(Radiology. 1999;213:51-57.)
© RSNA, 1999
Ventilation-Perfusion Lung Scintigraphy as a Guide for Pulmonary Angiography in the Localization of Pulmonary Emboli1
Neil C. Davey, MD 2,
Tony P. Smith, MD,
Michael W. Hanson, MD,
Vivian S. Lee, MD, PhD 3,
Daniel J. Stackhouse, MD 4 and
R. Edward Coleman, MD
1 From the Department of Radiology, Duke University Medical Center, Rm 1502, Box 3808, Erwin Rd, Durham, NC 27710. From the 1996 RSNA scientific assembly. Received September 29, 1998; revision requested November 11; final revision received January 25, 1999; accepted April 8. Address reprint requests to T.P.S. (e-mail: smith146 @mc.duke.edu).
 |
Abstract
|
|---|
PURPOSE: To assess the appropriateness of ventilation-perfusion (V-P) scintigraphic abnormalities as a guide to pulmonary angiography for the diagnosis of pulmonary embolism (PE).
MATERIALS AND METHODS: V-P scintigrams and pulmonary angiograms of 104 patients with angiographically proved PE were reviewed by two nuclear medicine physicians and two interventional radiologists. For V-P scintigrams, the lung with the larger amount of perfusion abnormality was determined followed by identification of specific lobes. Pulmonary angiograms were similarly evaluated for lateralization and lobar distribution of PE. Conclusions were initially reached independently and subsequently by consensus.
RESULTS: Interobserver agreement for lateralization was 88% (
= 0.75) for V-P scintigraphy and 98% ( = 0.96) for pulmonary angiography. In 72 patients, V-P scintigrams predicted unilateral embolus; 64 patients underwent pulmonary angiography of the suspected side. Eight patients underwent contralateral angiography only. Of the 64 patients, 61 (95%) had PE on the predicted side at angiography. V-P scintigrams predicted lobar distribution in 55 patients. Of these, PE was found in the predicted lobe in 42 (76%).
CONCLUSION: Localization of perfusion abnormalities at V-P scintigraphy provides useful information for the interventional radiologist and serves as an accurate guide for determining the initial approach for pulmonary angiography.
Index terms: Embolism, pulmonary, 60.72 • Lung, radionuclide studies, 60.12171 • Pulmonary angiography, 60.124
|
Introduction
|
|---|
Pulmonary angiography is still considered the definitive study for the diagnosis of pulmonary embolus (1). Pulmonary angiography is an invasive procedure, however, and therefore is reserved for cases in which a confident diagnosis cannot be reached by use of other imaging modalities, such as ventilation-perfusion (V-P) lung scintigraphy. Some have suggested that V-P scintigrams may serve as a guide for pulmonary angiography (2). On the basis of the V-P scintigram, the interventional radiologist can select which lung (right or left) to inject first and can concentrate on specific areas (lobes, segments, subsegments) within this lung that are considered the most suspicious at scintigraphy and therefore the most likely to have thrombus. In a recent survey of members of the Society of Cardiovascular and Interventional Radiology, or SCVIR, 73% of the angiographers stated they would obtain V-P scintigrams in all patients prior to pulmonary angiography, and 59% would attempt to catheterize the most suspicious side first (3). Although V-P scintigraphy appears to be an accepted practice as a guide for pulmonary angiography, sparse data exist as to whether such scanning actually represents a reasonable and accurate guide. We retrospectively reviewed a series of patients with angiographically proved pulmonary embolism (PE) to determine whether V-P scintigraphy could be used as a guide in the performance of pulmonary angiography for the diagnosis of PE.
|
MATERIALS AND METHODS
|
|---|
From a computer database of all pulmonary angiograms (n = 507) obtained during a 3-year interval, 108 consecutive patients with angiographically proved PE were identified. During this same 3-year interval, 2,968 V-P scintigrams were obtained for the diagnosis of PE in our institution (Duke University Medical Center, Durham, NC). In four patients, either angiographic or V-P images could not be located for review. The remaining 104 patients (55 men, 49 women; age range, 27–86 years; mean, 61.5 years) formed the cohort for this study. The following original radiographic imaging studies were reviewed in all patients: chest radiograph obtained on the same date as the V-P scintigram, V-P scintigram, and pulmonary angiogram (Figs 1, 2). The time between V-P scintigraphy and pulmonary angiography was 0–4 days, with a mean of less than 1 day (15 hours).
