Eur J Cardiothorac Surg 2008;33:633-638. doi:10.1016/j.ejcts.2007.12.039
Copyright © 2008, European Association for Cardio-thoracic Surgery. Published by Elsevier. All rights reserved.
Can we perform coronary artery bypass grafting on the basis of computed tomographic angiography alone? A comparison with conventional coronary angiography
Harinder Singh Bedia,*,
Jeet Amol Singh Gillb,
Sandeep Singh Bakshib
a Ludhiana Mediciti, Ferozpur Road, Ludhiana 142001, Punjab, India
b Delta Heart Centre, Ludhiana, Punjab, India
Received 4 September 2007;
received in revised form 5 December 2007;
accepted 11 December 2007.
* Corresponding author. Tel.: +91 161 5099510; fax: +91 161 5099506. (Email: drhsbedi{at}hotmail.com).
 |
Abstract
|
|---|
Objectives: Multislice (64 slice) computed tomography (MSCT) angiography has recently emerged as a potential technique that can evaluate the coronaries in an accurate yet non-invasive manner. It has still not been shown whether the accuracy of the anatomy shown is precise enough to operate on the basis of MSCT alone. The aim of this prospective clinical trial is to compare MSCT to conventional coronary angiography (CCA), and to conclude whether MSCT alone is adequate for proceeding for CABG. Methods: 50 patients with proven severe coronary artery disease (CAD) on CCA for elective CABG underwent MSCT prior to CABG. The MSCT images were compared with CCA and the accuracy, sensitivity and specificity of detecting significant stenosis cross checked. Lesion-by-lesion analysis was made. CCA was used as the reference standard for location and degree of stenosis. Results: An excellent correlation was found between the CCA and MSCT findings. The overall sensitivity, specificity, positive (PPV) and negative (NPV) predictive values for quantitative assessment of stenosis >70% by MSCT compared to CCA were 98.5, 99.1, 82.3 and 99.8%, respectively. Comparing the maximal percent diameter luminal stenosis by MSCT versus CCA, the Pearson's correlation coefficient between the two modalities was 0.994 (p
< 0.0001). Bland-Altman analysis demonstrated a mean difference in percent stenosis of 0.05 ± 2.42% (p
= 0.753). There was no significant correlation between stenosis difference and stenosis severity (Pearson's correlation coefficient = –0.027, p
= 0.695). 192 out of 207 (92.8%) of the observations were within ±1.96 SD (4.8 to –4.7% stenosis difference). Conclusions: The improved spatial and temporal resolution of the 64 row scanner provides an excellent correlation of MSCT with CCA. MSCT is a valuable tool in the armamentarium of the cardiac scientist. For the cardiac surgeon performing off pump CABG it helps in precise planning of the procedure and pre-judging the length of the conduit required. On the basis of our findings, in selected patients, we recommend the consideration of MSCT as a sole criteria for proceeding for CABG without CCA.
Key Words: Computed tomography • Conventional coronary angiography • Coronary artery disease
|
1. Introduction
|
|---|
Rapid advances in multislice computed tomography (MSCT) have facilitated increasingly accurate non-invasive coronary imaging. The improved spatial and temporal resolution generating 64 slices per rotation and covering the entire volume of the heart in 6–8 s promises a significant improvement of image quality that may allow a more precise evaluation of coronary stenosis than the previous 16 slice scanner. The present study was designed to assess the accuracy of the 64 slice MSCT scanner in patients with coronary artery disease (CAD) diagnosed on conventional coronary angiography (CCA) and to conclude whether or not MSCT angiography alone could give sufficient and accurate information regarding presence and quantification of atherosclerotic coronary lesions to proceed with coronary artery bypass grafting (CABG).
|
2. Methods
|
|---|
2.1 Study population
55 consecutive patients with stable angina scheduled for elective CABG for severe coronary artery disease (CAD) diagnosed by CCA underwent MSCT within 30 days of CCA. 5 patients were excluded—the exclusion criteria being irregular heart rate (1 patient), morbid obesity (1 patient) and calcium score >600 units (3 patients). The mean age was 60 ± 10 years and 70% were males. 78% had diabetes mellitus, 64% were chronic smokers, 76% were hypertensives and 66% had dyslipidemia. The human investigation committee approved the study protocol and all patients gave informed consent.
