As the prevalence of diabetes continues to rise rapidly in Korea, diabetes and its many complications have become a major contributor to increased morbidity and mortality. According to Kim et al. [1
], diabetes is the fifth-leading cause of death in Korea, with a two-fold increased death rate among diabetics compared to nondiabetics; moreover, diabetes-related mortality was found to be 19.6 per 100,000 persons in 2009. It is well known that cardiovascular disease is the most common cause of death (30.6%); therefore, the need for appropriate diagnosis and cost-effective management of CAD has become a rising concern. Although the gold standard for CAD detection is known to be coronary angiography, there are several limitations to this procedure: 1) invasiveness, 2) expense, 3) small but definite risk of complications, and 4) the need for a skillful operator and special facilities. Given these limitations, multiple noninvasive methods have been developed and are now available. These can be divided into functional and anatomical imaging. Functional imaging examines the hemodynamic changes, whereas anatomical imaging permits direct visualization of the coronary arteries [6
]. There are widely used functional imaging techniques to detect CAD such as positron emission tomography or single photon emission computed tomography and myocardial contrast or stress echocardiography [6
]. These tests are useful in defining ischemic territories and predicting myocardium reversibility, but they have several limitations: That methods have some risks because of using pharmacologic stress inducing perfusion defects manner. Moreover, they do not permit direct visualization of the coronary arteries. Specificity of such tests tends to be lower, thereby occasionally requiring other tests for further evaluation.
A 64-slice MDCT is a newer, more widely available method to assess coronary artery disease more directly irrespective of symptoms [9
]. Previous studies in the general population reported MDCT sensitivity and specificity as 95% to 99% (95% credible interval [Crl]) and 83% to 94% (95% Crl), respectively [14
]. Miller et al. [15
] provided a diagnostic performance of a 64-slice MDCT with 291 symptomatic patients (23% of whom have diabetes), and its sensitivity, specificity, PPV, and NPV were 85%, 90%, 91%, and 83%, respectively in a patient-based analysis. They concluded that although MDCT cannot be used as a replacement for ICA, it may help guide clinician decisions as to whether a patient needs further invasive studies [16
]. In our study, 64-slice MDCT demonstrated high quality noninvasive coronary angiograms that accurately detected significant lesions of symptomatic patients with type 2 diabetes. Only a few studies on 64-slice MDCT performance in diabetic patients have been done; one of them by Schuijf et al. [17
] reported that MDCT is a feasible noninvasive test in type 2 diabetes patients with both a 95% sensitivity and specificity of interpretable segments (220/256, 86%). When uninterpretable segments were included, sensitivity and specificity dropped to 81% and 82%, respectively. The present study shows relatively higher quality image (93.4% above adequate image), with 98.4% sensitivity and 100% specificity for diabetic patients regardless of imaging quality. This is consistent with meta-analyses results, which showed that the sensitivity of a 64-slice CT (98%) was significantly higher than that of a 16-slice CT (95%) in a patient-based analysis [18
]. Because a 64-slice MDCT has increased slices per gantry rotation (64 vs. 16) and a faster gantry speed (330 vs. 375 ms/rotation), the quality of image is increased; therefore, we were able to identify more accurate images compared with former studies. At the bedside, these results may have important clinical implications. Patients with diabetes can be delay proper diagnosis and management because they present atypical chest pain often confused with gastrointestinal or pulmonary symptoms, so high specificity and NPV of MDCT could be useful in reducing unnecessary invasive studies. MDCT could be also beneficial in the preoperative assessment of a noncardiac surgery, rapid triage in an emergency center or outpatient department, and evaluation of equivocal stress test results. Therefore, the risks and costs of invasive coronary angiogram could be avoided in a substantial number of diabetic patients by ruling out significant CAD with MDCT.
