0353-07_ActaCardio_62_4_04_2881-07
06-08-2007
Acta Cardiol 2007; 62(4): 339-344
09:37
Pagina 339
doi: 10.2143/AC.62.4.2022276
339
Diagnostic accuracy of computed tomography coronary angiography in routine practice
Laurent DAVIN, Patrizio LANCELLOTTI, Pierre-Julien BRUYERE*, Olivier GACH, Luc PIERARD, Victor LEGRAND Division of Cardiology and Radiology*, University Hospital of Liège, Belgium. Objectives — The recent newer advances in computed tomography have dramatically changed our approach to imaging cardiac disease. This study sought to compare the diagnostic value of 16-multi-detector spiral computed tomography (MSCT) for detecting coronary artery stenosis. Methods — A total of 88 consecutive patients (52∞∞men, mean age 68∞∞±∞∞8∞∞years) with atypical chest pain, stable angina or suspicion of ischaemia at stress test were studied by MSCT and invasive coronary angiography (ICA). The MSCT images and multiplanar reconstructions were analysed regarding the presence of∞∞≥∞∞50% coronary artery lesion. Results — All 88∞∞scans obtained at a mean heart rate of 68∞∞±∞∞8∞∞beats/min were interpretable. Sixteen coronary segments were evaluated in each patient. Of the 1320∞∞segments examined, 148 (11%) showed poor image quality.A total of 150∞∞significant lesions were detected using ICA, and 80 of 150 (53%) were detected by MSCT. Sensitivity, specificity, positive and negative predictive values were as follows: 53%, 97%, 68%, and 94%. Fifty-four patients had∞∞≥∞∞50% coronary stenosis.The diagnosis was confirmed by MSCT in 42∞∞patients and correctly ruled out in 30. By patient-based analysis, positive and negative predictive values were 91% and 71%. Conclusion — Although its specificity is high, the sensitivity of 16-slice MSCT for detecting∞∞≥∞∞50% coronary stenosis in non-selected patients submitted to ICA is rather low suggesting that for daily practice the diagnostic value of this technique should be improved. Keywords: coronary angiography – MSCT – coronary artery disease.
Introduction Imaging the heart has always been technically challenging because of the heart‘s continuous motion. Since 1999, mechanical multi-detector spiral computed tomography systems have been available for non-invasive detection of coronary artery disease (CAD)1-3. Since 2002 the use of a technically improved 16-row multislice spiral computed tomography (MSCT) with higher spatial and temporal resolution allowed better detection of significant obstructive coronary lesions45. Initial reports on the ability of contrast-enhanced MSCT to visualize coronary atherosclerotic stenosis have since received widespread attention6-8. Even if its
negative predictive value is quite high, recent studies have reported that MSCT has a poor sensitivity in detecting and quantifying coronary artery stenosis9-10. Therefore, MSCT should be used as a diagnostic tool in ruling out coronary atherosclerosis in patients with a low prevalence of CAD. The absence of significant coronary lesions on MSCT could indeed exclude the need for an invasive diagnostic strategy. This study was thus undertaken to assess whether cardiac MSCT has an acceptable diagnostic value in detecting and/or excluding CAD in a non-selected patient population submitted to invasive coronary angiography (ICA).
Methods STUDY Address for correspondence: Prof. Patrizio LANCELLOTTI, M.D., Ph.D., Department of Cardiology, University Hospital of Liège, B-4000 Liège, Belgium. E-mail:
[email protected] Received March 20, 2007; revision accepted for publication June 7, 2007.
POPULATION
During a period of 12∞∞months, from October 2005 to September 2006, 88∞∞patients (mean age: 68∞∞±∞ 8∞ years) at low to moderate cardiovascular risk were prospec-
0353-07_ActaCardio_62_4_04_2881-07
340
06-08-2007
09:37
Pagina 340
L. Davin et al.
tively studied by MSCT. Exclusion criteria were: irregular rhythm, contraindications to iodinated contrast material (e.g. impaired renal function, thyroid disorders or known allergy). Half of the patients were treated by beta blockers. All patients underwent a diagnostic invasive coronary angiography (ICA) within one week before or after MSCT. All patients gave informed consent and the study was approved by the hospital ethics committee.
