Author + information
- Received December 3, 2010
- Revision received February 25, 2011
- Accepted February 28, 2011
- Published online July 1, 2011.
- Clerio F. de Azevedo, MD, PhD⁎,†,⁎ (, )
- Marcelo S. Hadlich, MD⁎,
- Sabrina G. Bezerra, MD⁎,
- João L. Petriz, MD⁎,
- Rogério R. Alves, MD, PhD†,
- Olga de Souza, MD, PhD⁎,
- Miguel Rati, MD⁎,
- Denilson C. Albuquerque, MD, PhD⁎,† and
- Jorge Moll, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Clerio F. de Azevedo, D’Or Institute for Research and Education, Rua Diniz Cordeiro, 30 - Rio de Janeiro, Rio de Janeiro 22281-100, Brazil
Objectives We attempted to determine the prognostic value of coronary computed tomographic angiography (CTA) in patients with inconclusive functional stress tests.
Background Patients with suspected coronary artery disease (CAD) and inconclusive noninvasive cardiac stress tests represent a frequent management challenge.
Methods We examined 529 consecutive patients with suspected CAD and prior inconclusive functional stress tests. All patients underwent a coronary CTA scan using a 64-slice multidetector row scanner. CAD severity by coronary CTA was categorized as: 1) no evidence of CAD; 2) nonobstructive coronary plaques (<30%); 3) mild stenosis (30% to 49%); 4) moderate stenosis (50% to 69%); and 5) severe stenosis (≥70%). Patients were also categorized according to a modified Duke prognostic CAD index. Survival analyses were performed using Cox proportional hazards models adjusted for baseline risk factors and coronary artery calcium score. The primary outcome of the study was the combined endpoint of all-cause mortality and nonfatal myocardial infarction.
Results Among patients with inconclusive stress tests, the large majority (69%) did not demonstrate significant CAD by coronary CTA. During a mean follow-up of 30.1 ± 11.1 months, there were 20 (3.8%) deaths and 17 (3.2%) nonfatal myocardial infarctions. Multivariable Cox regression analysis revealed that the presence of increasing degrees of obstructive CAD by CTA was an independent predictor of adverse events (hazard ratio [HR]: 1.66 [95% confidence interval (CI): 1.23 to 2.23], p = 0.001). Indeed, the presence of ≥50% coronary stenosis was associated with an increased risk of events (HR: 3.15 [95% CI: 1.26 to 7.89], p = 0.01). Likewise, the Duke prognostic CAD index was also found to be an independent predictor of events (HR: 1.54 [95% CI: 1.20 to 1.97], p = 0.001).
Conclusions Among patients with inconclusive functional stress tests, the noninvasive assessment of CAD severity by coronary CTA has been shown to provide incremental prognostic information beyond the evaluation of traditional risk factors and coronary artery calcium score.
Patients with low or intermediate pre-test probability of significant coronary artery disease (CAD) and inconclusive noninvasive cardiac stress tests represent a frequent management challenge. Although the post-test probability of CAD remains in the low or, more likely, the intermediate level in most of these patients, a large majority end up undergoing invasive coronary angiography (ICA) to exclude the presence of significant CAD. Indeed, a recent large-scale multicenter study demonstrated that, based on current algorithms used for the assessment of patients with suspected CAD, as many as 40% of all diagnostic ICA do not demonstrate any coronary artery obstruction and, therefore, do not lead to any therapeutic intervention (1). It is important to recognize, however, that the population included in that registry consisted of patients without known CAD who were undergoing elective catheterization, but not necessarily with a previous cardiac stress test.
In recent years, numerous single-center (2) and several multicenter studies (3–5) have shown that multidetector coronary computed tomographic angiography (CTA) is an accurate imaging modality for the identification or the exclusion of significant CAD. In addition, it has been demonstrated that the evaluation of CAD by coronary CTA provides incremental prognostic information over the assessment of traditional risk factors and coronary artery calcification (6–8). Furthermore, recent studies have shown that coronary CTA and functional stress tests provide complementary rather than overlapping diagnostic and prognostic information (9–11). Accordingly, the 2010 Appropriateness Criteria consensus document (12) defined the use of coronary CTA in patients with equivocal stress tests as an appropriate indication. Nonetheless, studies examining the role of coronary CTA specifically in this patient population are lacking. Therefore, in the present study, we attempted to determine the prevalence of significant CAD by coronary CTA in patients with inconclusive noninvasive cardiac stress tests. In addition, we sought to determine whether the use of coronary CTA in this patient population would provide incremental prognostic information for the prediction of major adverse cardiac events.