2.2 Patient preparation
In the absence of contraindications, patients not already on beta-blockers were given 50 mg of oral metaprolol for heart rates over 75 beats/min 1 h before MSCT. Heart rate, electrocardiogram (ECG) and blood pressure were monitored and additional intravenous metaprolol (in incremental doses of 5 mg) administered to achieve a target heart rate of <75 beats/min.
2.3 Scan protocol and image reconstruction
All patients were scanned on a 64 slice scanner (GE, Lightspeed, VCT, WI, USA). An initial non-enhanced ECG gated scan was performed for calcium scoring and 3 patients with high calcium scores excluded at this stage. Coronary angiographic scan was done with real time contrast monitoring and initiation of scan at peak of opacification of root of aorta was performed using ‘smart prep’ technique. 65–70 ml of contrast injection (Iomeron, Braco Italy, iodine content 400 mg/ml) was administered at the rate of 5 ml/s followed by a saline chaser of 40 ml. As soon as the signal density in the ascending aorta reached a predefined threshold of 100 Hounsfield units (HU) the patient was instructed to maintain an end-inspiratory breath hold during which the CT volume data set and ECG trace were acquired. The scan parameters were: detector configuration 64 mm x 0.625 mm slices, collimation of 40 mm, rotation time 350 ms, tube voltage 120 kV, ECG modulated effective 200–800 mA, volumetric CT index 59 mGy, resolution of 0.4 mm, a slice thickness of 0.6 mm, a temporal resolution of 175 ms with single sector reconstruction, with an estimated effective radiation dose 13 mSv for men and 18 mSv for women. Electrocardiographically gated datasets were reconstructed automatically at 75% of the R-R cycle length and 35% of the R-R cycle length to approximate end-systole and end-diastole, respectively. The simultaneous acquisition of 64 parallel cross sections enabled the imaging of the entire coronary tree in a single breath hold of about 6–8 s.
2.4 MSCT analysis
CTA were analysed on a three-dimensional workstation (Advantage Windows 4.2, GE Healthcare, WI, USA) using a combination of axial images, multiplanar reconstruction, curvilinear reconstruction with vessel tracking, straight vessel views, cross sectional views and 3D volume rendered images. Each vessel was interrogated one at a time from cranial to caudal direction. As the artery is a 3D structure which changes direction during its course, it is necessary to scroll back and forth as one follows it from proximal to distal end. Once a location of luminal narrowing was identified in the vessel straight view—diameters were measured in cross sectional view using electronic caliper across the best cross section of the vessel for verification and quantification of stenosis (Fig. 1
). A previously described 15 segment American Heart Association model of the coronary tree was employed [1]. Analysis was by consensus of two observers unaware of the clinical data and the CCA findings. Each lesion was identified and examined using the above mentioned multiplanar reconstruction techniques along multiple longitudinal axes and transversely. Lesions were classified by the maximal luminal diameter stenosis seen in any plane. Lesion sites with clear well defined borders were measured quantitatively. Quantitative CT angiographic analysis was performed on the most severe well-defined lesion in each segment. In the case of multiple lesions in a given segment, the segment was classified by the worst lesion. In the case of multiple abnormal segments per-artery, the vessel was classified by the worst segment. In this study patients were classified as positive for the presence of significant CAD if there was diameter stenosis 
70% in any artery. Volume rendering techniques provided a general overview of the cardiac and coronary anatomy. These images give a good 3D virtual image and can help in planning the length of the conduit to be harvested as the curved distance to the aorta for a free graft can be measured (Fig. 2
). Anomalous coronary arteries were easily identified in two cases—in one this had been missed on CCA. Segments smaller than 1.0 mm lumen diameter were not evaluated as we are not presently grafting vessels of this size.
View larger version (151K):
[in this window]
[in a new window]
|
Fig. 1. Method of quantifying stenosis in MSCT. (A) CCA image showing stenosis in mid RCA (arrow); (B) MSCT multiplanar reconstruction showing exactly the sane findings; (C) MSCT straight vessel view; (D) diameter in cross sectional view of normal proximal segment; (E) diameter in cross sectional view of stenosis.
|
|
View larger version (118K):
[in this window]
[in a new window]
|
Fig. 2. MSCT volume rendered image showing the pre-judging the length of the conduit required for bypassing the obtuse marginal coronary artery.
|
|
2.5 CCA analysis
CCA was performed with the Judkins trans-femoral technique. 2D imaging of the coronary arteries was done in three biplane projections. Segmental disease was analysed in each vessel using the same 15 segment model employed for MSCT analysis. Stenosis severity in each segment was classified according to the maximum luminal diameter stenosis present in each segment. Lesions were examined in orthogonal views and stenosis severity determined using an automated edge-detection system.
