrast material could also contribute to the build up of slightly higher right atrial pressures. The result is a short-term right-to-left shunt through an existing PFO that causes contrast material to cross the foramen ovale into the left atrium. At spiral CT angiography, this amount of contrast material is lost for pulmonary artery contrast. However, 18.5% of abnormal patterns lie slightly below the lower range of overall PFO incidence (1). Therefore, not all PFOs cause flow characteristics with important hemodynamic consequences that influence imaging properties. Another reason why a smaller than anticipated number of patients had such enhancement patterns might be that some of the study population did not follow the breathing command or did not perform a Valsalva maneuver.
One might argue that suboptimal pulmonary artery contrast and more intense aortic enhancement could be due to an error in scan delay, which usually accounts for some problems in PE diagnosis. However, it is unlikely that strong aortic contrast in the presence of weak pulmonary artery contrast and streak artifact–producing contrast material in the superior vena cava and right atrium with a scan delay of 15 seconds is a result of missing or scanning behind the bolus.
Regarding the results of the qualitative contrast evaluation in groups 2 and 3, there was a tendency toward better enhancement ratings during the early phase of the examination than in the later phases. The fact that some contrast material reached the pulmonary circulation during the scan delay, when the patient was still breathing normally, might explain this finding. The highest number of poor contrast ratings was found throughout the middle phase; this finding corresponds with the fact that the Valsalva maneuver is still strong. Toward the end of the examinations, a trend to more fair and intermediate enhancement patterns is evident. This finding has two explanations. First, patients were not able to maintain deep inspiration any longer; therefore, shunt flow is decreased and blood with a higher concentration of contrast media can once again reach the pulmonary arteries. Second, there might be additional concentrated contrast material that has been flowing through subclavian or axillary venous collateral vessels and has finally reached the superior vena cava, where it adds to the contrast enhancement. This part of the explanation also depends on the Valsalva maneuver being weaker after about 20 seconds of scan time. The evidence of shunt flow during the examination is also seen in the qualitatively assessed disparity in left atrial and pulmonary venous attenuation (Fig 4). In addition, echocardiographic proof of a shunt at the atrium level was obtained in all patients with such images.
We believe that the results of this study add to the understanding of contrast enhancement and the dynamics of pulmonary arteries in patients suspected of having PE. On the basis of our results, consequences must be considered. The most important consequence is the potential inability to make a diagnosis during the whole CT examination regarding clinically suspected PE. In the present study, results of the evaluation of spiral CT angiographic images concerning the evidence and the possibility of detecting PE show that in 27 (11%) of 244 patients, no diagnosis was possible at the more distal lobar, segmental, and subsegmental levels (30). We believe, that severe central PE in the main, right, and left pulmonary arteries can probably be diagnosed or ruled out even with images with fair or poor contrast. However, this is not possible in a more peripheral location.
The second consequence is clinical. A PFO bears the risk of potentially fatal paradoxic embolism (5–7,10,31), which occurs during an additional sudden increase in right atrial pressure in the presence of free-floating thrombi in the right atrium (32). In this sense, deep inspiration and the Valsalva maneuver may act as an abrupt pressure-increasing tool. A PFO may serve as an important predictor of adverse outcome in patients with significant PE (31). In the study by Konstantinides et al (31), the authors found that the death rate for patients with PE and PFO was more than twice as high as that for patients without a PFO. Also, the overall risk of a complicated in-hospital course was 5.2 times higher in patients with than in those without PFO, as a result of ischemic stroke and arterial embolisms (31). In their study, the authors conclude that the diagnosis of a PFO means the need for more aggressive treatment, including thrombolysis, interventional procedures, or both to restore normal hemodynamics in the right side of the heart. In a comparison of our results with those in other studies, some authors mentioned anecdotally that a PFO might explain poor pulmonary artery enhancement (15,16,18). In their interpretation, however, the elevation of right atrial pressure that leads to shunt flow is caused by an increase in pulmonary vascular resistance, as may occur in massive PE. The fact that deep inspiration and the Valsalva maneuver could also play a role in faint pulmonary artery enhancement as a result of a right-to-left shunt through a PFO has not been evaluated before, to our knowledge.