From May 2006 to October 2008, we evaluated 529 consecutive patients who were physician-referred for investigation of suspected CAD. All patients had a previous inconclusive noninvasive cardiac stress test performed within the preceding 3 months. Stress tests included exercise electrocardiography (ECG) (n = 230), myocardial perfusion imaging (MPI) using single-photon emission tomography (n = 270), and stress echocardiography (n = 29). Imaging studies were performed with exercise stress in 178 patients (60%) and with pharmacological stress in 121 patients (40%).
Cardiac stress tests were defined as inconclusive when the results were considered to be unreliable and/or when the post-test probability of significant CAD remained at the intermediate level after testing. So, among patients that underwent only an exercise ECG, inconclusive results were due to a submaximal stress test (patients' incapacity to reach at least 85% of the age-predicted maximum heart rate) or due to a discordance between pre-test likelihood of CAD and exercise ECG results (e.g., patients with a low pre-test probability and a positive exercise ECG, or patients with a high pre-test probability and a negative exercise ECG). Among patients that underwent imaging stress tests, inconclusive results were due to submaximal stress tests or to the presence of artifacts and/or technical difficulties, such as attenuation artifacts in the case of MPI and poor acoustic windows in the case of stress echocardiography. In addition, imaging studies were also considered inconclusive if there was discordance between exercise ECG and imaging results, or discordance between pre-test likelihood of CAD and MPI or stress echocardiography results (e.g., patients with a low pre-test probability and a positive imaging study, or patients with a high pre-test probability and a normal imaging study). Functional stress tests exhibiting extensive abnormalities suggestive of a high risk of cardiac events (13) were not considered inconclusive.
Patients were excluded if found to have an irregular heart rate, documented history of allergy to iodinated contrast, or impaired renal function. Additionally, patients with known prior CAD, including previous myocardial infarction or revascularization, were also excluded. The study was approved by the Institutional Ethics Committee, and all patients gave written informed consent.
Before coronary CTA examination, the presence of the following cardiovascular risk factors was prospectively assessed: 1) diabetes mellitus (previous diagnosis by a physician and/or use of insulin or oral hypoglycemic agents); 2) dyslipidemia (known but untreated dyslipidemia or current treatment with lipid-lowering medication); 3) hypertension (documented history of high blood pressure or treatment with antihypertensive medication); 4) smoking (current smoking or cessation of smoking within 3 months); and 5) history of premature CAD in a first-degree relative (men, age <55 years; women, age <65 years). Additionally, pre-test likelihood of significant CAD was determined based on American College of Cardiology/American Heart Association previously published criteria (13,14).
Scan protocol and image reconstruction
All coronary CTA scans were performed with a 64-slice multidetector row scanner (Brilliance 64, Philips Medical Systems, Best, the Netherlands). Individuals presenting with baseline heart rates >65 beats/min were given 5 mg of intravenous metoprolol at 5-min intervals to a total possible dose of 20 mg to achieve a resting heart rate <65 beats/min.
Before coronary CTA acquisition, a nonenhanced, prospectively ECG-triggered scan was performed to measure the coronary artery calcium score (40 × 0.625 mm collimation, 2.5-mm slice thickness, 420-ms rotation time, tube voltage 120 kV, and tube current 55 to 100 mAs). During the CTA acquisition, 60 ml to 130 ml of iodinated contrast (Xenetix 350, Guerbet, Aulnay sous Bois, France) was injected followed by a 30-ml saline flush. A bolus tracking technique was used to synchronize the start of image acquisition with the arrival of contrast agent in the coronary arteries. The retrospectively ECG-gated, helical contrast-enhanced coronary CTA scan was performed with the following parameters: 64 × 0.625 mm collimation, 420-ms rotation time, tube voltage 120 kV, and tube current 800 to 1,000 mAs. Radiation reduction algorithms employing ECG-based tube current modulation were used, and estimated radiation doses ranged from 8 to 17 mSv (dose-length products ranged from 565 to 1,224 mGy · cm).
Coronary CTA images were reconstructed immediately after scan completion in a consistent manner in order to identify motion-free coronary artery images. ECG-gated datasets were reconstructed at 40%, 50%, 70%, and 80% of the cardiac cycle. In case of insufficient image quality, additional phases were reconstructed at 5% increments. Multiple phases were used for image interpretation if the phase of minimum coronary artery motion was different for different arteries.
