| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ultrasonography |
1 From the Dept of Radiology, Ochsner Foundation Hosp, 1514 Jefferson Hwy, New Orleans, LA 70121-2484 (E.I.B., L.A.T., L.A.K., M.A.S., D.H.S.); Research Dept, American College of Radiology, Reston, Va (J.H.S., P.E.C., J.H.B.); Dept of Radiology, The Western Pennsylvania Hosp, Pittsburgh (J.B.L., H.L.N.); Dept of Family and Community Medicine, Medical College of Wisconsin, Milwaukee (C.A.B.); and Dept of Medicine, Palo Alto VA Health Care System, Stanford Univ, Calif (P.A.H.). Received Jun 7, 1999; revision requested Aug 9; revision received Oct 11; accepted Nov 10. Supported in part by the American College of Radiology and in-kind contributions from Ochsner Foundation Hosp, The Western Pennsylvania Hosp, and the American College of Radiology. Address correspondence to E.I.B.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: In the principal pilot study, a prospective, blinded comparison of power Doppler imaging with duplex Doppler imaging, the reference-standard method, was conducted in 100 consecutive patients routinely referred for carotid artery imaging at a large, private multispecialty clinic. In the validation pilot study, a prospective, blinded comparison of power Doppler imaging with digital subtraction angiography, the reference-standard method, was conducted in 20 consecutive patients routinely referred at a teaching hospital. Using conservative assumptions, the authors performed cost-effectiveness analysis.
RESULTS: Power Doppler imaging produced diagnostic-quality images in 89% of patients. When the images of the patients with nondiagnostic examinations were regarded as positive, power Doppler imaging had an area under the receiver operating characteristic curve, Az, of 0.87, sensitivity of 70%, and specificity of 91%. The validation study results were very similar. The cost-effectiveness of screening and, as indicated, duplex Doppler imaging as the definitive diagnostic examination and endarterectomy was $47,000 per quality-adjusted life-year.
CONCLUSION: The Az value for power Doppler imaging compares well with that for mammography, a generally accepted screening examination, and with most other imaging examinations. Power Doppler imaging is likely to be a reasonably accurate and cost-effective screening examination for carotid artery stenosis in asymptomatic populations.
Index terms: Carotid arteries, angiography, 172.1248, 908.122 • Carotid arteries, stenosis or obstruction, 172.721, 908.721 • Carotid arteries, US, 172.12983, 172.12984, 172.12989 • Ultrasound (US), Doppler studies, 172.12983, 172.12984, 172.12989 • Ultrasound (US), power Doppler studies, 172.12989
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
|
The purpose of our research was to ascertain whether power Doppler imaging appears to be sufficiently promising in diagnostic accuracy and cost-effectiveness to merit a much larger and therefore much more expensive multicenter trial in a purely screening population to definitively evaluate its value in screening.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Power Doppler imaging was used only to demonstrate flow in the vessel and to assess the degree of stenosis; the characteristics and morphology of the plaque causing the stenosis were not evaluated or classified, and Doppler spectral analysis was not performed. The degree of stenosis was assessed by comparing the narrowed area of the internal carotid artery with a more normal distal portion of the depicted internal carotid artery by using North American Symptomatic Carotid Endarterectomy Trial methodology.
The criteria for a technically adequate examination were (a) both the internal carotid artery and the common carotid artery were well depicted and (b) the lumen was assessed appropriately (25,26). When calcification with shadowing was present, the technologist was instructed to try scanning other areas of the neck to obtain an image without shadowing. When, despite these attempts, shadowing obscured adequate visualization of the lumen, the examination was reported as nondiagnostic. All ultrasonographic (US) examinations were performed at facilities certified by the Intersocietal Commission for the Accreditation of Vascular Laboratories in performing extracranial carotid artery studies.
Principal Pilot Study
Under the direction of one of the authors (E.I.B.), 100 patients were examined between April and September 1997 with power Doppler imaging screening. Sagittal and transverse images of the right and left common carotid and internal carotid arteries were obtained (Figs 1, 2). The power Doppler imaging study took place either before or after a complete duplex Doppler examination of the extracranial carotid arteries, which had been ordered routinely according to usual indications. Duplex Doppler imaging served as the reference-standard examination. Each of the two imaging studies was performed independently by a different technologist and was interpreted offline by a different radiologist, without knowledge of the results of the other study. Four radiologists (including E.I.B., M.A.S., and D.H.S.) each interpreted at least 15 power Doppler imaging studies. The total time the patient was on the examining table for the entire power Doppler imaging examination was recorded by the technologist. The degree of stenosis on the power Doppler image was estimated visually by the interpreting radiologist.
