| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Brigham and Women's Hospital, 75 Francis St., Boston,
MA 02115.
2 Department of Biostatistics, Children's Hospital Boston, Boston, MA.
Received May 14, 2007;
accepted after revision August 21, 2007.
Address correspondence to A. R. Hunsaker.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Eight hundred twenty-nine consecutive patients (281 men and 548 women) underwent CTA–CTV for pulmonary embolism. Reports were evaluated as follows: positive or negative for pulmonary embolism with or without deep venous thrombosis (DVT) or with nondiagnostic CTV. Coexisting factors of malignancy, previous venous thromboembolism (VTE), recent surgery, and cardiovascular disease comprised the high-risk group of 446 patients. The remaining 383 patients formed the low-risk group. Statistical analysis included four binary predictors (previous VTE, malignancy, cardiovascular disease, and surgery) and three binary outcome variables (pulmonary embolism, DVT, and VTE). Chi-square test and univariate and multivariate regression analyses were performed.
RESULTS. VTE, pulmonary embolism, and DVT occurred in 152 (18.3%), 124 (15.0%), and 61 (7.3%) of 829 patients, respectively. Between the high-risk and low-risk groups, prevalence of VTE was 114 (25.6%) of 446 and 38 (9.9%) of 383 patients, respectively (p < 0.001); prevalence of pulmonary embolism was 92 (20.6%) of 446 and 32 (8.3%) of 383 patients, respectively (p < 0.001). Isolated DVT was found in 28 (3.4%) of 829 patients. The incremental value of CTV for the entire cohort was 3.4%, 0.72% in the low-risk group (six of 829) and 2.6% (22 of 829) in the high-risk group. For outcome variable VTE, malignancy and previous VTE were statistically significant (p = 0.04 and p < 0.001, respectively); for pulmonary embolism, malignancy and previous VTE were statistically significant (p = 0.03 and p = 0.005, respectively); for DVT, only previous VTE was statistically significant (p < 0.001).
CONCLUSION. CTV should not be performed routinely in all patients evaluated for pulmonary embolism and may only be useful in patients with a high probability of pulmonary embolism, including those with a history of VTE and possible malignancy.
Keywords: CT pulmonary angiography • CT venography • deep venous thrombosis • pulmonary embolism • venous thromboembolism
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
However, indirect CTV may not provide incremental diagnostic value in all patients with suspected pulmonary embolism, and the additional cost and radiation associated with this procedure may therefore be unwarranted. The purpose of our study was two-fold. First, we sought to assess the utility of performing routine pelvic and lower extremity CTV in all patients being evaluated for pulmonary embolism. Second, if indeed there is usefulness for the procedure, we attempted to identify the groups who may benefit from the combination study.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study Population
This was a retrospective single-center cohort study. We reviewed the
finalized reports of consecutive patients who were suspected of having
pulmonary embolism and who were therefore evaluated with combined pulmonary
CTA and indirect CTV to assess for pulmonary embolism between January 1, 2005,
and April 30, 2005. For this retrospective study, we searched the hospital's
radiology information systems database to identify all patients presenting to
the emergency department, outpatient clinics, or inpatient wards who underwent
CTA in combination with indirect lower extremity and pelvic CTV
(CTA–CTV) for evaluation of VTE. A search was also made for patients who
also underwent lower extremity sonography, but in all instances, none was
performed. Eight hundred forty-seven patients were identified, and 18 were
eliminated because both the CTA and the CTV were interpreted as nondiagnostic,
resulting in a study population of 829 patients.
Of the 829 patients (age range, 21–100 years; mean age, 60 years), 281 were men and 548 were women. Of these, 505 (61%) were outpatients and 324 (34%) were inpatients. Using the hospital's computer database, we extracted the following history: malignancy (including bronchogenic, ovarian, endometrial, testicular, prostate, thyroid, and colon carcinomas and melanoma, lymphoma, and sarcoma); previous VTE; cardiovascular disease (including atherosclerotic and coronary artery disease such as stroke, peripheral vascular disease, and cardiac disease); and surgery within 4 weeks of the CTA–CTV. Surgeries included resection of malignancies, open reduction and internal fixation for fracture, joint replacement, bowel resection for ischemia, lung transplantation, pneumonectomy, liver resection, spine surgery, and cerebral aneurysm resection. Symptoms of pleuritic chest pain, tachypnea, tachycardia, shortness of breath, and hemoptysis were also recorded.