All scans were analyzed independently by 2 cardiologists with experience interpreting several thousand coronary CTA scans. All analyses were performed on a dedicated 3-dimensional workstation (Brilliance, Philips Medical Systems). The coronary artery calcium score was calculated using the Agatston method. For the assessment of coronary CTA images, the readers were allowed to use any of the available post-processing algorithms. In case of discordant results, the final diagnosis was based on the consensus opinion of both readers.
Coronary arteries were scored using a modified 16-segment American Heart Association coronary artery classification, as previously described (15). In all individuals, irrespective of image quality, every arterial segment was scored. If a coronary artery segment was considered uninterpretable, the unevaluable segment was scored similarly to the most proximal evaluable segment. Coronary atherosclerotic lesions were evaluated semiquantitatively by visual estimation. CAD severity was graded as: 1) no CAD; 2) nonobstructive coronary plaques (<30% obstruction); 3) mild stenosis (30% to 49% obstruction); 4) moderate stenosis (50% to 69% obstruction); and 5) severe stenosis (≥70% obstruction) (Fig. 1). Patients were also classified according to a modified Duke prognostic CAD index (6,16): 1: <50% stenosis; 2: ≥2 mild stenoses (30% to 49% obstruction) including 1 artery with proximal disease, or 1 moderate stenosis (50% to 69% obstruction); 3: 2 moderate stenoses or 1 severe stenosis (≥70% stenosis); 4: 3 moderate stenoses, or 2 severe stenoses, or 1 proximal left anterior descending artery severe stenosis; 5: 3 severe stenoses or 2 severe stenoses including 1 proximal left anterior descending artery stenosis; and 6: >50% stenosis in the left main coronary artery
Epidemiological methods for follow-up included ascertainment of events blinded to coronary CTA results. Clinical information was obtained from telephone interviews, medical records, or by contacting patients' physicians. The primary outcome of the study was the combined endpoint of all-cause mortality and nonfatal myocardial infarction. Nonfatal infarction was defined based on the criteria of typical chest pain, elevated cardiac enzyme levels, and typical ECG changes.
All continuous variables are presented as mean ± SD, and all categorical variables are reported as a percentage or absolute number. Continuous variables were compared using Student t test for independent samples or analysis of variance tests. The normal distribution of continuous variables was tested and confirmed using the Shapiro-Wilk test. Categorical variables were compared using the chi-square statistic. The Kaplan-Meier technique was used to evaluate survival times and the log-rank test was used to compare survival curves. Cox proportional hazards models were used to examine the effects of several continuous and categorical predictors of adverse events. First, univariable analysis of baseline characteristics and coronary CTA results was performed. Then, multivariable analysis with stepwise selection was used to determine independent predictors of events after adjusting for baseline cardiac risk factors and coronary artery calcium score. Multivariable models were limited to no more than 4 variables to minimize model overfitting. In order to further evaluate the incremental prognostic value of coronary CTA, receiver-operator characteristic (ROC) analysis was performed and the area under the curve (AUC) was calculated and reported with 95% confidence intervals (CI). Analyses were performed with Stata software, version 10.0 (StataCorp, College Station, Texas). All tests were 2-tailed, and a value of p < 0.05 was considered indicative of statistical significance.
Overall, the study population consisted of 529 patients; 61% were men and the mean age was 58 ± 12 years (range: 22 to 87 years). Coronary CTA results revealed that, among this cohort of patients with prior inconclusive functional stress tests, the large majority did not demonstrate significant CAD (Fig. 2). Overall, 200 patients (38%) did not exhibit any evidence of CAD, 100 (19%) had nonobstructive plaques (<30% obstruction), and 63 (12%) demonstrated mild stenosis (30% to 49% obstruction). Only 62 patients (12%) exhibited moderate stenosis (50% to 69% obstruction), and 104 (20%) displayed at least 1 vessel with severe stenosis (≥70% obstruction). When patients were categorized according to the Duke prognostic CAD index, 320 patients (60%) were classified in the Duke 1 subcategory, 74 (14%) in Duke 2, 52 (10%) in Duke 3, 41 (8%) in Duke 4, 34 (6%) in Duke 5, and 8 (2%) in the Duke 6 subcategory.