SCAN
STATISTICAL
ANALYSIS
Data are expressed as mean ± SD. Sensitivity, specificity and likelihood ratios were calculated using 2∞∞≈∞∞2 table for dichotomous tests (positive or negative). For the analysis of the diagnostic accuracy of MSCT per coronary artery segments, all segments adequately assessed or not by coronary angiography were considered. All calculations were done with the use of a commercially statistical programme (Statistica version 6).
PROTOCOL, IMAGE RECONSTRUCTION AND
EVALUATION
Results All patients were studied with the same equipment (Sensation16 Straton, Siemens Medical Solution). The scan parameters were: 16∞∞≈∞ 0.75∞∞mm collimation; rotation time 370∞∞ms; table feed 2.6 to 3∞∞m/rotation; tube voltage 120 kV, effective mA: 500 to 700 (depending on age and weight/length); tube with modulation of irradiation. A bolus of 80∞∞ml contrast material with iodine content of 350∞∞mg/ml was injected through an arm vein (rate: 4∞∞ml/s). We used a bolus tracking to synchronize the arrival of contrast in the coronary arteries and the initiation of the scan data were acquired during a breath-hold of∞∞±∞∞20∞∞sec. Image reconstruction was retrospectively gated to the electrocardiogram and position of the reconstruction window within the cardiac cycle was individually optimized to get the best image quality. All scans were independently analysed by one observer unaware of the invasive coronary angiography results. Coronary arteries were evaluated irrespective of image quality but only branches larger than 1.5∞∞mm were studied. Maximum intensity projections were used to identify lesions and multi-planar reconstructions to classify lesions as non-significant or significant (∞∞≥∞ 50% diameter stenosis by visual assessment). Lesion location was allocated to coronary segments, according to the modified American Heart Association classification11.
QUANTITATIVE
CORONARY ANGIOGRAPHY
Coronary angiography was performed using Judkins technique with 6 F coronary catheters by the femoral approach. All angiograms were evaluated by two blinded independent reviewers. Lesions with∞∞≥∞ 50% diameter stenosis at visual evaluation were quantified using commercially available software (Integris HM 3000 Philips Medical System, The Netherlands) and were considered as significant. Angiographic measurements assessed by quantitative coronary angiography (QCA) were considered as gold standard and compared with the corresponding MSCT visual evaluation.
Clinical presentation was atypical chest pain in 28∞∞patients (33%), stable angina in 8 (10%), silent ischaemia in 20 (22%) and equivocal stress test in 32 (35%). Twenty patients were smokers (22%). Dyslipidaemia, hypertension and diabetes were diagnosed and treated in 48 (55%), 48 (55%) and 16 (18%) patients, respectively. Based on the European SCORE scale which evaluates the 10-year risk of fatal cardiovascular disease, a ten-year risk of 1%, 2%, 3-4% and 5-9% were calculated in 42, 16, 16, and 14 patients, respectively. MSCT was performed without complication in all 88∞∞patients. The mean heart rate at scan time was 78∞∞bpm (range 63 to 92∞∞bpm).
SEGMENT-BASED
ANALYSIS
Coronary angiography identified 150∞∞significant lesions (∞∞>∞∞50%) in 1320∞∞segments (prevalence: 11%). Proximal lesions (n∞∞=∞∞64) in segments 1-2, 5-7 and 11 (20, 36, and 8 lesions, respectively) were less frequent than distal lesions (n∞∞=∞∞86) in segments 3-4, 8-10, 1215 (24, 26, and 36 lesions, respectively). Overall, MSCT detected 80 lesions and failed to detect stenosis in 70∞∞segments with∞∞>∞∞50% stenosis (sensitivity 53%). The specificity, positive predictive value and negative predictive values were 97%, 68% and 94%, respectively (figure 1A). The sensitivity increased from 53% to 66% for detection of stenosis ≥ 70% without reduction in specificity (96%) (figure 1B). Poor image quality precluded accurate evaluation of coronary anatomy by MSCT in 148∞∞segments (11%). Exclusion of these segments from the analysis improved the diagnostic accuracy of MSCT, yielding a sensitivity of 66%, a specificity of 96%, a positive predictive value of 64% and a negative predictive value of 96% (figure 2A). MSCT analysis restricted to the proximal part of vessels (segments 1-2-5-6-7-11) also improved sensitivity (80%) and specificity (93%) (figure 2B). The positive and negative predictive values were 62% and 97%.