The study patients were consecutively recruited from patients referred to the US section of the Department of Radiology at the Ochsner Clinic for imaging of the carotid arteries on those days when the full technological staff was present.
The screening study results were compared with the complete duplex Doppler examination results (Figs 1, 2) and interpreted according to commonly accepted criteria (25,26). For duplex Doppler imaging, these criteria provide an estimate of the lower and upper limits of the degree of stenosis. The upper value for the degree of stenosis was used to place the reference-standard estimate of stenosis into the same five categories used for the power Doppler imaging evaluation as follows: stenosis of 0%–39%, 40%–59%, 60%–79%, and 80%–99%, and complete occlusion.
Data were double entered into an analytic file and verified. Diagnostic accuracy was assessed by using the ROC curve and Az; sensitivity, specificity, and predictive values; and likelihood ratios (27). These accuracy measures were calculated for each of two definitions of disease: stenosis of 60% or greater but no occlusion and stenosis of 40% or greater but no occlusion. The first definition is based on ACAS criteria. The second definition was studied because future clinical trials may yet again broaden the suggested indications for surgery and because short-interval reexamination may be desirable for patients who are near but still below the disease threshold at which active treatment is recommended.
Ninety-five percent CIs were computed by using "jackknifing" to estimate standard errors (28). Jackknifing accounts for the correlation of data arising from two examinations (left and right) performed in the same patient and thus yields accurate CIs. Some patients had nondiagnostic-quality power Doppler imaging examinations, and it is not clear whether the mechanism that leads to the identification of examinations as nondiagnostic meets the statistical assumptions of randomness that underlie the calculation of CIs. Therefore, CIs were calculated only in those patients with diagnostic-quality examinations. Az values were estimated nonparametrically (29).
Validation Pilot Study
Under the direction of another member of the study team (J.B.L.), 20 patients were examined between May and October 1997 with a power Doppler imaging screening examination after they had undergone selective left and right carotid digital subtraction angiography (DSA), which served as the reference-standard examination. DSA was performed in the interventional radiology section of the radiology department at The Western Pennsylvania Hospital, and the power Doppler imaging screening examination was performed in the US section of the department by a separate, blinded staff. The power Doppler imaging evaluation consisted of only longitudinal imaging of both extracranial systems, with measurements obtained from both the anterior and posterolateral windows. These patients were consecutively recruited from patients referred to the interventional radiology section for DSA of the carotid arteries on weekdays.
The degrees of stenosis depicted on the power Doppler and DSA images were determined by means of measurement with a ruler. For power Doppler imaging, the on-screen electronic ruler on the US scanner was used, the degree of stenosis was calculated in both views, and the higher degree of stenosis was then recorded by the radiologist.
Data were entered into an analytic file and verified. We defined the error of power Doppler imaging as the percentage of stenosis measured by using power Doppler imaging minus the percentage of stenosis measured by using DSA, and then we studied this error statistic. To compare these findings with those of the principal pilot study, we categorized the degree of stenosis as measured by power Doppler imaging and DSA into the same five categories and then calculated the same diagnostic accuracy statistics. Ninety-five percent CIs were again computed by using jackknifing. There were no nondiagnostic-quality power Doppler imaging examinations, but the statistical assumptions underlying the CIs for an Az necessitate that larger categories of stenosis be associated with disease. Therefore, for ROC curves only, completely occluded vessels were treated as "diseased" even though they would not be considered for surgery.
Cost-effectiveness Analysis
One of the authors (P.A.H.), who conducted the numeric analysis in the previously published cost-effectiveness model (11), reran the same Markov model in complete original detail, making only a required minimum of modifications. In brief, this Lee et al (11) model, which is fully described in the published article, uses empirical data from the ACAS on the risks of angiography (the definitive diagnostic examination used), carotid endarterectomy, and stroke of varying degrees of severity in operated on and non–operated on patients. Beyond the 5-year maximum postsurgical follow-up period of the ACAS, the model makes the conservative assumption that the risk of stroke in operated on patients gradually increases to the risk in non–operated on patients, so that the risks are equal at 10 years and beyond. The model assumes that screening is applied to a general population of 65-year-old men (not the populations in our pilot studies, which had a higher prevalence of disease). Future health effects and future costs are discounted at a 3% annual rate.