Baseline clinical characteristics are categorized according to the severity of coronary CTA diagnosed CAD (Table 1). Patients with increasing severity of obstructive CAD tended to be older (p < 0.001), male (p < 0.001), hypertensive (p < 0.001), diabetic (p < 0.001), and hyperlipidemic (p < 0.001). Also, individuals with higher pre-test likelihood of significant CAD exhibited increasing degrees of CAD severity by coronary CTA (p = 0.02). The coronary artery calcium score also increased with severity of coronary CTA diagnosed CAD (p < 0.001). The results of the noninvasive cardiac stress tests are summarized in Table 2.
During a mean follow-up of 30.1 ± 11.1 months, there were 20 (3.8%) deaths and 17 (3.2%) nonfatal myocardial infarctions. The baseline clinical and coronary CTA characteristics of patients with and without events are summarized in Table 3. Thirty-seven patients (7.0%) were lost during follow-up (31 men, age 53.4 ± 13.1 years). A total of 68 patients underwent revascularization 81 ± 176 days after the coronary CTA study: none (0%) with normal coronary arteries, 2 patients (1%) with <50% coronary obstruction, and 66 patients (40%) with ≥50% coronary stenosis
Clinical and coronary CTA predictors of adverse events: univariable analyses
The results of univariable Cox regression analyses are summarized in Tables 4 and 5.⇓⇓ Among baseline clinical characteristics, older age (hazard ratio [HR]: 1.06 [95% CI: 1.03 to 1.09], p < 0.001), and the presence of diabetes (HR: 2.74 [95% CI: 1.32 to 5.65], p = 0.007) were associated with increased risk for adverse events. The other traditional cardiovascular risk factors for CAD were not significantly associated with increased risk of events. The presence of an increased coronary artery calcium score was also an important predictor of adverse events on univariable analysis (HR: 1.88 [95% CI: 1.40 to 2.51], p < 0.001). Importantly, patients with increasing degrees of obstructive CAD by coronary CTA demonstrated significantly higher risk for adverse events (HR: 1.82 [95% CI: 1.44 to 2.30], p < 0.001). Likewise, patients with increasing CAD severity by the Duke prognostic CAD index also exhibited significantly higher risk of events during follow-up (HR: 1.68 [95% CI: 1.40 to 2.02], p < 0.001). When patients were partitioned into subgroups with <50% and ≥50% stenosis, the presence of more obstructive CAD was also found to be a strong predictor of adverse events (HR: 5.27 [95% CI: 2.60 to 10.67], p < 0.001)
Coronary CTA predictors of adverse events: multivariable analyses
In the first multivariable Cox regression analysis, after adjusting for age, diabetes, and coronary artery calcium score, the presence of increasing degrees of obstructive CAD was found to be a significant predictor of adverse events (HR: 1.66 [95% CI: 1.23 to 2.23], p = 0.001) (Table 5). Kaplan-Meier analyses revealed that the event-free survival probability was 98% for patients with no evidence of CAD, 93% for nonobstructive plaques, and 89% for mild, 84% for moderate, and 82% for those with severe stenosis (log-rank test chi-square = 31.8, p < 0.0001) (Fig. 3). In the second multivariable Cox regression analysis, after adjusting for the same covariates, the Duke prognostic CAD index was also found to be a significant predictor of events (HR: 1.54 [95% CI: 1.20 to 1.97], p = 0.001) (Table 5). The Kaplan-Meier event-free survival probabilities were 95%, 92%, 84%, 82%, 74%, and 73% for patients in Duke subcategories 1 to 6, respectively (log-rank test chi-square = 38.0, p < 0.0001) (Fig. 4). Finally, in the third multivariable Cox regression analysis, also adjusted for the same covariates, the presence of ≥50% coronary stenosis was a strong predictor of adverse events (HR: 3.15 [95% CI: 1.26 to 7.89], p = 0.01) (Table 5). The event-free survival probability was 95% for patients without and 83% for patients with at least 1 coronary lesion with ≥50% obstruction (log-rank test chi-square = 26.8, p < 0.0001) (Fig. 5).
Incremental prognostic value of coronary CTA
In order to further examine the incremental prognostic value of coronary CTA, beyond the assessment of coronary artery calcification, we performed a ROC analysis and compared the areas under the curve (AUC) for coronary artery calcium score alone and for the association of coronary artery calcium score and coronary CTA as predictors of adverse events. The AUC for the assessment of coronary artery calcification alone was 0.72 (95% CI: 0.64 to 0.79). When we added the information provided by the assessment of obstructive CAD severity coronary CTA, there was a trend towards a higher prognostic power for the combined analysis (AUC: 0.77 [95% CI: 0.70 to 0.83], p = 0.06) (Fig. 6A). Likewise, when we added the information provided by coronary CTA as the Duke prognostic CAD index, there was a significant increase in the capacity of the model to predict adverse events (AUC: 0.78 [95% CI: 0.71 to 0.85], p = 0.03) (Fig. 6B).
