| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Emergency Radiology |
1 From the Departments of Radiology (E.C., P.G., L.G., F.H.), Emergency Medicine (F.V., F.Z., P.M., M.S.R.), Nuclear Medicine (A.K.), Internal Medicine (P.H.), and Clinical Biology Haemostasis Unit (E.L.), Cliniques Universitaires St-Luc, Université Catholique de Louvain, Av Hippocrate 10, 1200 Brussels, Belgium. Received July 16, 2002; revision requested September 3; final revision received March 6, 2003, accepted April 3. Address correspondence to E.C. (e-mail: coche@rdgn.ucl.ac.be).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Ninety-four nonconsecutive patients, in whom acute PE was suspected, underwent thin-collimation multi–detector row CT (collimation, 4 x 1 mm; pitch, 1.25; scanning time, 0.5 second) and V-P scintigraphy. Concordance between CT and scintigraphic images was used in the diagnosis of PE. Pulmonary angiography was performed within 24 hours if interpretations of V-P and spiral CT images were inconclusive or discordant. Sensitivity and specificity values were calculated for V-P scintigrams and CT scans of the lungs. The rates of conclusive results for scintigraphy and CT were compared.
RESULTS: The sensitivity of thin-collimation multi–detector row CT and V-P scintigraphy for the detection of PE was 96% (27 of 28; CI: 82%, 99%) and 98% (65 of 66; CI: 92%, 99%), respectively. The specificity of CT and V-P scintigraphy was 86% (24 of 28; CI: 67%, 96%) and 88% (58 of 66; CI: 77%, 94%), respectively. Seven V-P scintigrams were of intermediate probability, and one spiral CT study was indeterminate. Examinations with spiral CT yielded conclusive results more often than examinations with planar V-P scintigraphy (P < .05). Five V-P scintigrams and spiral CT scans were discordant. Twelve pulmonary angiographic examinations were performed. Angiographic findings were concordant in 10 (91%) of 11 patients with conclusive CT scans in whom pulmonary angiography was attempted. CT was used to establish an alternative diagnosis in 19 (29%) of 66 patients in whom PE was excluded.
CONCLUSION: Thin-collimation multi–detector row CT is more accurate than V-P scintigraphy in the diagnosis of acute PE in outpatients. Furthermore, CT provides alternative diagnoses for patients without PE on high-quality transverse or near-isotropic reformatted images.
© RSNA, 2003
Index terms: Computed tomography (CT), multi–detector row, 60.12115 • Embolism, pulmonary, 60.72 • Lung, radionuclide studies, 60.1216
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The recent advent of multi–detector row spiral CT, which can be used to examine the entire chest with 1-mm section thickness during one breath hold, further improves the accuracy of spiral CT in the diagnosis of PE. Several authors (10–12) have reported a marked improvement in the analysis of subsegmental arteries in routine clinical practice with better interobserver agreement for the diagnosis of PE (10). To our knowledge, there is no published clinical study that compares thin-collimation multi–detector row CT with ventilation-perfusion (V-P) scintigraphy in patients who present to the emergency department. The purpose of this prospective study was to compare multi–detector row CT and V-P scintigraphy in the diagnosis of acute PE in patients who were cared for in the emergency department.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The physician on service in the emergency department at the time of presentation clinically assessed all patients. A pretest clinical probability was performed and categorized as low, moderate, or high according to the clinical scoring validated by Wells et al (13). Plasma D-dimer assays (Vidas D-dimer; BioMérieux, Marcy l’Etoile, France) were performed for each patient in undiluted plasma samples and analyzed in the coagulation laboratory by a technician who was unaware of any clinical information. The results were communicated by phone to the physician in charge of patient care within 1 hour.
Of the 841 consecutive patients, 747 (434 women and 313 men; mean age, 60 years ± 18) were excluded from the study for various clinical and personal reasons, including D-dimer test that was negative (n = 403), as recommended by Perrier et al (14); clinical signs of deep venous thrombosis (n = 45); D-dimer values that were positive with an obvious alternative diagnosis (n = 151); incomplete study protocols (n = 76); contraindications to spiral CT (n = 40); patient transfer (n = 11); death (n = 2); and patient refusal or inability to participate (n = 19). Thus, the final study group consisted of 94 nonconsecutive patients (66 women and 28 men). Mean patient age was 62 years ± 18 (range, 21–91 years). Among the final study group, 16 (17%) of 94 patients returned home on the same day, 16 (17%) were admitted to the 1-day clinic, and 62 (66%) were subsequently admitted as inpatients to complete their work-up and treatment. All studies were performed in accordance with the guidelines of the institutional review board. Written consent was obtained from all patients.