View larger version (17K):
[in this window]
[in a new window]
|
Figure 1. Flowchart of 104 patients with known PE at pulmonary angiography shows the number of V-P scintigrams reviewed by two nuclear medicine physicians for lateralization (right or left side) and the number of pulmonary angiograms reviewed by two interventional radiologists for location of thrombus. The results represent consensus between reviewers. PA = pulmonary angiography.
|
|
View larger version (115K):
[in this window]
[in a new window]
|
Figure 2a. Images obtained in a 61-year-old woman with a history of ischemic cardiomyopathy and new onset of pleuritic chest pain. (a) Anteroposterior chest radiograph demonstrates a left lower lobe retrocardiac opacity, which was improving from prior images obtained during the same hospitalization and was thought to represent atelectasis. There is blunting of the left costophrenic angle, which suggests a pleural effusion. (b) Posterior breath-hold ventilation image shows decreased ventilation medially at the lung bases (arrows). Washout images revealed abnormal retention in the left upper lung. The right lung was normal. (c) Posterior perfusion image shows hypoperfusion in the left upper lung and right medial lung (arrowheads), which was thought to reflect obstructive lung disease. There is also decreased perfusion in the left base (arrows) in the region of the radiographic abnormality. The study was given an intermediate probability for pulmonary embolus. At review for this study, both observers thought that the left lung, and in particular, the left upper lobe, was more likely to have pulmonary embolus. (d-f) Frontal view of the (d) right and (e) left lungs and (f) lateral view of the left lung at pulmonary angiography. There was no thrombus identified on the right. Thrombus is present in the proximal left lower segmental branches (arrows). For the current study, this was observed individually and by consensus by the two observers. No thrombus was noted in the left upper lobe by either observer or by consensus.
|
|
View larger version (132K):
[in this window]
[in a new window]
|
Figure 2b. Images obtained in a 61-year-old woman with a history of ischemic cardiomyopathy and new onset of pleuritic chest pain. (a) Anteroposterior chest radiograph demonstrates a left lower lobe retrocardiac opacity, which was improving from prior images obtained during the same hospitalization and was thought to represent atelectasis. There is blunting of the left costophrenic angle, which suggests a pleural effusion. (b) Posterior breath-hold ventilation image shows decreased ventilation medially at the lung bases (arrows). Washout images revealed abnormal retention in the left upper lung. The right lung was normal. (c) Posterior perfusion image shows hypoperfusion in the left upper lung and right medial lung (arrowheads), which was thought to reflect obstructive lung disease. There is also decreased perfusion in the left base (arrows) in the region of the radiographic abnormality. The study was given an intermediate probability for pulmonary embolus. At review for this study, both observers thought that the left lung, and in particular, the left upper lobe, was more likely to have pulmonary embolus. (d-f) Frontal view of the (d) right and (e) left lungs and (f) lateral view of the left lung at pulmonary angiography. There was no thrombus identified on the right. Thrombus is present in the proximal left lower segmental branches (arrows). For the current study, this was observed individually and by consensus by the two observers. No thrombus was noted in the left upper lobe by either observer or by consensus.
|
|
View larger version (132K):
[in this window]
[in a new window]
|
Figure 2c. Images obtained in a 61-year-old woman with a history of ischemic cardiomyopathy and new onset of pleuritic chest pain. (a) Anteroposterior chest radiograph demonstrates a left lower lobe retrocardiac opacity, which was improving from prior images obtained during the same hospitalization and was thought to represent atelectasis. There is blunting of the left costophrenic angle, which suggests a pleural effusion. (b) Posterior breath-hold ventilation image shows decreased ventilation medially at the lung bases (arrows). Washout images revealed abnormal retention in the left upper lung. The right lung was normal. (c) Posterior perfusion image shows hypoperfusion in the left upper lung and right medial lung (arrowheads), which was thought to reflect obstructive lung disease. There is also decreased perfusion in the left base (arrows) in the region of the radiographic abnormality. The study was given an intermediate probability for pulmonary embolus. At review for this study, both observers thought that the left lung, and in particular, the left upper lobe, was more likely to have pulmonary embolus. (d-f) Frontal view of the (d) right and (e) left lungs and (f) lateral view of the left lung at pulmonary angiography. There was no thrombus identified on the right. Thrombus is present in the proximal left lower segmental branches (arrows). For the current study, this was observed individually and by consensus by the two observers. No thrombus was noted in the left upper lobe by either observer or by consensus.
|
|
View larger version (125K):
[in this window]
[in a new window]
|
Figure 2d. Images obtained in a 61-year-old woman with a history of ischemic cardiomyopathy and new onset of pleuritic chest pain. (a) Anteroposterior chest radiograph demonstrates a left lower lobe retrocardiac opacity, which was improving from prior images obtained during the same hospitalization and was thought to represent atelectasis. There is blunting of the left costophrenic angle, which suggests a pleural effusion. (b) Posterior breath-hold ventilation image shows decreased ventilation medially at the lung bases (arrows). Washout images revealed abnormal retention in the left upper lung. The right lung was normal. (c) Posterior perfusion image shows hypoperfusion in the left upper lung and right medial lung (arrowheads), which was thought to reflect obstructive lung disease. There is also decreased perfusion in the left base (arrows) in the region of the radiographic abnormality. The study was given an intermediate probability for pulmonary embolus. At review for this study, both observers thought that the left lung, and in particular, the left upper lobe, was more likely to have pulmonary embolus. (d-f) Frontal view of the (d) right and (e) left lungs and (f) lateral view of the left lung at pulmonary angiography. There was no thrombus identified on the right. Thrombus is present in the proximal left lower segmental branches (arrows). For the current study, this was observed individually and by consensus by the two observers. No thrombus was noted in the left upper lobe by either observer or by consensus.