0353-07_ActaCardio_62_4_04_2881-07
06-08-2007
09:37
Pagina 341
Diagnostic accuracy of MSCT in routine practice
341
Fig. 1. – Sensitivity and specificity for the detection of coronary artery disease (CAD). A:∞∞≥∞∞50% stenosis detection; B: different degree of stenosis.
Fig. 2. – Sensitivity and specificity for the detection of∞∞≥∞∞50% coronary artery stenosis. A: analysis restricted to analysable segments by MSCT; B: analysis limited to proximal segments.
PATIENT-BASED
ANALYSIS
CAUSES
OF FALSE-POSITIVE AND FALSE-NEGATIVE
RESULTS
Fifty-four patients (61%) had significant CAD (1vessel disease in 26∞∞patients, 2-vessel disease in 18 and 3-vessel disease in 10). The diagnosis of significant CAD was confirmed by MSCT in 42∞∞patients and correctly ruled out in 30∞∞patients. Twelve patients with angiographic stenosis∞∞>∞∞50% were not identified (circumflex artery lesion in 2, mid-left anterior descending artery and diagonal lesions in 6, mid and distalright coronary artery stenosis in 4). In 4∞∞patients, a significant CAD was suspected by MSCT (2 proximal and 2∞∞mid-left anterior descending artery) despite the absence of significant narrowing on the coronary angiogram (2 proximal and 2∞∞mid-LAD). The sensitivity, specificity, positive and negative predictive values for detecting significant CAD were 78%, 88%, 91% and 71%, respectively.
Underlying causes of misinterpretation were analysed with regard to the presence of calcifications, motion artifacts and image contrast-to-noise ratio in the 70∞∞segments with false-negative and the 36∞∞segments with false-positive MSCT results. The presence of coronary calcifications was associated with the majority of false-positive findings (89%). Impaired image quality was another reason for false positive diagnosis. False negative results were due to the presence of calcifications (57%), image noise and suboptimal contrast enhancement (26%) or motion artifacts (17%). Overall, pitfalls were independent of vessel type. However, the right coronary artery and proximal-left anterior descending artery segments were the most frequent locations for false-positive findings, whereas diagonal
0353-07_ActaCardio_62_4_04_2881-07
342
06-08-2007
09:37
Pagina 342
L. Davin et al.
Fig. 3. – Example of concordance between MSCT and invasive coronary angiography. There is a significant lesion in the mid-segment of the right coronary artery (arrow).
Fig. 4. – Example of false negative result of MSCT due to calcifications. Suspicion of the presence of a significant lesion in the mid-segment of the right coronary artery (arrow) by MSCT, but presence of a non-significant lesion on the coronary angiogram.
branches and the circumflex artery were the most frequent locations for false-negative findings.