Imaging Evaluation
Each patient enrolled in the study underwent thin-collimation multi–detector row spiral CT, chest radiography, and planar V-P scintigraphy within 24 hours of each other. When findings at V-P scintigraphy or spiral CT did not indicate a diagnosis or when findings were discordant, pulmonary angiography was performed (n = 12). The entire diagnostic work-up was performed within a 48-hour period.
Multi–detector row spiral CT.—Spiral CT angiography was performed with a multi– detector row spiral CT scanner (Mx8000; Philips Medical Systems, Cleveland, Ohio) with a collimation of 4 x 1 mm, a pitch of 1.25, a scan time of 0.5 second, 120 kV, and 144 mAs. Overlapping reconstructions were performed for all data sets. Transverse images were reconstructed with an interval of 0.6 mm by using a 180° linear-interpolation algorithm with a medium sharp filter and a matrix of 512 x 512 pixels. All studies were performed with 70–100 mL of iobitridol (Xenetix 350; Guerbet, Aulnay-sous-bois, France), diluted with 20–30 mL of saline and administered at a rate of 3 mL/sec with a variable start delay. The iodinated contrast material was administered with an automatic injector (CT9000 ADV; Liebel-Flarsheim, Cincinnati, Ohio) through a peripheral 18-gauge venous access needle that was placed in an antecubital vein while the arm was in supination. Start delay time was determined with a test injection of 25 mL of contrast material at a rate of 3 mL/sec. Time-attenuation curves over the right pulmonary artery were produced from images (120 kV, 30 mAs, 10-mm section thickness) obtained every second for 20 seconds. Start delay was determined by adding 2 seconds to the time-to-peak value. The patients were instructed to hold their breath for the duration of the study. If they were unable to do so, they were instructed to breathe as quietly as possible.
All patients underwent craniocaudal imaging. The z-axis coverage and the field of view were chosen to include the entire thorax, from the apex to the base of the lungs. The field of view was fixed at 430 mm.
Pulmonary scintigraphy.—Perfusion scintigraphy of the lungs was performed less than 24 hours after referral and after intravenous administration of 150 MBq of technetium 99m (99mTc)–labeled macroaggregated albumin (Maascint; Bristol-Meyers Squibb Pharma, Brussels, Belgium) with the patient in the supine position on the camera bed. Immediately after the tracer was injected, images were acquired in six standard projections (anterior, posterior, right and left anterior oblique, and right and left posterior oblique) by rotating the head of the camera (400AC; GE Medical Systems, Milwaukee, Wis) around the thorax of the patient. The camera was equipped with a low-energy high-resolution collimator. A window of 20% centered on the energy peak of 99mTc (140 keV) and a matrix of 128 x 128 pixels was selected for the accumulation of at least 400,000 counts per view.
Ventilation scintigraphy with krypton 81m (81mKr) (Mallinckrodt, Petten, the Netherlands) was performed immediately after each perfusion study. The radioactive gas was administered at a rate of 1–3 L/min through a face mask (Bennet Medical Equipment, Los Angeles, Calif). Parameters were the same as those used for the perfusion study, except the 20% window was shifted and centered on the energy peak of 81mKr (190 keV). If ventilation scintigraphy with 81mKr could not be performed on the same day as V-P scintigraphy, it was performed either on the following day or within 24 hours.
Pulmonary digital subtraction angiography.—When indicated, pulmonary angiography (n = 12) was performed by one of two experienced vascular radiologists (P.G. or F.H.) with a digital subtraction angiography unit (Integris V3000; Philips Medical Systems, Best, the Netherlands). All procedures were performed with a femoral approach. Sequential catheterization of the main, right, and left pulmonary arteries was routinely performed with a 6-F pulmonary pigtail catheter (Grollman; Cook, Bjaeverskov, Denmark). Undiluted nonionic contrast material (Ultravist; Schering, Berlin, Germany) was injected into the main artery at a rate of 20 mL/sec and into selective arteries at a rate of 10 mL/sec, with a maximum total volume of 30 and 18 mL per injection, respectively. Images were obtained during suspended respiration at a rate of two to three images per second until 10–15 images were obtained. Five additional images were subsequently obtained at a rate of one per second. Anteroposterior and ipsilateral anterior projections were systematically obtained after each selective catheterization. Additional lateral or posteroanterior projections were obtained whenever they were deemed necessary. Examinations after selective and superselective catheterization in either the lobar or segmental branches were performed when the initial findings were insufficient for diagnosis. Digital subtraction angiography was performed within 24 hours of V-P and multi–detector row spiral CT examinations. The operating physician printed each image from the monitor immediately after completion of digital subtraction angiography.