In the present study, we were able to demonstrate that, among patients with suspected CAD and inconclusive noninvasive cardiac stress tests, the large majority does not demonstrate significant CAD when evaluated by coronary CTA. Less than one-third exhibited coronary lesions with ≥50% luminal obstruction and only one-fifth exhibited lesions with ≥70% obstruction. Furthermore, this study demonstrated that, in this patient population, the presence of increasing degrees of CAD severity by coronary CTA is associated with an adverse intermediate-term cardiac prognosis. Importantly, patients without significant CAD by coronary CTA were found to have a low risk of adverse events during follow-up.
As previously mentioned, the application of current investigation algorithms in “real-world” clinical practice results in a large number of ICAs that do not demonstrate any significant coronary obstruction and, thus, do not lead to any therapeutic intervention (1). The results of 2 recent studies suggest that the strategy of using coronary CTA as a gatekeeper to ICA could be cost saving in patients with equivocal noninvasive stress tests (17,18). Accordingly, the ACCF/SCCT/ACR/AHA/ASE/ASNC/SCAI/SCMR 2010 Appropriate Use Criteria for Cardiac Computed Tomography (12) defined the use of coronary CTA in patients with equivocal or discordant stress tests as an appropriate indication. It is important to recognize, however, that there is a paucity of data examining the role of coronary CTA specifically in this patient population.
Several previous studies have investigated the prognostic value of coronary CTA in patients with suspected CAD, but not necessarily with prior inconclusive stress tests (6–9,19). Using a 16-slice computed tomography scanner, Min et al. (6) have reported on the relationship between all-cause mortality and coronary CTA results in 1,127 patients with suspected CAD. In agreement with our results, they found that the modified Duke prognostic CAD index was a strong independent predictor of all-cause mortality. Similarly, Hadamitzky et al. (7) followed 2,223 patients with suspected CAD for a median of 2.3 years and demonstrated that the presence of significant CAD was a strong predictor of adverse cardiac events. In another important study, Ostrom et al. (8) examined 2,538 patients by coronary CTA using electron-beam tomography and, after a mean follow-up of 6.5 years, also found that the extent of CAD was a significant predictor of all-cause mortality. In these studies, the prognostic value of coronary CTA was incremental to traditional risk factors and coronary artery calcium scoring, which is in agreement with our findings. In a recent study, van Werkhoven et al. (9) evaluated the prognostic value of both coronary CTA and MPI in 439 patients and demonstrated that the presence of significant CAD added independent prognostic information to baseline risk factors and MPI results.
It is important to note that, in the present study, the proportion of patients with low pre-test likelihood of significant CAD (62%) was higher than in these previous reports (18% to 32%) (6,7,9,19). Nevertheless, we believe this difference was counterbalanced by the fact that all patients in the present study had a prior functional stress test that was either equivocal or inconclusive. Indeed, the proportion of patients with significant obstructive CAD in this study (31%) was comparable to the proportion observed in these previous reports (18% to 56%) (6–9,19).
There are a few previous studies that began to investigate the role of coronary CTA in patients with prior nondiagnostic stress tests (10,20,21). In a recent retrospective study, Chan et al. (21) evaluated 151 patients with equivocal stress tests and demonstrated that the majority (71%) did not exhibit ≥50% obstructive CAD, which is very close to the rate observed in the present study (69%). In that study, after a mean follow-up of 1.4 years, none of the patients without significant CAD required revascularization, whereas among patients with ≥50% obstruction, 34% underwent revascularization procedures. Similarly, in a prospective study that included 56 patients with equivocal and 81 patients with abnormal cardiac stress tests, Abidov et al. (10) also found a significant relationship between the severity of coronary CTA–diagnosed CAD and the rate of revascularization interventions during follow-up. Although these previous reports are in agreement with our results, it is important to highlight that the present study was the first to demonstrate the incremental prognostic value of coronary CTA as a predictor of major adverse events in patients with prior inconclusive noninvasive stress tests.