were examined. Accordingly, the overall accuracy of MSCT for identifying∞∞≥∞∞50% coronary stenosis was moderate and lower than previously reported. However, the negative predictive value to exclude significant coronary lesion (94%) remained high. Conversely, when the analysis was limited to confirm the diagnosis of significant CAD per patient, the predictive value of MSCT improved (91%) while its value to exclude a severe atherosclerosis remained moderate (71%). In routine practice, the accuracy of coronary anatomy assessment with 16-slice MSCT is therefore less precise than what has been reported by some authors. Some studies have previously reported similar observations with a less good sensitivity of the technique9. The spatial resolution inherent to the 16-slice methodology may limit the interpretation of small coronary arteries, notably the distal part of the coronary arteries. As shown in the present study, the diagnostic accuracy of MSCT is also improved when the analysis is limited to the proximal part of the coronary arteries. As complete visualization of the coronary tree cannot be anticipated with this methodology, further technical improvements are required until MSCT might challenge invasive coronary angiography as reference imaging modality in patients with CAD suspicion. Because the last generation of 64slice MSCT has a higher temporal and spatial resolution and allows shorter breath hold, the images achieved with this new technology may compete with those obtained by invasive coronary angiography17-19. Our data emphasize some of the previously recognized limitations of CT-scanner angiography for the quantification of coronary artery stenosis severity9,20. In fact, the main cause of misdiagnosis was related to the presence of vascular calcifications as well as to image distortion. These factors account for 89% of false positive results. Other factors, such as tachycardia, arrhythmias, inadequate breath holding or motion artifacts may also play a role, but these factors were not relevant in this study which included
Discussion Non-invasive visualisation of the coronary arteries is challenging. High spatial and high temporal resolution is needed to visualise small, tortuous, and rapidly moving coronary arteries12. This has become possible with the development of 16-slice CT angiography and initial reports have demonstrated that MSCT is able to detect coronary artery stenosis and allows for examination of coronary atherosclerotic plaques6-8,13-14. Observations from the present study confirm that 16slice MSCT scanner could provide good-quality noninvasive coronary arteriograms that accurately delineate the presence or absence of significant∞∞≥∞∞50% coronary stenosis in unselected patients with suspected CAD.
COMPARISON
WITH PREVIOUS STUDIES
Hitherto, most initial trials have limited their recruitment to selected patients with optimal image acquisition and quality, slow heart rate (eventually after beta blocker administration) who were trained to take breaths with minimal motion4,15. Moreover, some of these studies have assessed disease severity per patient basis, per vessel or per coronary segment. Comparison between those is thus uncertain and their results regarding sensitivity and specificity should be interpreted with caution. In addition, in some cases, up to 16% of segments have been excluded from the final analysis 16, which may dramatically change the likelihood calculations. By comparison, our cohort of patients was consecutive and not selected and many technical parameters or rules for patients’ preparation to get the best image quality were not considered12,15. In addition, all segments
0353-07_ActaCardio_62_4_04_2881-07
06-08-2007
09:37
Pagina 343
Diagnostic accuracy of MSCT in routine practice
patients in sinus rhythm without chronic obstructive pulmonary disease, half of them being treated by beta blockers.
PLACE
OF
MSCT IN
ROUTINE PRACTICE
The place of CT-scanner angiography in the algorithm of cardiac evaluation is still debated. Before using this technique routinely, it is imperative to precise if the coronary tree anatomy is the most important with an accuracy description and location of significant stenosis in coronary segments. In this case, the high negative predictive value of the test can be applied and at present there is a consensus to recommend MSCT evaluation in patients with a low pretest probability of CAD21. On the contrary, if the objective is detection of CAD in a patient without detailed description of coronary atherosclerosis, the good positive predictive value seems to be useful. Actually, the position of the added value of MSCT in the decisional tree for diagnosis of CAD is yet unknown. Other non-invasive tests such as stress echocardiography or nuclear imaging have proved their ability to exclude the presence of significant coronary disease. Sensitivity and specificity of MSCT would be evaluated compared with these techniques to precise the best place of MSCT in the cardiac evaluation algorithm. Moreover, there is no data about the prognostic value of MSCT. The poor positive predictive value indicates that coronary MSCT cannot be used as a screening technique. According to recent reports, it seems that patients who get the best advantages for MSCT are those in whom a pre-operative evaluation for non-cardiac surgery is required. This approach is also legitimate in patients with suspected CAD, if the stress test is inconclusive or cannot be performed. Other indications could be the evaluation of venous and arterial conduits in patients with previous bypass coronary surgery or imaging an anomalous course of one coronary artery22-24. The patients with a high pre-test probability of CAD should not be evaluated by MSCT. The invasive strategy will be required and a double exposition to X-ray radiation and iodinated contrast is not recommended.