Imaging
All CTA–CTV studies were performed on either a 16- or 64-MDCT scanner
(Somatom Sensation 64 and Emotion, Siemens Medical Solutions). On both
scanners, the lungs were scanned from base to apex and images displayed from
apex to base. Scanning parameters for the 16-MDCT scanner included 0.75-mm
collimation helical acquisitions, 1-mm-thick reconstruction, 0.7-mm
reconstruction increment, 0.5-second rotation time, 100–140 mAs, and 120
kVp. The 64-MDCT scanning parameters included 0.6-mm collimation, 1-mm-thick
reconstruction, 0.5-mm reconstruction increment, 0.5-second rotation time,
100–140 mAs, and 120 kVp. All studies were followed with images of the
pelvis from a level just above the iliac crest down to the popliteal fossa 3
minutes after completion of the CTA. For all studies, 125 mL of 370 mg I/mL
iopromide nonionic contrast material (Ultravist 370, Bayer HealthCare) was
administered IV at an injection rate of 3 mL/s. For both the 16- and 64-MDCT
scanners, the automatic bolus-triggering software program was used, with a
region of interest placed over the main pulmonary artery, triggering the
scanner at an enhancement level of 200 H. No additional contrast material was
administered for the indirect CTV.
Images were reconstructed as follows: contiguous 5-mm-thick transverse slices in lung window settings (width, 2,000; level, –600), contiguous 1-mm-thick slices in mediastinal windows (width, 360; level, 40), and contiguous 3-mm-thick coronal slices in mediastinal windows. Images of the pelvis and lower extremities were displayed at 10-mm thickness at 30-mm intervals in soft-tissue windows (width, 350; level, 40). All images were reviewed and interpreted on PACS workstations (Impax, Agfa or Centricity, GE Healthcare).
Interpretation of CT Scans
Findings from all 829 examinations were interpreted by three
fellowship-trained chest radiologists with 12, 12, and 14 years of experience,
respectively, and by two fellowshiptrained emergency radiologists with at
least 8 years of experience each. The interpretations were retrospectively
collected by two of the authors who reviewed the reports.
The reports were categorized according to final interpretation as follows: positive for pulmonary embolism and for DVT, positive for pulmonary embolism and negative for DVT, positive for pulmonary embolism and technically inadequate CTV, negative for pulmonary embolism and positive for DVT, negative for pulmonary embolism and for DVT, and negative for pulmonary embolism and technically inadequate CTV. Studies were considered inadequate if contrast opacification of either the pulmonary arterial or deep venous system was insufficient for the confident exclusion of thromboembolism. The investigators did not reinterpret the scans. A positive CTA study included saddle, lobar, segmental, or greater than one subsegmental filling defect within the vessels. Positive CTV studies included only those studies interpreted as filling defects within one or more of the pelvic or lower extremity veins.
Data Analysis
The frequencies of pulmonary embolism with or without DVT or pulmonary
embolism with nondiagnostic CTV studies and the presence of DVT alone were
recorded for the entire cohort. Using previously described risk factors for
VTE
[14–16],
patients were classified into two groups: low-risk and high-risk. Those in the
high-risk group showed one or more of the following: history of malignancy,
previous VTE disease, surgery, or cardiovascular disease. Those in the
low-risk group were those who presented with symptoms of chest pain, dyspnea,
or tachypnea without other comorbidities.
Statistical Analysis
Four binary predictors of outcome were recorded: cancer, cardiovascular
disease, history of VTE, and surgery within 4 weeks before the CTA. We set
VTE, pulmonary embolism, and DVT as our three binary outcome variables. First,
summary statistics including range, median, mean ± SD for the
continuous variable, age, were computed. For all binary variables, the
analysis of independent proportions was conducted with two-tailed z
statistics. Second, against each of the three outcome variables, we tabulated
the predictor variable by constructing a contingency table. Using this
two-by-two table, a univariate chi-square test was conducted to test whether
the predictor and the outcome were associated or independent. Finally,
multivariate logistic regression was used to test for an association between
the four identified high-risk variables (malignancy, cardiovascular disease,
postoperative state, and previous VTE) and diagnoses of VTE, pulmonary
embolism, and DVT. The 95% CI for means and proportions was calculated, and
p < 0.05 was considered significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Analysis by Group
The high-risk group (Table
2) had a significantly higher rate of VTE overall, accounting for
75% (114 of 152) of the positive cases. Results were similar for pulmonary
embolism and DVT, with the high-risk group accounting for 74.2% (92 of 124)
and 80% (49 of 61) of patients who had pulmonary embolism and DVT,
respectively. Of those with DVT, 46% (28 of 61) occurred in patients with no
evidence of pulmonary embolism, with 78.6% (22 of 28) of these cases occurring
in the high-risk group.