This is a single-center observational study of a cohort of patients that were physician-referred for coronary CTA. Decisions regarding further investigation and treatment strategies were defined by the patients' physicians. In fact, the data of the present study represent actual clinical practice patterns from numerous physicians referring patients with inconclusive stress tests for evaluation by coronary CTA. Therefore, it is not possible to determine whether event-free survivors fared better because of more optimal treatment and/or nonsurvivors fared poorly because of inappropriate treatment.
In the present study, all-cause mortality was included as a primary endpoint. We decided not to use cardiac death as an endpoint because we were unable to ascertain the cause of death in some patients. The disadvantage of using all-cause mortality is that some of the events included in our analyses were unrelated to atherosclerotic disease. The advantage is that the bias resulting from cause of death misclassification did not occur in the present model. Nevertheless, it is important to note that, in this type of patient cohort, cardiovascular disease remains the leading cause of death, accounting for approximately one-half of all fatalities. It is also important to highlight that, due to the relatively small number of events (20 deaths and 17 nonfatal myocardial infarctions), there is a degree of overfitting in our multivariable Cox regression models.
Another potential limitation is related to the issue of post-test referral bias. In the present study, the coronary CTA results probably had an influence on patient management, particularly referral to revascularization. Because revascularization also affects risk, the association between coronary CTA results and revascularization referral might have introduced a bias that lowered patients' risk in proportion to the severity of coronary CTA results. Despite this potential confounder, we decided not to censor from prognostic analyses any patient revascularized after the coronary CTA study. The selective removal of patients with higher CAD severity would have resulted in a relative underestimation of risk and a flattening of the coronary CTA abnormality–risk relationship in proportion to revascularization referral rates. It is important to note, however, that this post-test referral bias may have had the effect of weakening the prognostic value of coronary CTA. Therefore, without this confounder, the prognostic power of coronary CTA documented by the present study might have been even stronger.
Conclusions and Clinical Implications
In the present study, we demonstrated that, among patients with suspected CAD and inconclusive cardiac stress tests, the majority does not exhibit significant obstructive CAD when examined by coronary CTA. Most importantly, the noninvasive assessment of CAD severity by coronary CTA has been shown to provide incremental prognostic information beyond the evaluation of traditional risk factors and coronary artery calcium score. These results, taken in combination, highlight the potential of coronary CTA to represent a useful imaging modality that could help optimize investigation algorithms used in patients with suspected CAD and, therefore, avoid a significant number of invasive coronary angiographies that would not have led to any therapeutic intervention.
The study was partially supported by FAPERJ (Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro) grant E-26/190.2666/2008. The authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- area under the curve
- coronary artery disease
- confidence interval
- computed tomographic angiography
- hazard ratio
- invasive coronary angiography
- myocardial perfusion imaging
- receiver-operator characteristic
- Received December 3, 2010.
- Revision received February 25, 2011.
- Accepted February 28, 2011.
- American College of Cardiology Foundation
- Hamon M.,
- Biondi-Zoccai G.G.,
- Malagutti P.,
- Agostoni P.,
- Morello R.,
- Valgimigli M.
- Budoff M.J.,
- Dowe D.,
- Jollis J.G.,
- et al.
- Meijboom W.B.,
- Meijs M.F.,
- Schuijf J.D.,
- et al.
- Min J.K.,
- Shaw L.J.,
- Devereux R.B.,
- et al.
- Hadamitzky M.,
- Distler R.,
- Meyer T.,
- et al.
- Ostrom M.P.,
- Gopal A.,
- Ahmadi N.,
- et al.
- van Werkhoven J.M.,
- Schuijf J.D.,
- Gaemperli O.,
- et al.
- Abidov A.,
- Gallagher M.J.,
- Chinnaiyan K.M.,
- Mehta L.S.,
- Wegner J.H.,
- Raff G.L.
- Berman D.S.,
- Hachamovitch R.,
- Shaw L.J.,
- et al.
- Taylor A.J.,
- Cerqueira M.,
- Hodgson J.M.,
- et al.
- Gibbons R.J.,
- Abrams J.,
- Chatterjee K.,
- et al.
- Gibbons R.J.,
- Balady G.J.,
- Bricker J.T.,
- et al.
- Menon M.,
- Lesser J.R.,
- Hara H.,
- et al.
- Hadamitzky M.,
- Freissmuth B.,
- Meyer T.,
- et al.
- Danciu S.C.,
- Herrera C.J.,
- Stecy P.J.,
- Carell E.,
- Saltiel F.,
- Hines J.L.