2.6 Comparative analysis
CCA findings were taken as standard values and the MSCT findings were compared in terms of specificity, sensitivity, positive and negative predictive value. Quantitative per lesion analysis compared the maximal percent diameter stenosis of the most severe lesion in each segment by each imaging modality. The accuracy of CTA to detect CAD was compared to CCA according to the following analyses:
- (1) per-segment analysis, comparing each segment in every vessel;
- (2) per-artery analysis examining the presence of significant lesions in each of the major coronary vessels (right coronary artery, left circumflex, left anterior descending and left main).
2.7 Statistical analysis
All data were stored on Microsoft Excel and transferred to SPSS 13.0 (SPSS Inc., Chicago, IL). The diagnostic accuracy of MSCT to detect coronary lesions was evaluated using CCA as the reference standard. Paired Student's t-test was used for analysing difference from zero. Quantitative lesion severity was compared between MSCT and CCA using Pearson's correlation and Bland-Altman analysis, performed by plotting the stenosis difference between MSCT and CCA versus the CCA stenosis [2]. For analysis of sensitivity, specificity and positive and negative predictive accuracy, CCA lesion severity was used as the ‘gold standard’ and comparison was by 2 x 2 cross tabulation. A two-tailed p
< 0.05 was considered statistically significant.
|
3. Results
|
|---|
According to inclusion criteria MSCT was performed successfully without any complication or patient discomfort.
3.1 MSCT compared to CCA for quantitation of lesion severity
Altogether 740 coronary segments were correlated (10 segments were excluded because of reduced image quality). Of these, 207 of 740 had stenosis. Image quality and spatial resolution of even distal segments and side branches were excellent (Figs. 3–6
). MSCT was accurate in the quantitative assessment of lesion severity. Comparing the maximal percent diameter luminal stenosis by MSCT versus CCA, the Pearson's correlation coefficient between the two modalities was 0.994 (p
< 0.0001) (Fig. 7
).
View larger version (89K):
[in this window]
[in a new window]
|
Fig. 3. (A) Conventional coronary angiogram (CCA) showing multiple lesions (small arrows) in right coronary artery (RCA); (B) multislice computed tomogram (MSCT) showing the same picture. Additionally the MSCT shows areas of non-stenotic calcification (arrow heads).
|
|
View larger version (90K):
[in this window]
[in a new window]
|
Fig. 4. (A) CCA and (B) MSCT multiplanar reconstruction showing comparable lesions in mid-left anterior descending coronary artery.
|
|
View larger version (57K):
[in this window]
[in a new window]
|
Fig. 5. CCA image (A) showing an ostial left anterior descending coronary artery (LAD) stenosis (single arrow), lesion in circumflex artery (double arrow) and stenosis in obtuse marginal (arrow head); (B and C) comparable findings in multiplanar and volume rendered images of MSCT, respectively. Additionally an area of calcification is picked up by MSCT (white speck between two arrows).
|
|
View larger version (75K):
[in this window]
[in a new window]
|
Fig. 6. Multiple lesions in posterior descending coronary artery imaged exactly the same in CCA (A) and MSCT (B).
|
|
View larger version (9K):
[in this window]
[in a new window]
|
Fig. 7. Correlation between maximum percent diameter stenosis by quantitative analysis of MSCT compared with CCA. The Pearson's correlation showed an R-value of 0.994, p
< 0.001, n
= 207.
|
|
Bland-Altman analysis (Fig. 8
) demonstrated a mean difference in percent stenosis of 0.05 ± 2.42% (p
= 0.753). There was no significant correlation between stenosis difference and stenosis severity (Pearson's correlation coefficient = –0.027, p
= 0.695). 192 out of 207 (92.8%) of the observations were within ±1.96 SD (4.8 to –4.7% stenosis difference).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 8. Bland-Altman analysis of the differences of percent diameter stenosis measured by MSCT and CCA compared to the diameter stenosis by CCA.
|
|
3.2 CTA compared to CAG for detection of significant stenosis
Overall on a per-segment basis MSCT had a sensitivity of 98.5% in detecting significant lesions, and a specificity of 99.1%. The positive predictive value was 82.3% and the negative predictive value was 99.8%. On a per-artery basis the values were—sensitivity 99.4%, specificity 84.6%, positive predictive value 97.7% and negative predictive value 95.6%.