Chest radiography.—Every patient underwent posteroanterior and profile chest radiography within 24 hours. When a patient was unable to stand, bedside radiography of the chest was performed.
Image Analysis
All conventional chest radiographs, CT images, V-P scans, and digital subtraction angiograms were initially assessed by the on-call radiologist, nuclear physician, or angiographer, in a blinded fashion (reading 1). The images were later reviewed by one or two radiologists who worked independently (reading 2), were blinded to the imaging results of other specialties, and had experience in either conventional chest radiography (E.C., 8 years), chest CT (E.C., 8 years; L.G., 15 years), V-P scintigraphy (A.K., 20 years), or digital subtraction angiography (P.G., F.H.; both with 10 years). Any discrepancies concerning the presence or absence of PE were resolved by consensus reading (reading 3) of the CT scans (n = 4) and angiograms (n = 2). No consensus reading was available for images obtained with V-P scintigraphy. All three separate previously mentioned interpretations were recorded on separate spreadsheets and used for future reference. The initial CT and V-P readings were used to assess the need for pulmonary angiography. All reports were centralized by one coordinator (F.V., 5 years experience) who was responsible for patient care.
Multi–detector row spiral CT.—Image reading was preferentially performed at a workstation (O2; Silicon Graphics, Mountain View, Calif) at mediastinal (window width, 360 HU; window center, 60 HU) and lung (window width, 1,600 HU; window center, -600 HU) window settings. The radiologists could either vary the window and level settings or obtain multiplanar reformatted images in case of diagnostic difficulties. Images were graded for the quality of opacification at the mediastinal window setting. Image quality was graded as excellent when a high degree of opacification appeared in pulmonary arteries down to the fourth order, satisfactory when the images were sufficient for analysis of pulmonary arteries down to the fourth order, and poor when contrast enhancement did not allow confident analysis of pulmonary arteries down to the fourth order. The presence of PE was noted, as was the presence of any other abnormality in the mediastinum, chest wall, or lung parenchyma. PE were identified when either complete or partial filling defects within the main, lobar, segmental, or subsegmental arteries were identified (15). Visualized PE were localized to specific main, lobar, segmental, or subsegmental artery branches. To identify subsegmental arteries, we used the nomenclature proposed by Remy-Jardin et al (16) and later by Schoepf et al (17). Isolated subsegmental PE were specifically recorded. Alternative diagnoses suggested at CT were systematically mentioned.
V-P scintigraphy.—Interpretations of V-P lung scintigrams were fit into preestablished categories inspired by previous models (18,19). A diagnosis of PE was excluded when no perfusion defects of any kind were seen on scintigraphic images in at least the six classical views mentioned previously. A diagnosis of PE was unlikely but not fully excluded (low probability) when perfusion defects of any size were matched by equal or larger ventilation defects and were smaller or equal in size and shape to abnormalities that were eventually depicted on chest radiographs. A diagnosis of PE was made (high probability) when single or multiple large, wedge-shaped perfusion defects, often bordered by areas of overperfusion, coexisted with a normal distribution of ventilation. A diagnosis of PE was likely but not certain (intermediate probability) when perfusion defects did not fulfill the criteria for the other categories. In this instance, chest radiographs were either normal or abnormal. Radiographs were abnormal when the extension of perfusion defects on scintigraphic images was larger than the corresponding radiographic abnormalities.
Pulmonary angiography.—Pulmonary angiograms were reviewed at a computer workstation (Integris V3000). The criterion for diagnosis of acute PE was a partially occlusive filling defect within an arterial branch or a completely occlusive filling defect indicated by a meniscus of contrast material outlining the trailing edge of the PE (20). Any abnormalities that were suspicious for chronic PE were also recorded.