Some limitations of the present study include the predominantly retrospective manner of data collection, which accounts for differences in scan protocols, especially concerning scan delay and flow rate selection, and inconsistencies in the composition of the patient population. Consequently, a selection bias might have influenced our study results, especially concerning the high number of indeterminate images (11%) compared with findings in other studies of spiral CT angiography of the pulmonary circulation (19,23,33–36). In the literature, the percentage of inconclusive images ranges from 1% to 12% as a result of various causes besides pulmonary artery contrast for the lack of image quality. In a prospective study with clearly defined patient inclusion criteria and uniform examination protocols that include the use of a bolus tracking device (which was not available at the time this study was performed), the rate of nondiagnostic images might be considerably lower than that in the present study.
In conclusion, results of the present study indicate that in a large number of cases a PFO may lead to insufficient attenuation in the pulmonary arteries at spiral CT angiography, which would prohibit the diagnosis of PE if the examination was performed during deep inspiration. It is conceivable that this shortcoming could be overcome with a prior diagnosis of a PFO at echocardiography and subsequent CT examination during silent respiration or near end expiration. The selection of a longer scan delay to include the recirculation phase after completion of the first pass might help achieve a diagnostic scan (16). Especially since the introduction of subsecond and multi–detector row spiral CT scanners into clinical practice, the possibility of a decrease in imaging times will enable even dyspneic patients to hold their breath in expiration throughout the whole study.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, C.B.H.; study concepts, C.B.H., S.G., C.J.H., G.H.M.; study design, C.B.H., K.F.L., C.C., G.H.M.; literature research, C.B.H.; clinical studies, C.B.H., S.G., K.F.L.; data acquisition, C.B.H., S.G., G.H.M.; data analysis/interpretation, C.B.H., S.G., K.F.L., M.M.T.; statistical analysis, C.B.H., S.G.; manuscript preparation, C.B.H., S.G., C.J.H.; manuscript definition of intellectual content, C.B.H.; manuscript editing and revision/review, S.G., C.J.H., G.H.M.; manuscript final version approval, C.C., M.M.T., G.H.M., C.J.H.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. E. Williamson, J. Kirsch, P. A. Araoz, W. B. Edmister, D. D. Borgeson, J. F. Glockner, and J. F. Breen ECG-Gated Cardiac CT Angiography Using 64-MDCT for Detection of Patent Foramen Ovale Am. J. Roentgenol., April 1, 2008; 190(4): 929 - 933. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. D. Maldjian and F. S. Chew Imaging of pulmonary embolism: self-assessment module. Am. J. Roentgenol., March 1, 2006; 186(3 Suppl): S215 - S218. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. D. Maldjian, A. Anis, and M. Saric Radiological reasoning: pulmonary embolism--thinking beyond the clots. Am. J. Roentgenol., March 1, 2006; 186(3 Suppl): S219 - S223. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Merkle and R. C. Gilkeson Remnants of Fetal Circulation: Appearance on MDCT in Adults Am. J. Roentgenol., August 1, 2005; 185(2): 541 - 549. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Kim, J. B. Seo, K.-H. Do, S. Ko, S.-H. Lee, J. S. Lee, J. W. Song, T.-H. Lim, and K. S. Song Paradoxical Embolism Detected on CT Angiography and Treated With Temporary Inferior Vena Cava Filtration and Anticoagulation Am. J. Roentgenol., November 1, 2004; 183(5): 1244 - 1246. [Full Text] [PDF]
|
|
|
|
 |
|
 C. Wittram and C. B. Henk Pulmonary Artery Enhancement at CT Pulmonary Angiography [letter] * Drs Henk and Grampp respond: Radiology, December 1, 2003; 229(3): 932 - 933. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