On a patient basis all patients were correctly identified as having lesions which warranted CABG per se. In retrospect, the image quality was clear enough in all cases to be the sole diagnostic criteria before CABG.
Intramural segments of the coronary arteries (confirmed at surgery) were more frequently picked up by MSCT (n
= 4) than by CCA (n
= 1).
In 5 patients, distal vessels not visualised on CCA due to total occlusion were clearly seen on MSCT. In 3 of these patients additional lesions and calcified spots seen on the MSCT helped us to plan additional sequential grafts to revascularize all segments of the same vessel. Vessels seen to be totally occluded on MSCT were also totally occluded on CCA.
|
4. Discussion
|
|---|
The last few years have witnessed the emergence of MSCT as a minimally invasive coronary imaging tool. The technology has improved from the previous electron-beam CT and 4-row MSCT to the 16 row and now the 64 row MSCT scanners—that offer an isotropic voxel resolution of 0.4 mm x 0.4 mm x 0.4 mm. The 64 slice scanner is a major improvement due to the improved acquisition time and the fact that the entire scan time is significantly shortened to 6–8 s, thus making this technology more robust against respiratory and motion artifacts of the patient.
To become a clinically accepted modality for the cardiac surgeon, the mandatory pre-requisite is that MSCT must accurately visualise all the therapeutically relevant coronary artery segments, and the degree of stenosis found must correlate with that of CCA. We have found that the 64 slice MSCT scanner provides an excellent correlation with CCA and accurately and consistently delineates the presence or absence of significant lesions in the coronary arteries. It is able to give reliable information on the presence, severity and characteristics of luminal stenosis and atherosclerotic plaque.
Our findings are fairly consistent with previous studies [3–7] some of which have shown a 100% correlation [7]—using the 64 slice scanner, and no exclusion of any segment for analysis [5]. A very high specificity and NPV approaching 100% have been reported [8]. Even with the 16 slice MSCT correlations have consistently exceeded 95% for negative predictive value [9]. A number of studies have been able to analyse all segments [5,10–12] with 0% incidence of unanalyzable segments—which closely matches with our study. The 64 slice scanner in comparison to the previous 16 slice machines has increased slices per gantry rotation (64 vs 16) and faster gantry speed (500 ms/rotation vs 350) which translates into superior spatial resolution (0.35 vs 0.60 mm) and temporal resolution (44 vs 120 ms). The reduction in voxel size makes the distinction between hypointense soft plaque and blood pool contrast more evident. Smaller voxel size also reduces partial volume effects, minimising the degree of calcium blooming and beam hardening artifacts. Consequently this study was able to provide quantitative comparison of percent diameter stenosis of individual lesions by MSCT with CCA, showing a mean difference of 0.6 and a relatively small standard deviation of 2.3.
MSCT being non-invasive can be performed as an out-patient procedure. To date there has been no report of any major fatal complication. Due to the relative ease of performance, non-invasiveness, and short duration of examination with no need for hospital stay—it has obvious advantages for the patient and the clinician and will increase the detection of a disease which can be disabling or fatal if not detected in time. The cost of a MSCT in India is exactly the same as for a CCA—approximately US $250. As predicted by some [13], we also feel that it will gradually replace diagnostic invasive CCA as the primary imaging modality for the evaluation of coronary arteries. The scan is especially useful to the cardiac surgeon for besides being consistently accurate as shown in our study, it provides a virtual 3D image, gives important information about the vessel wall and its surrounding structures besides just the luminal diameter, shows the presence and situation of calcification so that the site of anastomosis can be pre-planned and also the length of conduit required can be accurately judged preoperatively so that only the appropriate length is harvested. This can have important bearing especially in off pump and in minimally invasive surgery.