Chest radiography.—Chest radiographic description included the evaluation and tabulation of abnormalities that could interfere with interpretation of V-P scintigrams, such as parenchymal consolidation, atelectasis, pleural effusion, and emphysema.
Patient Follow-up and Outcome
Baseline creatinine levels were measured in all patients before the spiral CT examination, and the creatinine level was monitored in hospitalized patients. Patients with no findings of PE on imaging studies were observed for 3–6 months. Any episodes of recurrent or new deep venous thrombosis or PE were recorded. Patients were interviewed via telephone. If no response was elicited, the patient’s charts in the outpatient clinic were reviewed. Finally, information was not available for a few patients, so the family physician was interviewed. For patients who died during follow-up, the chart was reviewed and the attending physician was interviewed to document the presence of concurrent venous thrombosis or PE. The emergency physician in charge of the study performed all these procedures. Autopsy evidence was also recorded.
Final Diagnosis of PE
The presence of PE was confirmed with concordance of positive results of a spiral CT scan and a high-probability V-P scintigram. In this situation, no further imaging was performed, and the patient was treated with anticoagulant therapy and did not undergo pulmonary angiography (Fig 1).
|
|
|
|
|
2 test, respectively. A P value of less than .05 was considered to indicate a statistically significant difference. Sensitivity and specificity values (including 95% CIs) were calculated for multi–detector row spiral CT and V-P scintigraphy. Presence of PE (sensitivity) was diagnosed either when the CT scan was positive and concordant with a high-probability V-P scintigram or when a pulmonary angiogram was positive and CT scans and V-P scintigrams were discordant. Absence of PE (specificity) was diagnosed either when the CT scan was negative and concordant with a normal or low-probability V-P scintigram or when a pulmonary angiogram was negative and the results were discordant. Absence of PE was confirmed with negative clinical follow-up at 6 months. Sensitivity of scintigraphy was calculated by considering high-probability tests as positive for PE and all other tests as negative. Specificity of scintigraphy was calculated by considering normal or low-probability images as negative for PE and all other images as positive. Conclusive results were defined as a positive or negative interpretation at CT and either a high-probability interpretation or a normal or low-probability interpretation at V-P. Inconclusive results were defined as an indeterminate interpretation at CT and an intermediate interpretation at V-P.
Differences in parameters were assessed with the McNemar test, with a level of significance set at .05. The 
coefficients for observer agreement of the imaging studies were calculated and interpreted as described by Altman (21). A coefficient less than 0.20 was interpreted as poor, a coefficient between 0.21 and 0.40 as fair, a coefficient between 0.41 and 0.60 as moderate, a coefficient between 0.61 and 0.80 as good, and a coefficient between 0.80 and 1.00 as very good.
coefficients were calculated for agreement between the two readers. Interobserver agreement was calculated for reading 2 of CT scans and angiograms and for readings 1 and 3 of CT scans, V-P scintigrams, and angiograms.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Image Analysis
Multi–detector row spiral CT.—Multi–detector row spiral CT provided adequate enhancement of pulmonary arteries in 93 of 94 patients. Multi–detector row spiral CT scans were judged to be of excellent enhancement in 75 of 94 patients and of satisfactory enhancement in 18. The technical quality of one CT examination was considered to be unsatisfactory because of poor opacification of the arteries of the lower lobe secondary to pulmonary fibrosis distorting pulmonary vessels. PE was diagnosed in 28 of 94 patients (Table 1). In 15 of the 28 patients, the emboli involved central and peripheral pulmonary arteries. In the remaining 13 patients, only segmental or subsegmental branches were obstructed. Isolated subsegmental PE were found in four (14%) of 28 patients with PE (Table 2). As clearly shown in Table 2, these four patients had varied clinical symptoms. Spiral CT scans were interpreted as positive for PE at the first reading in 27 of 28 patients with PE (sensitivity, 96%; CI: 82%, 99%) (Table 1). There was one false-negative CT study. In one patient, acute PE was located at the level of division of the right lower lobe artery into basal segmental arteries and was first revealed only with pulmonary angiography (Fig 3). This abnormality was retrospectively identified by two chest radiologists at the second reading of spiral CT scans. All four subsegmental PE were diagnosed at the first reading, and the sensitivity was 100% (Table 2, Fig 2).