|
|
View larger version (130K):
[in this window]
[in a new window]
|
Figure 2e. Images obtained in a 61-year-old woman with a history of ischemic cardiomyopathy and new onset of pleuritic chest pain. (a) Anteroposterior chest radiograph demonstrates a left lower lobe retrocardiac opacity, which was improving from prior images obtained during the same hospitalization and was thought to represent atelectasis. There is blunting of the left costophrenic angle, which suggests a pleural effusion. (b) Posterior breath-hold ventilation image shows decreased ventilation medially at the lung bases (arrows). Washout images revealed abnormal retention in the left upper lung. The right lung was normal. (c) Posterior perfusion image shows hypoperfusion in the left upper lung and right medial lung (arrowheads), which was thought to reflect obstructive lung disease. There is also decreased perfusion in the left base (arrows) in the region of the radiographic abnormality. The study was given an intermediate probability for pulmonary embolus. At review for this study, both observers thought that the left lung, and in particular, the left upper lobe, was more likely to have pulmonary embolus. (d-f) Frontal view of the (d) right and (e) left lungs and (f) lateral view of the left lung at pulmonary angiography. There was no thrombus identified on the right. Thrombus is present in the proximal left lower segmental branches (arrows). For the current study, this was observed individually and by consensus by the two observers. No thrombus was noted in the left upper lobe by either observer or by consensus.
|
|
View larger version (125K):
[in this window]
[in a new window]
|
Figure 2f. Images obtained in a 61-year-old woman with a history of ischemic cardiomyopathy and new onset of pleuritic chest pain. (a) Anteroposterior chest radiograph demonstrates a left lower lobe retrocardiac opacity, which was improving from prior images obtained during the same hospitalization and was thought to represent atelectasis. There is blunting of the left costophrenic angle, which suggests a pleural effusion. (b) Posterior breath-hold ventilation image shows decreased ventilation medially at the lung bases (arrows). Washout images revealed abnormal retention in the left upper lung. The right lung was normal. (c) Posterior perfusion image shows hypoperfusion in the left upper lung and right medial lung (arrowheads), which was thought to reflect obstructive lung disease. There is also decreased perfusion in the left base (arrows) in the region of the radiographic abnormality. The study was given an intermediate probability for pulmonary embolus. At review for this study, both observers thought that the left lung, and in particular, the left upper lobe, was more likely to have pulmonary embolus. (d-f) Frontal view of the (d) right and (e) left lungs and (f) lateral view of the left lung at pulmonary angiography. There was no thrombus identified on the right. Thrombus is present in the proximal left lower segmental branches (arrows). For the current study, this was observed individually and by consensus by the two observers. No thrombus was noted in the left upper lobe by either observer or by consensus.
|
|
V-P imaging was performed by using standard methods. Specifically, ventilation imaging was performed after the administration of approximately 10–20 mCi of xenon 133 gas by means of inhalation. Posterior images were obtained during breath-hold, equilibrium, and washout phases. Perfusion imaging was performed in at least six projections after an intravenous injection of 4 mCi of technetium 99m–macroaggregated albumin, or 99mTc-MAA.
Pulmonary angiography was performed by using conventional methods (4). All angiograms were obtained with use of selective injection by means of a 7-F pigtail catheter placed into either the right or left pulmonary artery. Pulmonary arterial pressures were measured, and angiographic images were obtained as determined by the attending interventional radiologist by using either screen-film or digital techniques or a combination of the two. Our routine clinical practice is to catheterize first the side with the greatest suspicion for PE at V-P scintigraphy. Pulmonary angiograms diagnostic for PE on the first side typically precluded performance of angiography in the contralateral side. Reasons for performing angiography in the contralateral side included the preference of the interventional radiologist, determination of thrombus burden for clinical purposes, or uncertainty of the diagnosis on the basis of the first side.