MSCT allowed differentiation of lesions on the basis of calcified, fatty or fibrous proliferation. Some calcified non-obstructive lesions not seen on CCA were easily visualised on MSCT (Figs. 3 and 5). This can help in the pre-planning of the site of anastomosis.
MSCT also exactly verifies the intra-thoracic anatomy and topographic relationship of intra-thoracic organs and the chest wall. This additional information can be important in the precise planning of re-do CABG (distance between graft and right ventricle and sternum) and MIDCAB procedures (spatial relationship between LAD and LIMA) [14] (none in this study).
MSCT can give information also about the pathogenesis of the stenosis and about plaque vulnerability. Pixel density above 130 HU indicates calcification, whereas measurement of less than 130 HU indicates fibrous or fatty lesions. Patients with inhomogenous atherosclerotic plaques with calcified and non-calcified components susceptible to rupture and thrombotic occlusions are theorised (unproven) to be at higher risk for myocardial infarction [14]. Such non-critical lesions in non-bypassed arteries might warrant a more aggressive secondary prevention post CABG than otherwise indicated.
Observations from our study demonstrate that the 64 slice MSCT scanner provides high quality non-invasive coronary angiograms that accurately delineates coronary stenosis in a manner comparable to CCA.
4.1 Study limitations
A shortcoming of our study is that we excluded patients with high calcium scores (>600 Agatston units), obesity (BMI >30 kg/m2) and high heart rates (>75 min–1)—which are a good proportion of ‘real world’ patients. A previous study [6] has shown an overall sensitivity of 95% and specificity of 90% for the detection of angiographically significant stenoses even in the presence of high coronary calcium scores (up to 1804 U), increased heart rates (up to 96 beats/min) or obesity (BMI up to 46 kg/m2). Also there is a bias in our study as we were only studying patients with known severe CAD. We are now studying patients in the excluded group to find the degree of correlation of MSCT with CCA in them. A continuing study to recruit a larger group of patients is also underway at present.
|
5. Conclusion
|
|---|
On the basis of our study, the excellent image quality and the high sensitivity and specificity in the detection and quantitative grading of CAD that is comparable to CCA—we recommend the use of contrast enhanced MSCT angiography in selected patients with slow heart rates in sinus rhythm as the sole imaging diagnostic criteria to proceed with CABG where indicated.
|
Appendix A
|
|---|
Conference discussion
Dr I. Kron (Charlottesville, VA): We in fact use this modality to rule out coronary disease in slightly younger patients who require valvular heart surgery. Your data is convincing. There is just one issue, and that issue is as you discussed: it relates to the calcified lesion. When I had this paper to review, I spoke with two of our cardiologists who are quite expert in this technology. They showed me examples of calcified lesions that looked severe by this imaging modality but in fact did not have flow-limiting lesions. As you well know, a significant stenosis is required to put a radial artery on a circumflex, for example.
Many in Virginia have calcified plaques and are obese. So how do I plan based on this modality? The second question that I think would be interesting to the audience is, what is the relative cost of this procedure compared to angiography?
Dr Bedi: The first question was, what do I do in obese patients with calcification? In lesions which are heavily calcified, this modality would not help us. We would have to have a conventional angio. In obese patients, although we did not have any obese patients in our study, but in the recent study from JACC 2005, they studied obese patients, they studied patients with calcium scores over 600, and they found a 95% relationship between the two modalities.
Your second question regarding cost, the cost to the patient is the same for a conventional angio as for a CT angio.
Dr Y. Balbaa (Cairo, Egypt): My question for you is an ethical issue. Your patients already had conventional coronary angiography and then they were scheduled for surgery, and then before surgery they had another modality, the 64 multislice CT. I am aware that this process is not without risk. It is major radiation for each patient with some minimal risk of malignancy. So my question is, how did you manage this ethically?
Dr Bedi: Basically at this point of time today there is really no risk, no additional risk of a CT angio in a patient who has normal renal function. So that is number one. If there is renal dysfunction, this would have been excluded. Other than that, really there is no risk in a patient with stable coronary artery disease. There is no additional risk we are putting the patient to.
Dr Balbaa: There is a chance of 1 in 2000 of malignancy, because one CT angiography with a 64 multislice is equivalent to having 90 chest X-rays as a radiation dose.
Dr Bedi: Yes, the radiation dose definitely is higher. That is true.