All the modifications that we made to the original model are detailed in Table 1. As Table 1 explains, we estimated the cost of power Doppler imaging by surveying informed radiologists about the resources consumed by this procedure relative to an established, relatively closely related procedure—duplex Doppler imaging. This is the standard method used to determine the relative value units and hence payment levels assigned to a new procedure. Because the same equipment is used in power Doppler imaging and duplex Doppler imaging, the comparison largely involved time. The analysis of time included set-up and clean-up times rather than merely the time to obtain images.
As is usual in cost-effectiveness analyses, we conducted a sensitivity analysis to analyze the effect of varying key assumptions of the model by using plausible alternatives (Table 2). Because our basic assumptions deliberately were generally conservative, the alternatives examined were usually more favorable.
|
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Considering only the 89 patients with diagnostic-quality power Doppler imaging examinations—or a total of 178 vessels—the Az value for the empirical ROC curve of power Doppler imaging was 0.90 (95% CI: 0.71, 1.00) for stenosis of 60% or greater (Fig 3) and 0.84 (95% CI: 0.66, 1.00) for stenosis of 40% or greater (Fig 4). In these patients, power Doppler imaging had a sensitivity of 56% and a specificity of 99% for the detection of stenosis 60% or greater (Table 3). The positive predictive value was 90%, and the negative predictive value was 96%. For the detection of stenosis 40% or greater, the sensitivity of power Doppler imaging was 76%; specificity, 88%; positive predictive value, 51%; and negative predictive value, 96%.
|
|
|
As ROC curves normally do (29), the ROC curves for power Doppler imaging in our study indicated that there were possibilities for gaining sensitivity at the cost of specificity, and vice versa. For example, for the detection of 60% or greater stenosis, one could refer patients for further work-up if their power Doppler imaging examinations were nondiagnostic or revealed stenosis of 40% or greater. This would result in 91% sensitivity and 79% specificity. Table 4 contains data from the principal pilot study—information from which the described combination and other possible combinations of sensitivity and specificity can be constructed.
|
The Az value for empirical ROC curve was 0.88 when disease was defined as stenosis of 60% or greater and 0.84 when disease was defined as stenosis of 40% or greater (Figs 3, 4). Power Doppler imaging had a sensitivity of 62% and a specificity of 100% for the detection of 60% or greater stenosis and a sensitivity of 75% and a specificity of 90% for the detection of 40% or greater stenosis (Table 3).
The error of power Doppler imaging, defined as the percentage of stenosis measured by using power Doppler imaging minus the true percentage of stenosis determined by using DSA, averaged –4 percentage points for the 40 vessels, with an SD of 21 percentage points. Thus, there was a slight tendency for power Doppler imaging to underestimate the true degree of stenosis. For two vessels, there was a large discrepancy between the power Doppler imaging reading and the DSA reading. Subsequent investigation disclosed that the location of stenosis that had been measured on the power Doppler image was not the location of maximum stenosis seen on the DSA image. When the correct area, as defined by using DSA, was remeasured on the power Doppler image, the agreement in readings was close. Considering only the other 38 vessels, the error averaged –1 percentage point, with an SD of 16 percentage points.
Cost-effectiveness Analysis
Our model indicated that in the base case, without an intervention consisting of screening and, if indicated, a definitive diagnosis and endarterectomy, the average member of the model cohort of 65-year-old men could look forward to slightly more than 11 quality-adjusted life-years (QALYs), with life-time expenditures for stroke-related health care of $5,671. (Table 5 presents all cost-effectiveness results; note that both future QALYs and future costs are discounted at 3% per year.)