Conclusions Non-invasive detection and evaluation of CAD remain a source of matter. Although its specificity is high, the sensitivity of 16-slice MDCT for detecting significant coronary artery stenosis in non-selected patients submitted to ICA is rather low suggesting that the diagnostic value of this technique should be first improved to be used in daily practice.
343
References 1. Achenbach S, Giesler T, Ropers D, Ulzheimer S, Derlien H, Schulte C, Wenkel E, Moshage W, Bautz W, Daniel WG, Kalender WA, Baum U. Detection of coronary artery stenoses by contrast-enhanced, retrospectively electrocardiographically-gated, multislice spiral computed tomography. Circulation 2001; 103: 2535-8. 2. Nieman K, Oudkerk M, Rensing BJ, van Ooijen P, Munne A, van Geuns RJ, de Feyter PJ. Coronary angiography with multi-slice computed tomography. Lancet 2001; 357: 599603. 3. Vogl TJ, Abolmaali ND, Diebold T, Engelmann K, Ay M, Dogan S, Wimmer-Greinecker G, Moritz A, Herzog C. Techniques for the detection of coronary atherosclerosis: multidetector row CT coronary angiography. Radiology 2002; 223: 212-20. 4. Becker CR, Knez A, Leber A, Treede H, Ohnesorge B, Schoepf UJ, Reiser MF. Detection of coronary artery stenoses with multislice helical CT angiography. J Comput Assist Tomogr 2002; 26: 750-5. 5. Kopp AF, Schroeder S, Kuettner A, Baumbach A, Georg C, Kuzo R, Heuschmid M, Ohnesorge B, Karsch KR, Claussen CD. Non-invasive coronary angiography with high resolution multidetector-row computed tomography. Results in 102∞∞patients. Eur Heart J 2002; 23: 1714-25. 6. Hoffmann U, Moselewski F, Cury RC, Ferencik M, Jang IK, Diaz LJ, Abbara S, Brady TJ, Achenbach S. Predictive value of 16-slice multidetector spiral computed tomography to detect significant obstructive coronary artery disease in patients at high risk for coronary artery disease: patient-versus segment-based analysis. Circulation 2004; 110: 2638-43. 7. Kuettner A, Beck T, Drosch T, Kettering K, Heuschmid M, Burgstahler C, Claussen CD, Kopp AF, Schroeder S. Diagnostic accuracy of noninvasive coronary imaging using 16detector slice spiral computed tomography with 188∞∞ms temporal resolution. J Am Coll Cardiol 2005; 45: 123-7. 8. Mollet NR, Cademartiri F, Krestin GP, McFadden EP, Arampatzis CA, Serruys PW, de Feyter PJ. Improved diagnostic accuracy with 16-row multi-slice computed tomography coronary angiography. J Am Coll Cardiol 2005; 45: 128-32. 9. Kaiser C, Bremerich J, Haller S, Brunner-La Rocca HP, Bongartz G, Pfisterer M, Buser P. Limited diagnostic yield of non-invasive coronary angiography by 16-slice multi-detector spiral computed tomography in routine patients referred for evaluation of coronary artery disease. Eur Heart J 2005; 26: 1987-92. 10. Aviram G, Finkelstein A, Herz I, Lessick J, Miller H, Graif M, Keren G. Clinical value of 16-slice multi-detector CT compared to invasive coronary angiography. Int J Cardiovasc Intervent 2005; 7: 21-8. 11. 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, American Heart Association. Circulation 1975; 51: 5-40. 12. Ferencik M, Moselewski F, Ropers D, Hoffmann U, Baum U, Anders K, Pomerantsev EV, Abbara S, Brady TJ, Achenbach S. Quantitative parameters of image quality in multidetector spiral computed tomographic coronary imaging with submillimeter collimation. Am J Cardiol 2003; 92: 1257-62.