|
Table 3 shows the prevalence of VTE in the high-risk group. Some overlap exists among the groups because some patients had more than one risk factor. The prevalence of VTE was highest in the group with previous VTE (42.6% [26 of 61]), although the prevalence of VTE was also high in the other subgroups. Prevalence of isolated DVT was also highest (14.7% [nine of 61]) in patients with previous VTE, and in this group, only 9.8% (six of 61) had DVT in the setting of pulmonary embolism. Of the 26 patients with previous VTE, four had a history of both pulmonary embolism and DVT and in our study, all four had pulmonary embolism (three without DVT and one with DVT), whereas none had isolated DVT. Eight of the 26 had previous pulmonary embolism (one had both pulmonary embolism and DVT, four had pulmonary embolism without DVT, and three had DVT alone). The largest subgroup in the 26 was that with previous DVT (14 patients), eight of whom had pulmonary embolism in our study (four with and four without DVT); the remaining six had DVT alone. Isolated DVT was found in approximately 4% in each of the remaining three groups.
|
Regression Analysis
On the basis of the univariate chi-square analysis as a binary predictor of
the outcome variable VTE, malignancy and history of previous VTE were
statistically significant (p = 0.04 and p < 0.001,
respectively). Cardiovascular disease and postoperative period were not
statistically significant (p = 0.26 and p = 0.15,
respectively). For the binary outcome variable pulmonary embolism, malignancy
and history of VTE were statistically significant (p = 0.03 and
p = 0.005, respectively), whereas cardiovascular disease and
postoperative period were not statistically significant (p = 0.30 and
p = 0.16, respectively). For the outcome variable DVT, only history
of VTE was significant (p < 0.001).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We subdivided our patient population into high-risk and low-risk groups in an attempt to answer our initial question of the necessity for routine indirect CTV in all patients with suspected pulmonary embolism. Our highrisk group included patients with malignancy, history of VTE, surgery within 4 weeks before the CTA–CTV study, and patients with cardiovascular disease, groups that have been previously defined in the literature [14–16]. Our low-risk group included patients who presented with chest pain, tachypnea, shortness of breath, and undocumented tachycardia or decreased oxygen saturation who did not have any of the features of the high-risk group. We found that of the 152 instances of VTE, 114 (75%) occurred in the high-risk group and 25% (38 of 152) were seen in the low-risk group. Pulmonary embolism was also much higher in the high-risk group (92 [74.2%] of 124 cases of pulmonary embolism), with 25.8% (32 of 124) occurring in the low-risk group. In addition, 49 (80.3%) of the 61 cases of DVT were seen in the high-risk group. Further subdividing the 49 cases of DVT in the high-risk group, 22 (44.9%) had isolated DVT, whereas 55% (27 of 49) cases were associated with pulmonary embolism. There were six cases of isolated DVT in the low-risk group as well as six cases of DVT in association with pulmonary embolism.
High incremental value of CTV appears to be the driving force for its popularity because it is usually calculated on the basis of the number of cases of VTE in the study rather than on the entire study population—those with and without VTE. On review of our data in these categories, the incremental value of indirect CTV appears to confirm previously published reports in the literature [10–13], which show incremental benefit of CTV as high as 38%. Indeed, we found incremental values of 15.4–17.4% in patients with cardiovascular disease (15.4%), recent surgery (16.7%), and malignancy (17.4%). Ghaye et al. [10] reported a 14.4% and 27.4% incremental value of CTV when using single-detector and MDCT scanners, respectively. These values were calculated according to the number of patients who had VTE, leading to a conclusion of increased diagnosis of VTE in 27% of patients. More recently, PIOPED II results were published [7]. In this large multicenter clinical trial (1,090 patients enrolled and 824 completed study), the false-negative rate for CTA alone was 17%, with the conclusion that CTA–CTV has a higher sensitivity for the diagnosis of pulmonary embolism than CTA alone.
In the PIOPED II study, among 632 patients with normal findings on composite reference standards, the absolute difference in positivity between the CTA alone and CTA–CTV was 0.75%. However, if one were to compare the differences between CTA alone and CTA–CTV among only those who were positive for VTE, one would find a 16.7% false-negative rate (25 positive with CTA alone and 30 positive with CTA–CTV). In the accompanying editorial to the PIOPED II article, Perrier and Bounameaux [17] made a similar point, noting that pulmonary embolism was found in 192 of 824 patients who received a reference diagnosis. This amounted to an absolute gain of only 2% in the negative predictive value by adding CTV (97% for CTA–CTV and 95% for CTA alone).
In a prospective study, Perrier et al. [18] showed that the improvement in overall detection of pulmonary embolism by lower extremity venous sonography was marginal (0.9%; 95% CI, 0.3–2.7). Thus they showed that in patients without a high clinical probability of pulmonary embolism, this diagnosis can be safely ruled out without lower extremity venous sonography. Similarly, the Christopher Study [19] showed that chest CT and D-dimer testing are sufficient in patients with a low clinical probability.