|
A number of variants studied in the sensitivity analysis were found to be considerably more cost-effective for the intervention—in the range of approximately $30,000–$40,000 per QALY (Table 5). These variants included dismissing patients with a nondiagnostic power Doppler imaging examination rather than referring them for further work-up and limiting screening to a population with a higher prevalence of stenosis. Differences in the cost of power Doppler imaging had a much smaller effect on the cost-effectiveness of the intervention.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Accuracy.—The Az is probably the single best measure of the accuracy of a diagnostic examination because, as described, the ROC curve encompasses the possibilities of exchanging sensitivity for specificity, and vice versa. The Az value for power Doppler imaging of the carotid arteries in all the patients in our study, 0.87, is greater than that for most diagnostic imaging examinations in use and compares well to mammography, a universally accepted imaging screening examination. The Az value for screening mammography was reported to be 0.84 for a large, systematic sample of radiologists accredited in mammography by the American College of Radiology (31) and to be 0.83–0.88 in smaller, less systematic studies (32–34). The Az value for power Doppler imaging that we observed is also similar to that for duplex Doppler imaging, which has been reported to be 0.91 (14), and duplex Doppler imaging is gaining widespread use as a definitive presurgical diagnostic examination for carotid artery stenosis (14, 35–37).
Economy.—Power Doppler imaging is similar to mammography not only in diagnostic accuracy. As well, our cost estimate of $80 per power Doppler imaging examination (Table 1) is almost identical to the cost of screening mammography. (We estimate mammography's cost, as with all physician services, at 125% of the average Medicare payment rate; see Table 1.) Before our study, duplex Doppler imaging had been the examination principally discussed as the screening examination for carotid artery stenosis. Our estimate is that power Doppler imaging costs approximately one-third as much as duplex Doppler imaging.
Cost-effectiveness analysis provides a more sophisticated viewpoint than does a simple cost comparison, because it enables one to compare health benefits gained with their costs. Health care services that cost up to approximately $50,000 per QALY are generally regarded as clearly acceptable, with the $50,000–$100,000 per QALY cost range regarded as borderline (11,12,38). In our base case, which was designed to err, if anything, toward a high cost per QALY, we found cost-effectiveness to be at the high end of the "clearly acceptable" range when duplex Doppler imaging or MR imaging was the definitive diagnostic examination. Moreover, we identified multiple strategies for reducing the cost to approximately $30,000–$40,000 per QALY (Tables 2 and 5).
Overall, then, our research results indicate that power Doppler imaging seems to be a good screening examination for carotid artery stenosis; this proved to be true from the standpoints of diagnostic accuracy, cost, cost-effectiveness, and speed.
Screening strategy.—We envisage that power Doppler imaging screening, like other screening examinations such as mammography, will be used for the initial categorization of patients as being nondiseased versus requiring further diagnostic evaluation. The latter group of patients can then be asked back to undergo further studies, such as duplex Doppler imaging or MR imaging.
Although our research was not a definitive study, it provided a number of suggestions about screening strategy. For one, the results of our analysis show, unsurprisingly, that cost-effectiveness is strongly dependent on the prevalence of stenosis among the individuals screened. Thus, it may be desirable to focus screening on those individuals who have at least one risk factor in addition to that of age—for example, smoking, which triples the risk of serious stenosis (30). Second, power Doppler imaging has reasonably good accuracy in depicting stenosis of 40% or greater. Hence, if future clinical trials show a benefit from some treatment in this broader population, or if the best care for individuals with stenosis in the 40%–59% range involves frequent follow-up, power Doppler imaging should be useful for screening. Third, as the findings of a pilot study should, our research results suggest ways to improve results in the future—principally, by exerting more effort to avoid nondiagnostic examinations and taking more care to identify the most stenotic region on images. Our cost-effectiveness analysis results show that these strategies should be valuable, even if the cost of screening increases moderately. We expect that with increasing experience, the technique can be made more effective.
Fourth, obtaining maximal cost-effectiveness—that is, achieving the lowest cost per QALY—necessitates following a screening strategy that, from a conventional clinical perspective, seems excessive in its single-minded pursuit of specificity and its consequent sacrifice of sensitivity. Most obviously questionable is the approach of dismissing patients whose screening examination is non-diagnostic rather than referring them for duplex Doppler imaging. In the United States, we are willing to spend more than a bare-bones minimum amount on health care, and perhaps we should choose an available point on the ROC curve for power Doppler imaging with the higher sensitivity and lower specificity that is typical of the operation of screening mammography (39,40). For example, as noted, if all patients with nondiagnostic power Doppler imaging examinations or power Doppler imaging findings of 40%–99% stenosis are referred for further work-up, then power Doppler imaging has a sensitivity of 91% and a specificity of 79%.