|
|
|
|
|
V-P scintigraphy.—High-probability V-P scintigrams of the lungs were obtained in 24 of 28 patients with PE (sensitivity, 86%; CI: 67%, 96%) (Table 1). The PE that were missed with scintigraphy in four patients were located in lobar and segmental arteries in two patients (Fig 3), in segmental arteries in one, and in one subsegmental arterey in one (Fig 2). Normal or low-probability V-P scintigrams were obtained in 58 of 66 subjects without PE (specificity, 88%; CI: 77%, 94%) (Table 3). Three (42%) of seven patients with intermediate-probability V-P scintigrams had PE, which is higher than the 32.6% prevalence found in the prospective investigation of pulmonary embolism diagnosis study (22). The other four patients had no evidence of acute PE. Among the seven patients with intermediate-probability V-P scintigrams, six had normal chest radiographs and one had emphysema.
Comparison of multi–detector row CT and V-P scintigraphy.—By comparing the findings of V-P scintigraphy and multi–detector row CT (Table 4), concordant positive results were found in 24 of 28 patients and concordant negative results were found in 58 of 66. Interpretations of multi–detector row CT scans and V-P scintigrams were discordant in five patients, indeterminate in one, and intermediate in seven (Table 4). The number of spiral CT examinations with conclusive results was significantly higher than the number of planar V-P scintigraphic examinations with conclusive results (P < .05).
|
|
coefficient was 0.94 for two readings of multi–detector row CT scans, 0.94 for V-P scintigrams, and 0.66 for angiograms.
Complications and Creatinine Levels
There were no serious complications associated with the iodinated contrast material–enhanced multi–detector row CT, V-P scintigraphic, or pulmonary angiographic examinations performed during the study. Two successive measurements of serum creatinine levels were available for 69 patients who were hospitalized after their initial diagnostic work-up. These patients had a baseline serum creatinine level of 1 ± 0.2 mg/dL (88.4 ± 17.7 µmol/L). Only one patient, a 57–year-old man, had an increase in serum creatinine level, from 1.5 to 2.2 mg/dL (132.6 to 194.5 µmol/L). This occurred 3 days after the spiral CT examination, and his serum creatinine level returned to normal 6 days later. Even patients who underwent both spiral CT and pulmonary angiography did not have a substantial increase in serum creatinine levels.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Our results show higher values of sensitivity and specificity with thin-collimation multi–detector row CT compared with the results of other studies. In most studies, however, the diagnostic performance of single–detector row spiral CT was evaluated in an unselected population that included both inpatients and outpatients (1–9). The presence of comorbid conditions may indeed decrease the accuracy of the test and is a factor that was not present in our patients. This may, in part, explain the better results obtained in our study.
A large prospective study (8) was undertaken to evaluate the diagnostic performance of single–detector row spiral CT in 299 outpatients. The authors found 4% of spiral CT scans to be inconclusive. Among patients with conclusive results, the sensitivity of single–detector row spiral CT was 70% and the specificity was 91%.
The overall prevalence of PE in our study population was 30% and was similar to the prevalence of PE in previous studies (22). Four (14%) of 28 patients had isolated subsegmental PE, a prevalence that is within the reported range of prevalence (6%–34%) in patients with acute PE (23–25). These patients had a wide range of symptoms, which clearly demonstrates the need to localize tiny emboli and thereby obviate further investigations. In our experience, thin-collimation multi–detector row spiral CT angiography is accurate even in subsegmental vessels. This modality allowed us to identify the four patients with isolated subsegmental PE at the time of the first reading. A possible false-positive reading of subsegmental PE occurred at the time of the second reading, when both chest-certified radiologists found an area of low attenuation located in a subsegmental branch of the superior lingular artery. This abnormality was not identified with a prospective and blinded reading of pulmonary angiograms. This patient did not receive anticoagulant therapy or develop any thromboembolic event at 6-month follow-up.
It is known that interobserver variability with pulmonary angiography, which is the standard diagnostic modality for PE, is much higher for these small emboli than it is for larger emboli (26,27). We found excellent agreement for the diagnosis of acute PE with thin-collimation multi–detector row CT ( = 0.94) and pulmonary angiography ( = 0.83) on a patient-by-patient basis. In the diagnosis of acute PE, our results show that the agreement between two readings of CT scans is comparable to the agreement between two readings of lung scintigrams ( = 0.94) but superior to the agreement between two readings of pulmonary angiograms ( = 0.66).