V-P scintigrams and corresponding chest radiographs were reviewed retrospectively by two nuclear medicine physicians (M.W.H., R.E.C.). The chest radiograph was used only for V-P scintigraphic interpretation. With the understanding that pulmonary angiography requires either the right or the left side to be studied initially, these physicians were instructed to determine whether one lung had a greater likelihood of PE on the basis of overall perfusion defects, areas of V-P mismatching, and correlation with the chest radiograph. The lung with the largest degree of abnormality was identified, or both lungs were determined to be equivalent in their extent of abnormality. When there was lateralization of abnormality to one lung, a determination was made as to whether the areas of abnormality were greatest in a particular lobe: on the right in the upper, middle, or lower lobes or on the left in the upper or lower lobes. If a single lobe did not demonstrate a greater degree of perfusion abnormality, then the affected lobes were considered to have equivalent likelihood for PE. Conclusions were reached independently and subsequently by consensus between the readers.
Pulmonary angiograms were reviewed retrospectively by two interventional radiologists (D.J.S., T.P.S.). As with the V-P scintigrams, conclusions were first reached independently and then subsequently by consensus. The pulmonary angiograms were assessed for the presence of emboli and for the location of emboli with respect to lung (right and left side) and to specific lobe. In 77 patients (74%), emboli were identified following unilateral pulmonary angiography, and the study was terminated. Bilateral pulmonary angiograms were obtained in 27 patients (26%). A total of 131 lungs (sides) and 341 lobes were studied angiographically.
The agreement between observers in determining lateralization of abnormalities was determined for V-P scintigrams and pulmonary angiograms. The statistic, an intraclass correlation coefficient, and its 95% CI were computed to take into consideration the probability that the agreement between pairs of observers was due to chance. The value assumes its maximum value of 1 for perfect agreement, and values greater than 0.75 were taken to represent agreement beyond chance (5). A P value less than .05 was considered to indicate a statistically significant difference.
Consensus interpretation of pulmonary angiographic findings was compared with consensus interpretation of the V-P scintigraphic findings. The positive predictive values of V-P scintigraphy in lateralizing abnormalities and localizing emboli to specific lobes was then determined by using consensus angiography as the standard.
|
RESULTS
|
|---|
At the time the V-P scintigrams were obtained, the probability of PE was estimated on the basis of published modified Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) criteria (6). The distribution of results reflects the patient population usually referred for pulmonary angiography in our institution with 22 (21%) high-, 74 (71%) intermediate-, and eight (8%) low-probability V-P scintigrams.
The results are presented in flowchart form in Figure 1. At scintigraphy, perfusion abnormalities were lateralized in 72 (69%) patients: to the right lung in 56 patients (54%) and to the left in 16 patients (15%) (Fig 1). Bilateral, equal perfusion abnormalities were observed in 32 patients (31%). Interobserver agreement for lateralization was calculated at 88% with a of 0.75 (P < .001), which represents agreement beyond chance. There was no statistically significant difference in interobserver agreement for right and left lung abnormalities ( = 0.67 and 0.73, respectively).
For the 27 patients in whom bilateral pulmonary angiography was performed, the interobserver agreement for angiographic lateralization of the abnormality was 98%; the corresponding of 0.96 confirmed near perfect agreement. Overall, at pulmonary angiography consensus reading, pulmonary emboli were identified on the right in 62 patients (60%), on the left in 34 patients (33%), and bilaterally in eight patients (8%).
At V-P scintigraphy, perfusion abnormalities were localized in 136 lungs—32 bilateral, 56 right, and 16 left. Of the 72 patients in whom V-P scintigraphy was used to predict unilateral embolus, 57 (79%) underwent unilateral pulmonary angiography of the suspected side, seven (10%) underwent bilateral pulmonary angiography, and eight (11%) had the procedure terminated after embolus was diagnosed at contralateral pulmonary angiography (Fig 1). Of the 64 patients with lateralizing V-P defects who underwent pulmonary angiography of the suspected lung, 61 (95%) were found at pulmonary angiography to have PE on the predicted side (Fig 2), and three patients (5%) had embolus in only the contralateral lung (Fig 3). Thus, the overall positive predictive value of V-P scintigraphy for lateralization of PE was 95%.
View larger version (156K):
[in this window]
[in a new window]
|
Figure 3a. Images obtained in a 27-year-old woman with shortness of breath several days following a motor vehicle accident. The chest radiograph obtained on the same day as the V-P images revealed air-space disease in the left lung base with a large left and a small right pleural effusion. (a) Posterior breath-hold ventilation image shows markedly diminished ventilation in the left lung. The patient was unable to cooperate fully due to shortness of breath and was breathing during image acquisition. (b) Posterior perfusion image shows bilateral perfusion abnormalities in both lung bases (arrows). There is also a subsegmental perfusion defect in the lateral aspect of the right lung (arrowhead). The study was placed in the intermediate category for pulmonary embolus. Because of the markedly decreased ventilation pattern in the left lung, and with the more focal subsegmental perfusion abnormalities in the right, both observers thought the right lung was more likely to demonstrate PE. Frontal pulmonary angiograms of the (c) right and (d) left lungs demonstrate no evidence of PE in the right and a filling defect representing PE in the posterior basal segment of the left lower lobe (arrows). These findings were made individually by the two observers and by consensus.