Mr K. Morcos (Cairo, Egypt): Can you please comment on when you have stents, how well does it see the stents and does this correlate well with the angio?
Dr Bedi: I am not the right person to answer that, but yes, it is one of the good modalities to look for instent restenosis. But exactly what is the correlation, I would not be knowing today.
|
Acknowledgments
|
|---|
We acknowledge with thanks the contribution of Mr Sudhir Shekhawat MSc for the statistical analysis.
|
Footnotes
|
|---|
Presented at the 21st Annual Meeting of the European Association for Cardio-thoracic Surgery, Geneva, Switzerland, 16–19 September 2007.
|
References
|
|---|
- Austen WG, Edwards JE, Frye RL, Gensini GG, Gott VL, Griffith LS, McGoon DC, Murphy ML, Roe BB. A reporting system on patients evaluated for coronary artery disease. Report of the Ad-Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery. Circulation 1975;51(Suppl.):5-40.[Medline]
- Altman DG, Bland JM. Measurement in medicine: the analysis of method comparison studies. Statistician 1983;32:307-317.[CrossRef]
- Hoffmann M, Shi H, Schmitz B, Schmid F, Lieberknecht M, Schulze R, Ludwig B, Kroschel U, Jahnke N, Haerer W, Brambs H, Aschoffet A. Noninvasive coronary imaging with multirow computed tomography. JAMA 2005;293:2471-2478.[Abstract/Free Full Text]
- Leber A, Knez A, von Ziegler F, Becker A, Nikolau K, Paul S, Wintersperger B, Reiser M, Becker CR, Steinbeck G, Boekstegers P. Quantification of obstructive and nonobstructive coronary lesions by 64 row computed tomography. J Am Coll Cardiol 2005;46:147-154.[Abstract/Free Full Text]
- Leschka S, Alkadhi H, Plass A, Desbiolles L, Grunefelder J, Marincek B, Wildermuth S. Accuracy of MSCT coronary angiography with 64-row technology: first experience. Eur Heart J 2005;26:1482-1487.[Abstract/Free Full Text]
- Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnostic accuracy of noninvasive coronary angiography using 64-slice spiral computed tomography. J Am Coll Cardiol 2005;46:552-557.[Abstract/Free Full Text]
- Vrtiska TJ, Flectcher JG, McCulough CH. State-of-the-art imaging with 64-channel multidetector computed tomography angiography. Perspect Vasc Surg Endovasc Ther 2005;17:T1.[CrossRef]
- Ong TK, Chin AP, Liew CK, Chan WL, Seyfarth MT, Liew HB, Rapace A, Fong YYA, Ang CK, Sim KH. Accuracy of 64-row multidetector computed tomography in detecting coronary artery disease in 134 symptomatic patients: influence of calcification. Am Heart J 2006;151:1323e1-1323e6.
- Becker CR, Coronary. CT angiography in symptomatic patients. Eur Radiol 2005;15(Suppl. 2):B33-B41.[CrossRef][Medline]
- Mollet NR, Cademartiri F, van Mieghem C, Runza G, McFadden E, Baks T, Serruys PW, Krestin GP, de Feyteret PJ. High resolution spiral computed tomography coronary angiography in patients referred for diagnostic conventional coronary angiography. Circulation 2005;112:2318-2323.[Abstract/Free Full Text]
- Pugliese F, Mollett NR, Runza G. Diagnostic accuracy of non-invasive 64 slice CT angiography in patients with stable angina pectoris. Eur Radiol 2006;16:575-582.[CrossRef][Medline]
- Gibbons RJ, Araoz PA, Williamson EE. The year in cardiac imaging. J Am Coll Cardiol 2006;48:2324-2339.[Free Full Text]
- Gaspar T, Dvir D, Peled N. The role of 16-slice computed tomography angiography in the diagnosis of coronary artery disease: large sample analysis. IMAJ 2005;7:424-427.[Medline]
- Treede H, Becker C, Reichenspurner H, Andreas K, Christian D, Maximilian R, Bruno R. Multidetector computed tomography in coronary surgery: first experiences with a new tool for diagnosis of coronary artery disease. Ann Thorac Surg 2002;74:S1398-S1402.[Abstract/Free Full Text]
Top
Abstract
1. Introduction