Study Limitations and Strengths
Like most research, our study had important limitations and strengths. Our total of 120 patients and 240 vessels was far larger than that typical in preliminary studies, but some of the CIs, particularly those for sensitivity and positive predictive value, were wide because of the limited number of patients with disease. In a definitive study, a much larger number of patients will be required.
The estimates of the Az, sensitivity, and specificity of power Doppler imaging in the validation pilot study were strikingly similar to those in the principal pilot study (Table 3; Figs 3, 4). Because these two studies deliberately differed in many respects, this strong similarity provides considerable confidence in the reliability of the results. Moreover, the ability of the staff in the validation study largely to replicate the technique used in the principal study with only brief instruction suggests that our power Doppler imaging technique for evaluating carotid arteries should be readily teachable and transferable. However, because of the somewhat different power Doppler imaging technique used in the two studies (transverse and longitudinal planes in one and longitudinal from two different windows in the other) and the different measurement techniques used, the results of the two studies are not fully comparable. Although this is a weakness of the research, the principal pilot study nonetheless stands as a fully uniform and consistent study with 100 patients, which is a relatively large number.
Duplex Doppler imaging is a less than perfect reference-standard examination and is perhaps especially suspicious because both it and power Doppler imaging are Doppler techniques. However, in the validation study, the results of conventional angiography, which is the best reference-standard method and is not similar to power Doppler imaging in technology, showed power Doppler imaging to have the same examination characteristics as duplex Doppler imaging showed it to have in the main study. This is strong, albeit indirect, evidence that duplex Doppler imaging is an accurate reference-standard method. This finding is beneficial, because the serious morbidity associated with conventional angiography would give rise to serious ethical questions if conventional angiography were used in an asymptomatic population solely for research.
Because US is notably operator dependent, careful attention to technique, specialized training, and certification by a relevant accrediting body (ie, the American College of Radiology, American Institute of Ultrasound in Medicine, or Intersocietal Commission for the Accreditation of Vascular Laboratories) are necessary.
In early evaluations of new imaging techniques, overly positive findings are often reported, in part because "the sickest of the sick," in whom disease is uncommonly easy to detect (41,42), are examined. In contrast, we followed the methodologically appropriate protocol of examining consecutive patients in an actual clinical population. With our principal pilot study having a population that was 71% asymptomatic (as defined by the ACAS) and in which 88% of the vessels were not diseased (that is, <60% stenotic), we avoided the bias of overloading the study population with patients who had severe illness. Furthermore, the strikingly similar results obtained in the validation study population, which had a prevalence of disease three times as high, suggest that power Doppler imaging is about equal in diagnostic accuracy across a broad range of disease severities. However, we included symptomatic patients to have an adequate number of patients with disease to measure sensitivity, and all of our study patients were referred for carotid artery studies. Only in a multiinstitutional study of a nonreferred population with many times more patients can the diagnostic accuracy of power Doppler imaging in a purely screening setting be definitively measured (43).
Because in the ACAS, patients were categorized by percentage of stenosis and endarterectomy was evaluated accordingly, we studied power Doppler imaging's capability to enable measurement of percentage of stenosis. Previously published reports (6,44,45) indicate that plaque morphology is indicative of stroke risk. If complete duplex Doppler imaging is used in postscreening work-up, it can help evaluate plaque morphology.
Cost-effectiveness findings are very sensitive to the assumptions those conducting the analysis have used. Most conspicuously, Lee et al (11) reached the conclusion that cost-effective screening is not possible; however, Yin and Carpenter (12) estimated the cost-effectiveness for the same pattern of care to be a favorable $40,000 per QALY. We, being deliberately conservative, used the model of Lee et al (11) and made a minimum number of emendations to it. In any case, the results of our work show that power Doppler imaging screening followed by duplex Doppler imaging or MR angiography as the definitive diagnostic examination is much more cost-effective than duplex Doppler imaging followed by conventional angiography, which is the pattern of diagnosis modeled by both Lee et al (11) and Yin and Carpenter (12).