|
|
View larger version (150K):
[in this window]
[in a new window]
|
Figure 3b. Images obtained in a 27-year-old woman with shortness of breath several days following a motor vehicle accident. The chest radiograph obtained on the same day as the V-P images revealed air-space disease in the left lung base with a large left and a small right pleural effusion. (a) Posterior breath-hold ventilation image shows markedly diminished ventilation in the left lung. The patient was unable to cooperate fully due to shortness of breath and was breathing during image acquisition. (b) Posterior perfusion image shows bilateral perfusion abnormalities in both lung bases (arrows). There is also a subsegmental perfusion defect in the lateral aspect of the right lung (arrowhead). The study was placed in the intermediate category for pulmonary embolus. Because of the markedly decreased ventilation pattern in the left lung, and with the more focal subsegmental perfusion abnormalities in the right, both observers thought the right lung was more likely to demonstrate PE. Frontal pulmonary angiograms of the (c) right and (d) left lungs demonstrate no evidence of PE in the right and a filling defect representing PE in the posterior basal segment of the left lower lobe (arrows). These findings were made individually by the two observers and by consensus.
|
|
View larger version (165K):
[in this window]
[in a new window]
|
Figure 3c. Images obtained in a 27-year-old woman with shortness of breath several days following a motor vehicle accident. The chest radiograph obtained on the same day as the V-P images revealed air-space disease in the left lung base with a large left and a small right pleural effusion. (a) Posterior breath-hold ventilation image shows markedly diminished ventilation in the left lung. The patient was unable to cooperate fully due to shortness of breath and was breathing during image acquisition. (b) Posterior perfusion image shows bilateral perfusion abnormalities in both lung bases (arrows). There is also a subsegmental perfusion defect in the lateral aspect of the right lung (arrowhead). The study was placed in the intermediate category for pulmonary embolus. Because of the markedly decreased ventilation pattern in the left lung, and with the more focal subsegmental perfusion abnormalities in the right, both observers thought the right lung was more likely to demonstrate PE. Frontal pulmonary angiograms of the (c) right and (d) left lungs demonstrate no evidence of PE in the right and a filling defect representing PE in the posterior basal segment of the left lower lobe (arrows). These findings were made individually by the two observers and by consensus.
|
|
View larger version (152K):
[in this window]
[in a new window]
|
Figure 3d. Images obtained in a 27-year-old woman with shortness of breath several days following a motor vehicle accident. The chest radiograph obtained on the same day as the V-P images revealed air-space disease in the left lung base with a large left and a small right pleural effusion. (a) Posterior breath-hold ventilation image shows markedly diminished ventilation in the left lung. The patient was unable to cooperate fully due to shortness of breath and was breathing during image acquisition. (b) Posterior perfusion image shows bilateral perfusion abnormalities in both lung bases (arrows). There is also a subsegmental perfusion defect in the lateral aspect of the right lung (arrowhead). The study was placed in the intermediate category for pulmonary embolus. Because of the markedly decreased ventilation pattern in the left lung, and with the more focal subsegmental perfusion abnormalities in the right, both observers thought the right lung was more likely to demonstrate PE. Frontal pulmonary angiograms of the (c) right and (d) left lungs demonstrate no evidence of PE in the right and a filling defect representing PE in the posterior basal segment of the left lower lobe (arrows). These findings were made individually by the two observers and by consensus.
|
|
Because these results may be biased by the exclusion of the eight patients who underwent only contralateral injection, we also calculated the positive predictive value assuming all eight had contralateral disease with no embolus in the lung lateralized at V-P scintigraphy. The lower limits of the positive predictive value for lateralization at V-P scintigraphy would then be 85% (61 of 72).
Twenty-seven patients underwent bilateral pulmonary angiography. V-P lung scintigraphic probability in this population was high in five, low in six, and intermediate in 16. V-P scintigraphy lateralized to a side in 16 patients, and both sides were thought to hold an equal likelihood of PE in 11. Pulmonary angiography showed PE limited to the right lung in 11 patients, limited to the left lung in eight, and bilaterally in eight. Of the 16 that lateralized, nine were to the right lung where pulmonary angiography indicated thrombus only on the right in six, bilaterally in two, and only in the left in one. Of the remaining seven that lateralized at V-P scintigraphy to the left, three had PE only in the left lung, two had PE bilaterally, and two had PE only in the right lung.