It is well known that patients with type 2 diabetes experience more diffuse, calcified, and extensive CAD, with significantly higher coronary calcium scores compared to controls. Calcified atherosclerosis has been identified as the main reason for poor image quality (46%, included multiple factors). However, there were no significant differences between sensitivity and specificity in both groups, probably due to small voxel size effect. Calcium deposits form a metal density that overwhelms the density of other tissues (calcium blooming effect), leading to misinterpreted adjacent plaques through artifacts attenuating low-energy X-rays (beam-hardening artifacts). Because small voxel size reduces and modifies these artifacts [19
], 64-slice MDCT ameliorates imaging challenges and is helpful in detecting CAD regardless of high coronary calcium scores in diabetic groups.
Raff et al. [19
] pointed out that factors related to poor image quality included not only calcium scores over 400, but also obesity (BMI ≥30 kg/m2
) and heart rates over 70 beats/min. In a recent study, Dewey et al. [20
] also reported that increased BMI, heart rate, and the presence of breathing artifact were associated with worse image quality.
Our study found that the aforementioned factors are not associated with imaging quality and diagnostic accuracy. Because BMI is a distinctive characteristic in this study compared with Caucasian-based research, we suggest that MDCT may be favorable in far-east Asians including Koreans. Asia-Oceanian use different BMI cutoffs for overweight (23 kg/m2
) and obese (25 kg/m2
) patients compared to Western measures of BMI and commonly tend to have a more lean body shape than Caucasians [13
]. Koreans generally have a higher total body fat content and higher abdominal fat distribution than Caucasians. It is well known that they develop obesity-related complications at a lower BMI level [21
]. This point implies that previous surveys about the applicability of MDCT for Caucasians cannot be applied to all ethnicities. In fact, because obesity is common in Western patients with diabetes, adequate adjustments such as scanner settings of kV and mA must be considered [20
]. Therefore, we believe that the ethnic differences in obesity, for example, are the reason for different image quality and accuracy. Dewey et al. [20
] showed that Asians have a tendency for better imaging quality compared to Caucasians (odds ratio, 2.68; P
=0.06) and significantly better quality than African Americans. Therefore, Koreans may be in a more favorable group with lesser variables affected by body size in comparison to Caucasians.
It is widely accepted that the accuracy of distal segments is lower than that of proximal segments because less blood flow due to small diameters reduces contrast-to-noise, making it difficult to interpret peripheral lesions [22
]. This study was consistent with prior results. The sensitivity of 2nd diagnostic and posterior descending artery (PDA) branches dropped sharply to 25% and 33.3%, respectively. Namgung et al. [27
] had reported conflicting results with higher sensitivity in distal segments (95%) compared to proximal segments (89%), as well as a greater degree of calcification and motion artifact at the proximal segment. According to Yun et al. [28
], diameters of the distal segment of left anterior descending artery and PDA of right coronary artery were 2.16±0.39 mm (median, 2.09 mm) and 2.09±0.48 mm (median, 2.02 mm), respectively for Koreans. Although it is believed that coronary plaques are more prevalent in the proximal segments of each vessel regardless of diabetes [29
], diabetic patients often present with equivocal symptoms that could lead to physician misdiagnosis. Additionally, this raises the concern that perhaps radiologists should be more thorough in interpreting peripheral lesions and selecting patients for further study.
Our study has several limitations. Given this is a cross-sectional, single-center study, the diagnostic accuracy of MDCT may not represent the true accuracy in all Koreans with diabetes. Since including all patients had chest pain possible to influence the results, thus selection bias may be not excluded in this study. Limitations of coronary 64-slice MDCT include radiation-exposure [30
], the use of contrast, the need for heart rate control, and anatomical imaging not directly indicative of ischemic changes.
In conclusion, our results demonstrate that there is no significant difference in accuracy between invasive coronary angiogram and MDCT regardless of diabetes. Thus, MDCT may be helpful in reducing unnecessary invasive studies for diabetic patients with chest pain, and may facilitate more reliable decisions for physicians in Korea.