0353-07_ActaCardio_62_4_04_2881-07
344
06-08-2007
09:37
Pagina 344
L. Davin et al.
13. Maruyama T, Yoshizumi T, Tamura R, Takashima S, Toyoshima H, Konishi I, Yamashita S, Yamasaki K. Comparison of eight- versus 16-slice multidetector-row computed tomography for visibility and image quality of coronary segments. Am J Cardiol 2004; 94: 1539-43. 14. Schroeder S, Kopp AF, Baumbach A, Meisner C, Kuettner A, Georg C, Ohnesorge B, Herdeg C, Claussen CD, Karsch KR. Noninvasive detection and evaluation of atherosclerotic coronary plaques with multislice computed tomography. J Am Coll Cardiol 2001; 37: 1430-5. 15. Schroeder S, Kopp AF, Kuettner A, Burgstahler C, Herdeg C, Heuschmid M, Baumbach A, Claussen CD, Karsch KR, Seipel L. Influence of heart rate on vessel visibility in noninvasive coronary angiography using new multislice computed tomography: experience in 94∞∞patients. Clin Imaging 2002; 26: 106-11. 16. Martuscelli E, Romagnoli A, D’Eliseo A, Razzini C, Tomassini M, Sperandio M, Simonetti G, Romeo F. Accuracy of thin-slice computed tomography in the detection of coronary stenoses. Eur Heart J 2004; 25: 1043-8. 17. Fine JJ, Hall PA, Richardson JH, Butterfield LO. 64-slice peripheral computed tomography angiography: a clinical accuracy evaluation. J Am Coll Cardiol 2006; 47: 1495-6. 18. Ropers D, Rixe J, Anders K, Kuttner A, Baum U, Bautz W, Daniel WG, Achenbach S. Usefulness of multidetector row spiral computed tomography with 64-∞∞≈∞ 0.6-mm collimation and 330-ms rotation for the noninvasive detection of significant coronary artery stenoses. Am J Cardiol 2006; 97: 3438. 19. Mollet NR, Cademartiri F, van Mieghem CA, Runza G, McFadden EP, Baks T, Serruys PW, Krestin GP, de Feyter
20.
21.
22.
23.
24.
PJ. High-resolution spiral computed tomography coronary angiography in patients referred for diagnostic conventional coronary angiography. Circulation 2005; 112: 2318-23. Cademartiri F, Runza G, Mollet NR, Luccichenti G, Belgrano M, Bartolotta TV, Galia M, Midiri M, Pozzi Mucelli R, Krestin GP. Impact of intravascular enhancement, heart rate, and calcium score on diagnostic accuracy in multislice computed tomography coronary angiography. Radiol Med (Torino) 2005; 110: 42-51. Budoff MJ, Achenbach S, Blumenthal RS, Carr JJ, Goldin JG, Greenland P, Guerci AD, Lima JA, Rader DJ, Rubin GD, Shaw LJ, Wiegers SE. Assessment of coronary artery disease by cardiac computed tomography: a scientific statement from the American Heart Association Committee on Cardiovascular Imaging and Intervention, Council on Cardiovascular Radiology and Intervention, and Committee on Cardiac Imaging, Council on Clinical Cardiology. Circulation 2006; 114: 1761-91. Nieman K, Pattynama PM, Rensing BJ, Van Geuns RJ, De Feyter PJ. Evaluation of patients after coronary artery bypass surgery: CT angiographic assessment of grafts and coronary arteries. Radiology 2003; 229: 749-56. Martuscelli E, Romagnoli A, D’Eliseo A, Tomassini M, Razzini C, Sperandio M, Simonetti G, Romeo F, Mehta JL. Evaluation of venous and arterial conduit patency by 16slice spiral computed tomography. Circulation 2004; 110: 3234-8. Deibler AR, Kuzo RS, Vohringer M, Page EE, Safford RE, Patron JN, Lane GE, Morin RL, Gerber TC. Imaging of congenital coronary anomalies with multislice computed tomography. Mayo Clin Proc 2004; 79: 1017-23.