Confident analysis of pulmonary arteries down to the subsegmental level with thin-collimation multi–detector row CT may be explained by several factors. First, the use of thin collimation (effective thickness, 1.3 mm) and overlapping reconstruction (interval of reconstruction, 0.6 mm) provides greater accuracy for analysis of pulmonary arteries down to the subsegmental level. Ghaye et al (12) reported that peripheral pulmonary arteries down to the fifth-order branches can be accurately depicted on hard-copies with reconstructed 1.25-mm sections by using multi–detector row spiral CT. Schoepf et al (10) also demonstrated that the use of thin-collimation multi–detector row spiral CT has improved detection of subsegmental PE. Second, workstation review provides a flexible environment for varying window and level settings and for cine image presentation. Third, the routine use of time-attenuation curves to determine the scanning delay helped to obtain optimal opacification of all pulmonary vessels down to the subsegmental level.
We obtained adequate opacification of central and peripheral pulmonary arteries in 99% of the patients. The only patient in whom opacification was inadequate had pulmonary fibrosis. V-P scintigrams were intermediate and chest radiographs were abnormal in this patient. We did not systematically record the body mass index and the breathing status of patients in the study population, but we presume that interpretation of thin-collimation contrast-enhanced spiral CT images should be more difficult in patients who are obese or have severe dyspnea. The spiral CT protocol should probably be adapted in this clinical setting with use of thicker sections and an increased tube current and pitch.
In 19 (29%) of 66 patients who did not have PE, thin-collimation multi–detector row spiral CT provided diagnostic information that was either suggestive of an alternative diagnosis or consistent with the final clinical diagnosis. This additional diagnostic information was not provided by V-P scintigraphy or pulmonary angiography. This represents a serious advantage of spiral CT in these patients (6,7). As reported by Schoepf et al (28), the use of thin-collimation multi–detector row spiral CT would increase both sensitivity and specificity in the diagnosis of pulmonary lesions. Moreover, Honda et al (29) reported similar image quality for scans obtained with multi–detector row CT (4 x 1-mm) compared with scans obtained with single–detector row thin-section CT.
Additional postprocessing techniques will offer radiologists the possibility to reconstruct renderings of the entire chest with multiplanar or three-dimensional techniques. Multiplanar reformations should avoid pitfalls in the detection of PE in central and peripheral pulmonary arteries (30). Further studies are warranted, however, to establish the radiation dose induced by thin-collimation multi–detector row CT and the cost-effectiveness of this new method compared with other imaging methods currently used in the diagnostic work-up of PE.
This study has several limitations. First, the study did not include all consecutive patients in whom PE was suspected. This was partially dictated by the study design, which included patients in whom spiral CT and V-P scintigraphy could be performed within 24 hours. Difficult access to the imaging modalities during the weekend, lack of time for some patients to be included during rush hours, and opportunity to delay the investigations during hospitalization were all reasons to explain the large number of patients who were not included in this study. Also excluded were patients with raised D-dimer levels who were not taken into consideration because of an obvious alternative diagnosis to PE (n = 151) or contraindications to spiral CT (n = 40). On the other hand, we encountered neither technical problems nor contraindications to V-P scintigraphy. Our demographic data analysis did not reveal any statistically significant age difference between subjects included in and those excluded from the study; however, the proportion of women was greater in the group of patients included in the study (66 [70%] of 94 patients) than in the group of patients excluded from the study (434 [58%] of 747 patients). It is well known that oral contraceptives play a role in the pathogenesis of coagulation disorders, thereby raising the clinical probability of PE. This may explain the sex ratio difference found in our study group.
Second, we used a discrepant analysis of two tests to evaluate the performance of multi–detector row CT and V-P scintigraphy, with use of pulmonary angiography as the standard in case of discrepant results. This design is more ethical than a study that compares the results of multi–detector row CT and scintigraphy with the results of pulmonary angiography in every patient. Furthermore, a study that compared 1-mm collimation single–detector row spiral CT and pulmonary angiography in pigs with methacrylate casts as an independent standard demonstrated that angiography cannot be accepted as the standard of reference in all instances (31). By using this imperfect standard, interpretation will always be reported as the errors of the techniques being compared with this method of reference. The design we used in the present study had already been used in another study comparing the use of single–detector row CT and V-P scintigraphy in the detection of PE (4); however, this method of evaluation could have introduced a bias concerning the measurement of sensitivity and specificity with overestimation of these results for both techniques (32). Concerning sensitivity, the probability that both spiral CT and V-P scintigraphy induced a false-positive diagnosis of PE in the same patient does exist; however, the chance of making a wrong diagnosis by using both diagnostic tests together is low. Furthermore, it has not been proved that the error in our experimental results would be greater than the error of interpreting pulmonary angiograms alone. Specificity should not be affected in the present study because all patients with negative results were checked at 6-month follow-up, which represents a guarantee that a false-negative diagnosis of PE did not occur.