In the three patients in whom the V-P scintigram predicted an abnormality in one lung and pulmonary angiography showed embolus in the contralateral lung, no common characteristics were identified. There were two women and one man aged 27, 57, and 58 years. Two scintigrams were interpreted as intermediate- and one as low-probability for pulmonary embolus. Two of the pulmonary angiograms were obtained within 1 day of the V-P scintigram, and one was obtained within 2 days. The PE identified at pulmonary angiography was on the left side in two patients and on the right in one. In all three patients, PE was identified as a filling defect in the contrast material–filled vessel and was therefore not completely occlusive at the levels visualized at angiography.
The value of scintigraphy in predicting lobar distribution of pulmonary emboli was examined. All readers independently classified perfusion or angiographic abnormalities as involving one or more of the five pulmonary lobes. An interobserver agreement of 91% was recorded for lobar localization at V-P scintigraphy (n = 520 lobes examined) with a of 0.67 (P < .001), which indicates reasonable agreement. For pulmonary angiography, an interobserver agreement of 95% was recorded (n = 341 lobes examined) with a of 0.90 (P < .001), which represents excellent agreement. There was no significant difference in agreement with either technique when data for right and left lungs were evaluated separately.
Although agreement between observers was significant ( = 0.67; P < .001), V-P scintigraphy was somewhat less predictive of lobar distribution of pulmonary emboli. Sixty-two (60%) of 104 patients had pulmonary emboli lateralized to a specific lobe by consensus interpretation of V-P scintigrams. Excluding the seven (11%) patients who underwent angiography of only the contralateral lung, embolus was found in the predicted lobe in 42 of 55 patients for a positive predictive value of 76%. PE was not present in the predicted lobe in 13 (24%). In the eight patients in whom lateralization was achieved without localization to a specific lobe, PE was found in more than one lobe in six patients. The other two patients underwent only contralateral angiography.
The interpretation of high, intermediate, or low probability of PE on the basis of V-P scintigrams was not a significant factor in the accuracy of scintigraphy for predicting lobar distribution of emboli (83%, 69%, and 65%, respectively), although a trend was observed.
|
DISCUSSION
|
|---|
PE continues to be a condition of high morbidity and mortality implicated in approximately 50,000 deaths annually in the United States (7). The diagnosis of PE remains difficult. Stein and Henry (8) noted that among all patients who had PE at autopsy in a general hospital, only 12% had an antemortem suspicion for or diagnosis of PE. This is in part because the clinical signs and symptoms of PE are nonspecific (9,10). There is certainly a need for a reliable diagnostic test. Contrast material–enhanced computed tomography (CT) is becoming popular, and results of early studies (11,12) confirm its accuracy at least to the segmental level. Magnetic resonance (MR) angiography has begun to show promising results as well (13). Nonetheless, V-P scintigraphy remains the most used initial imaging study for the diagnosis of pulmonary embolus, and pulmonary angiography remains the definitive diagnostic test.
Interventional radiologists request V-P scintigraphy prior to pulmonary angiography for two reasons. The first reason is that V-P scintigrams may be sufficiently diagnostic to allow treatment to be initiated confidently without angiography. This is particularly true when studies can be categorized as showing normal or high probability of PE and when they confirm clinical suspicion. Two situations often arise, however, in which additional diagnostic confirmation is sought: nondiagnostic V-P scintigrams and those in which the scintigram does not correlate with clinical suspicion. Regarding the former, 71% of the patients in our series had intermediate-probability interpretations of their V-P scintigrams, and angiography was performed for the diagnosis of PE. The other 27% of the V-P scintigrams had a high or low suspicion for PE, and angiography was requested on the basis of a high clinical suspicion. The second reason for V-P scintigraphy prior to angiography is that interventional radiologists believe V-P scintigraphy can serve as a guide for pulmonary angiography and thereby theoretically improve the accuracy and decrease the extent of pulmonary angiography.
The use of V-P scintigraphy as a guide for pulmonary angiography is not based on a great deal of published data. Meyer et al (14) examined 34 consecutive patients before and after thrombolytic therapy for PE to determine the relationship between perfusion scintigram defects and angiographic severity. Their study results showed perfusion scintigraphy to be a reliable method of assessing pulmonary vascular obstruction. However, when embolic involvement was extensive, showing greater than 50% obstruction at angiography, the angiographic defects were underestimated on the scintigram. It should be noted that these findings were not in keeping with earlier thrombolytic data in which a lack of agreement between V-P scintigraphy and pulmonary angiography was observed (15–17).