Our study showed a 15% prevalence of pulmonary embolism with or without DVT and a 3.4% prevalence of isolated DVT in both the high- and low-risk groups combined. However, the prevalence of isolated DVT in the low-risk group was 0.72% whereas that in the high-risk group was 2.6% when the entire population was taken into account. Therefore, the incremental value of CTV for the entire cohort was 3.4% but was only 0.72% in the low-risk group and 2.6% in the high-risk group when the whole population was taken into account. In the high-risk patients, the incremental value of CTV was 2.5% (11 of 446), 0.45% (two of 446), 2.0% (nine of 446), and 1.1% (five of 446) in the subgroups with malignancy, cardiovascular disease, previous VTE, and surgery, respectively. Only a history of VTE predicted the outcome variable DVT. In calculating the incremental value of CTV on the basis of the entire study population rather than on only those with VTE, we gain a better idea of the prevalence of the disease, isolated DVT, in all patients being evaluated for pulmonary embolism and can better determine the usefulness of CTV.
On the basis of our study, which may reflect the types of referrals of patients with suspected pulmonary embolism, CTV may be of benefit in patients who have a high likelihood of pulmonary embolism, particularly those with previous VTE and possibly those with malignancy, but in others the value appears less significant. The small incremental value of 1–2% with CTV for all patients referred for CT may not be worth either the financial cost or the added radiation dose. In a study to assess radiation exposure, Rademaker et al. [20] calculated organ, effective, and gonadal doses in six patients who underwent CTA–CTV using a single-detector helical CT scanner and found that the effective dose for combined CTA–CTV was 4.75 mSv, 50% of this being from the CTV. More specifically, they found that the radiation dose for the ovaries increased by a factor of 500 and by 2,000 in the testes, higher than a dual-phase helical CT of the liver. In another study [21] comparing CTV with Doppler sonography, radiation doses were also calculated for patients receiving CTA–CTV on a 4-MDCT scanner. A median cumulative effective dose of 2.81 mSv for women with CTA alone with an increase to 8.31 mSv with CTA–CTV using a 4 x 2.5 mm collimation with a table speed of 12.5 mm was reported. The effective gonadal dose was 3.87 mSv. For men, the results were similar, with the median cumulative effective dose at 2.06 mSv for CTA alone, rising to 8.24 mSv (gonadal dose, 3.91 mSv).
Limitations of our study include its retrospective nature. Selection bias, however, was not a contributing factor because all consecutive patients over the study period were used. Second, we did not use a reference standard for validation of negative or positive CTA–CTV results. Because of this, false-positive and false-negative cases may have been included. Sensitivity and specificity could not be calculated because of the lack of a reference standard [22]. But, as in other similar retrospective studies, the incremental value is an important tool. Third, in patients whose CTV was interpreted as technically inadequate, Doppler sonography of the lower extremities was recommended in all cases but not performed. We think this may have reflected the degree of clinical suspicion on the part of the requesting physician. Fourth, our CTV study was performed using a 3-cm gap between slices; however, this is the method used by several reports in the literature [3, 4, 13, 23] in which interslice gaps of 2–5 cm are routinely used. Lastly, data regarding the predictive value of the presence or absence of symptoms and signs of DVT in this population are lacking; hence, this area deserves further study.
In summary, the results of this study seem to suggest that routine CTV in all patients with suspected pulmonary embolism may, in fact, be unnecessary. Prospective studies may need to be performed to further assess whether indirect CTV could have more costs (both radiation and financial) than benefits in unselected patients. We recommend performing CTV in those patients with a high probability of pulmonary embolism, namely patients with previous VTE and possibly those with malignancy. Of course, all patients presenting with signs of DVT, such as leg swelling, should also undergo CTV.
Acknowledgments
We thank Laura J. Freeman for her help in preparation of this
manuscript.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Nazaroglu, C. A. Ozmen, H. O. Akay, I. Kilinc, and A. Bilici 64-MDCT Pulmonary Angiography and CT Venography in the Diagnosis of Thromboembolic Disease Am. J. Roentgenol., March 1, 2009; 192(3): 654 - 661. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. R. Goodman, H. D. Sostman, P. D. Stein, and P. K. Woodard CT Venography: A Necessary Adjunct to CT Pulmonary Angiography or a Waste of Time, Money, and Radiation? Radiology, February 1, 2009; 250(2): 327 - 330. [Full Text] [PDF]
|
|
|
|
 |
|
 S. Konstantinides Acute Pulmonary Embolism N. Engl. J. Med., December 25, 2008; 359(26): 2804 - 2813. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