Third, our diagnostic examinations were performed in emergency outpatients; therefore, the results may not apply to hospitalized patients. Thus, we evaluated spiral CT scans and V-P scintigrams under better conditions than if we had drawn our subjects from an unselected population of hospitalized patients. Previous studies have reported the difficulty of using single–detector row spiral CT in the detection of acute PE in patients in the intensive care unit (9) or in patients with underlying pulmonary diseases. Other studies are needed to evaluate the performance of multi–detector row CT in unselected patients.
In summary, our results indicate thin-collimation multi–detector row CT is a powerful imaging technique for use in the detection of PE down to the subsegmental level in the evaluation of outpatients. Compared with V-P scintigraphy, multi–detector row CT has greater diagnostic accuracy and significantly higher rates of conclusive results.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, E.C., F.V.; study concepts and design, E.C., F.V., P.H.; literature research, E.C.; clinical studies, E.C., A.K., F.V.; data acquisition, E.C., A.K., F.V.; data analysis/interpretation, E.C., F.V.; statistical analysis, E.C., F.Z.; manuscript preparation, definition of intellectual content, and editing, E.C.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. M. Heyer, P. S. Mohr, S. P. Lemburg, S. A. Peters, and V. Nicolas Image Quality and Radiation Exposure at Pulmonary CT Angiography with 100- or 120-kVp Protocol: Prospective Randomized Study Radiology, November 1, 2007; 245(2): 577 - 583. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Remy-Jardin, M. Pistolesi, L. R. Goodman, W. B. Gefter, A. Gottschalk, J. R. Mayo, and H. D. Sostman Management of Suspected Acute Pulmonary Embolism in the Era of CT Angiography: A Statement from the Fleischner Society Radiology, November 1, 2007; 245(2): 315 - 329. [Full Text] [PDF]
|
|
|
|
 |
|
 N. Tunariu, S. J.R. Gibbs, Z. Win, W. Gin-Sing, A. Graham, P. Gishen, and A. AL-Nahhas Ventilation-Perfusion Scintigraphy Is More Sensitive than Multidetector CTPA in Detecting Chronic Thromboembolic Pulmonary Disease as a Treatable Cause of Pulmonary Hypertension J. Nucl. Med., May 1, 2007; 48(5): 680 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Coche, S. Vynckier, and M. Octave-Prignot Pulmonary Embolism: Radiation Dose with Multi-Detector Row CT and Digital Angiography for Diagnosis Radiology, September 1, 2006; 240(3): 690 - 697. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Ghaye, A. Nchimi, C. T. Noukoua, and R. F. Dondelinger Does Multi-Detector Row CT Pulmonary Angiography Reduce the Incremental Value of Indirect CT Venography Compared with Single-Detector Row CT Pulmonary Angiography? Radiology, July 1, 2006; 240(1): 256 - 262. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Reinartz, H.-J. Kaiser, J. E. Wildberger, C. Gordji, B. Nowak, and U. Buell SPECT Imaging in the Diagnosis of Pulmonary Embolism: Automated Detection of Match and Mismatch Defects by Means of Image-Processing Techniques J. Nucl. Med., June 1, 2006; 47(6): 968 - 973. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. D. Stein, S. E. Fowler, L. R. Goodman, A. Gottschalk, C. A. Hales, R. D. Hull, K. V. Leeper Jr., J. Popovich Jr., D. A. Quinn, T. A. Sos, et al. Multidetector computed tomography for acute pulmonary embolism. N. Engl. J. Med., June 1, 2006; 354(22): 2317 - 2327. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. T. Johnson and E. K. Fishman IV Contrast Selection for MDCT: Current Thoughts and Practice Am. J. Roentgenol., February 1, 2006; 186(2): 406 - 415. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. E. Jones and C. Wittram The Indeterminate CT Pulmonary Angiogram: Imaging Characteristics and Patient Clinical Outcome Radiology, October 1, 2005; 237(1): 329 - 337. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