Gottschalk et al (18) reviewed the PIOPED data and found that, excluding high-probability studies, if PE was excluded by means of angiography on the side of the abnormal V-P scintigram, the probability of PE on the side that appeared normal on the V-P scintigram was 5%. These conclusions, however, are based on a small number of patients (19 patients), despite the large number of patients in the PIOPED study. Even prior to the publication of these results, the survey results from 1993 of the Society of Cardiovascular and Interventional Radiology suggest that many interventional radiologists based their practice on similar conclusions (3). In this survey, 39% of interventional radiologists said they would terminate pulmonary angiography once PE was excluded on the scintigraphically abnormal side. In the current study, lateralization at V-P scintigraphy agreed with pulmonary angiographic findings in 61 of 64 patients who underwent angiography of the most suspicious lung, which resulted in an overall positive predictive value of 95%. There were three patients in whom pulmonary angiography demonstrated PE in the contralateral, or less abnormal, lung at scintigraphy. On the more abnormal side at scintigraphy, pulmonary angiography showed chronic PE in one patient and normal pulmonary arteries in the other two. Our results, therefore, also suggest that in a small number of patients, angiography limited to the most suspicious side by means of V-P scintigraphy could lead to false-negative diagnoses at pulmonary angiography.
The retrospective nature of this study results in two main limitations. First, eight patients underwent angiography that showed PE in the lung contralateral to the one that showed greater perfusion abnormality at scintigraphy, and in these patients, the second side was not injected, as the findings would not have altered patient care. These incomplete data limit precise determination of the positive predictive value for V-P lateralization of pulmonary embolus in this study. Second, definitive angiography for the total distribution of PE within the lungs was not performed. Once thrombus was diagnosed, the study was usually terminated, and no efforts were made to visualize all segments in all lobes. Although there was excellent interobserver agreement for lobar localization of abnormalities at both V-P scintigraphy and pulmonary angiography, the overall positive predictive value was only 76% for V-P scintigraphy in predicting lobar distribution of PE. This may be in part due to limited angiographic evaluation.
It is important to consider the frequency with which PE is a bilateral disease, because, in these cases, V-P guidance would be unnecessary. While it is often presumed that PE is always a bilateral disease, in their review of 500 autopsies in cases of PE, Morpurgo and Schmid (19) found PE to be unilateral in 38.5% of cases, and infarction occurred unilaterally in 56.8%. These findings also have been noted angiographically. In a recent review of the anatomic distribution of PE at pulmonary angiography, Oser et al (20) found that of 33 patients who underwent bilateral pulmonary angiography, emboli were present unilaterally in 54% of patients.
Only patients with angiographically proved PE were entered in this study. From this information, therefore, one cannot predict the lateralization patterns in patients without PE, that is, how often V-P scintigrams of those without PE would lateralize. This, however, was not an issue in the current study, nor should it be one for the practicing interventional radiologist. Because current angiographic techniques usually are limited to injection of one lung at a time, the interventional radiologist must decide which lung to examine first. This situation may not be as much of an issue in the future, as other imaging modalities such as CT and MR imaging are applied more readily to the diagnosis of PE.
During this study, CT was not being used for the diagnosis of PE in our institution. It is unclear, therefore, how CT would have effected the results of this study. Since the completion of this study, however, CT has become very popular for the diagnosis of PE (11,12). It is used most often when V-P scintigraphy is expected to be nondiagnostic, such as with an abnormal chest radiograph or where it is difficult to perform good ventilation imaging due to poor breath holding or patients receiving ventilatory support (21). In patients who already have undergone V-P scintigraphy and in whom CT is being performed for further diagnostic clarification, the results presented here show that V-P scintigraphy could be used as a guide to areas suspicious for PE. Such information does not change current CT techniques but could guide radiologists to areas for greater scrutiny. CT has been reported to have difficulties with the depiction of segmental and subsegmental emboli, and whether the data presented here could improve the diagnostic capabilities of CT for this type of embolic disease is unknown (22). In addition, further advances in other imaging modalities such as MR may be of greater diagnostic value if the study can concentrate on areas of greatest scintigraphic abnormality.
We found that in 95% of patients, PE was diagnosed on the most suspicious side as demonstrated at V-P scintigraphy. In addition, PE also was present in the predicted lobe in 76% of those patients in whom there was lateralization to a side. Our results show that V-P scintigraphy can provide an accurate guide to the selection of the first side to study by means of pulmonary angiography. In particular, when the interventional radiologist is faced with limiting pulmonary angiography to a single lung, such as in patients who are very ill and who may not tolerate the study well or in whom limited contrast material or fluid can be administered, V-P scintigraphic guidance may prove vital.