Because our clinical research dealt with screening, we modeled the definitive diagnostic examination simply as a single imaging examination. However, detailed modeling of diagnostic examinations for symptomatic patients indicates that the best health outcomes result from a combination of imaging examinations (46). A combination strategy for definitive diagnostic examinations may also be superior in asymptomatic screening populations.
A discussion of the (mostly surgical) controversy (5–10) about the benefits of endarterectomy for asymptomatic patients is beyond the scope of this article; however, this issue is obviously fundamental to the optimal treatment of these patients. We note here that (a) resolution of this controversy may come only from further, more varied randomized clinical trials, in which a good screening examination would be valuable for identifying patients and (b) for those who judge endarterectomy to be beneficial, a good screening examination is needed.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, E.I.B.; study concepts, E.I.B., J.H.S., C.A.B.; study design, E.I.B., J.H.S., H.L.N.; definition of intellectual content, E.I.B., J.H.S.; literature research, E.I.B., P.E.C.; clinical studies, E.I.B., J.B.L., H.L.N.; data acquisition, E.I.B., J.B.L., L.A.T., L.A.K., M.A.S., D.H.S.; data analysis, C.A.B, P.E.C., P.A.H., J.H.B.; statistical analysis, C.A.B., P.E.C.; manuscript preparation, E.I.B., J.H.S.; manuscript editing, J.H.S., E.I.B., P.E.C.; manuscript review, all authors.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. R. Bates, C. J. D. Babb, D. E. Casey, C. U. Cates, G. R. Duckwiler, T. E. Feldman, W. A. Gray, K. Ouriel, E. D. Peterson, K. Rosenfield, et al. ACCF/SCAI/SVMB/SIR/ASITN 2007 Clinical Expert Consensus Document on Carotid Stenting: A Report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents (ACCF/SCAI/SVMB/SIR/ASITN Clinical Expert Consensus Document Committee on Carotid Stenting) Vascular Medicine, February 1, 2007; 12(1): 35 - 83. [PDF]
|
|
|
|
 |
|
 American Society of Interventional & Therapeutic N, Society for Cardiovascular Angiography and Interve, Society for Vascular Medicine and Biology, Society of Interventional Radiology, E. R. Bates, J. D. Babb, D. E. Casey Jr, C. U. Cates, G. R. Duckwiler, T. E. Feldman, et al. ACCF/SCAI/SVMB/SIR/ASITN 2007 Clinical Expert Consensus Document on Carotid Stenting: A Report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents (ACCF/SCAI/SVMB/SIR/ASITN Clinical Expert Consensus Document Committee on Carotid Stenting) J. Am. Coll. Cardiol., January 2, 2007; 49(1): 126 - 170. [Full Text] [PDF]
|
|
|
|
 |
|
 P. S. Mullenix, M. J. Martin, S. R. Steele, G. S. Lavenson Jr, B. W. Starnes, N. C. Hadro, R. P. Peterson, and C. A. Andersen Rapid High-Volume Population Screening for Three Major Risk Factors of Future Stroke: Phase I Results Vascular and Endovascular Surgery, May 1, 2006; 40(3): 177 - 187. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Krejza, J. Kochanowicz, Z. Mariak, J. Lewko, and E. R. Melhem Middle Cerebral Artery Spasm after Subarachnoid Hemorrhage: Detection with Transcranial Color-coded Duplex US Radiology, August 1, 2005; 236(2): 621 - 629. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Wessels, J. U. Harrer, S. Stetter, M. Mull, and C. Klotzsch Three-Dimensional Assessment of Extracranial Doppler Sonography in Carotid Artery Stenosis Compared With Digital Subtraction Angiography Stroke, August 1, 2004; 35(8): 1847 - 1851. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Nederkoorn, Y. van der Graaf, and M.G. M. Hunink Duplex Ultrasound and Magnetic Resonance Angiography Compared With Digital Subtraction Angiography in Carotid Artery Stenosis: A Systematic Review Stroke, May 1, 2003; 34(5): 1324 - 1331. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. I. Bluth, M. Claudon, D. Winninger, and P. Pesque Power Doppler Imaging to Evaluate Flow-limiting Stenoses Dr Claudon and colleagues respond: Radiology, November 1, 2001; 221(2): 557 - 561. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) pilot studies. The ROC curves and Az values (AUC) in each study are similar.