|
Footnotes
|
|---|
2 Current address: Department of Medical Imaging, St Alphonsus Regional Medical Center, Boise, Idaho. 
3 Current address: Department of Radiology-MRI, New York University Medical Center, NY.
4 Current address: Charlotte Radiology, Charlotte, NC.
Abbreviations: PE = pulmonary embolism
PIOPED = Prospective Investigation of Pulmonary Embolism Diagnosis
V-P = ventilation-perfusion
Author contributions: Guarantor of integrity of entire study, T.P.S.; study concepts, T.P.S., R.E.C.; study design, N.C.D., T.P.S., R.E.C.; definition of intellectual content, T.P.S.; literature research, N.C.D., T.P.S.; clinical studies, N.C.D., T.P.S.; data acquisition, T.P.S., M.W.H., D.J.S., R.E.C.; data analysis, N.C.D., T.P.S., V.S.L., R.E.C.; statistical analysis, V.S.L.; manuscript preparation, N.C.D., T.P.S., V.S.L.; manuscript editing and review, all authors
|
References
|
|---|
-
Greenspan RH. Pulmonary angiography and the diagnosis of pulmonary embolism. Prog Cardiovasc Dis 1994; 37:93-106.[Medline]
-
Newman GE. Pulmonary angiography in pulmonary embolic disease. J Thorac Imaging 1989; 4:28-39.[Medline]
-
Tulchinsky M, Eggli DF, Mahraj RPM, Cardella JF. Influence of V/Q scan results on pulmonary angiography procedure (abstr). J Nucl Med 1993; 34:121P.
-
Hudson ER, Smith TP, McDermott VG, et al. Pulmonary angiography performed with iopamidol: complications in 1,434 patients. Radiology 1996; 198:61-65.[Abstract/Free Full Text]
-
Fleiss JL. The measurement of interrater agreement. Statistical methods for rates and proportions 2nd ed. New York, NY: Wiley, 1981; 212-236.
-
Sostman HD, Coleman RE, DeLong DM, Newman GE, Paine S. Evaluation of revised criteria for ventilation-perfusion scintigraphy in patients with suspected pulmonary embolism. Radiology 1994; 193:103-107.[Abstract/Free Full Text]
-
Alpert JS, Dalen JE. Progress in cardiovascular diseases. Prog Cardiovasc Dis 1994; 36:417-422.[Medline]
-
Stein PD, Henry JW. Prevalence of acute pulmonary embolism among patients in a general hospital and at autopsy. Chest 1995; 108:978-981.[Abstract/Free Full Text]
-
Morgenthaler TI, Ryu JH. Clinical characteristics of fatal pulmonary embolism in a referral hospital. Mayo Clin Proc 1995; 70:417-424.[Medline]
-
The PIOPED Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism. JAMA 1990; 263:2753-2759.[Abstract]
-
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]
-
van Rossum AB, Pattynama PMT, Tjin ER, et al. Pulmonary embolism: validation of spiral CT angiography in 149 patients. Radiology 1996; 201:467-470.[Abstract/Free Full Text]
-
Meaney JF, Weg JG, Chenvert TL, Stafford-Johnson D, Hamilton BH, Prince MR. Diagnosis of pulmonary embolism with magnetic resonance angiography. N Engl J Med 1997; 333:1422-1427.
-
Meyer G, Collignon MA, Guinet F, Jeffrey AA, Barritault L, Sors H. Comparison of perfusion lung scanning and angiography in the estimation of vascular obstruction in acute pulmonary embolism. Eur J Nucl Med 1990; 17:315-319.[Medline]
-
UPET Investigators [Urokinase Pulmonary Embolism Trial]. The urokinase pulmonary embolism trial: a national cooperative study. Circulation 1973; 47(suppl 2):1-108.[Free Full Text]
-
Goldhaber SZ, Kessler CM, Heit J, et al. Randomized controlled trial of recombinant tissue plasminogen activator versus urokinase in the treatment of acute pulmonary embolism. Lancet 1988; 8606:293-298.
-
Parker JA, Markis JE, Palla A, et al. Pulmonary perfusion after rt-PA therapy for acute embolism: early improvement assessed with segmental perfusion scanning. Radiology 1988; 166:441-445.[Abstract/Free Full Text]
-
Gottschalk A, Stein PD, Henry JW, Relyea B. Can pulmonary angiography be limited to the most suspicious side if the contralateral side appears normal on the ventilation/perfusion lung scan?. Chest 1996; 110:392-394.[Abstract/Free Full Text]
-
Morpurgo M, Schmid C. Clinico-pathological correlations in pulmonary embolism: a posteriori evaluation. Prog Respir Res 1980; 13:8-15.
-
Oser RF, Zuckerman DA, Gutierrez FR, Brink JA. Anatomic distribution of pulmonary emboli at pulmonary angiography: implications for cross-sectional imaging. Radiology 1996; 199:31-35.[Abstract/Free Full Text]
-
Mayo JR, Remy-Jardin M, Müller NL, et al. Pulmonary embolism: prospective comparison of spiral CT with ventilation-perfusion scintigraphy. Radiology 1997; 205:447-452.[Abstract/Free Full Text]
-
Drucker EA, Rivitz SM, Shepard JAO, et al. Acute pulmonary embolism: assessment of helical CT for diagnosis. Radiology 1998; 209:235-241.[Abstract/Free Full Text]
TOP
Abstract
Introduction
