Diabetes mellitus (DM) has been widely considered a coronary artery disease (CAD) equivalent, which implies a 10-year cardiovascular risk of 20% for every diabetic patient [1]; however, recently published data refute this notion [2, 3]. According to medical convention, patients with DM are at a higher risk for developing CAD, despite lacking any pertinent symptoms [4]. Thus, it remains controversial whether asymptomatic diabetic patients should be screened for CAD and receive early optimized medical therapy even there is low coronary calcium score [5, 6].Previous studies have demonstrated that diabetic patients with initially asymptomatic CAD show accelerated plaque progression, despite the use of optimized medical treatment. Eventually, patients with DM develop more major adverse cardiac events (MACE) and evidence of permanent myocardial injury than patients without [7, 8]. Recent studies have highlighted the incremental increase in event rates associated with coronary artery calcification [9]. These investigations showed significantly more adverse outcomes in diabetic patients with a computed tomographic (CT) coronary artery calcium score (CACS) ≥10 than in patients without DM [10]. Thus, effective screening strategies for appropriate cardiovascular risk stratification of asymptomatic diabetic patients warrant further study.

Coronary CT angiography (CCTA) allows for the direct quantification and characterization of atherosclerotic plaque for potential risk evaluation beyond the assessment of stenosis [11‐13]. Furthermore, CCTA enables serial non-invasive monitoring of plaque progression [14‐16] and therefore may help to identify patients at risk for future adverse cardiovascular events [17, 18].

The aim of the present study was to monitor the progression of the atherosclerotic coronary plaque burden by serial cardiac CT and evaluate its relationship with cardiac events in asymptomatic diabetic patients.

A total of 225 patients who underwent CCTA were included. Eighteen patients were excluded due to poor image quality, resulting in a final study population of 197 patients. All patients had DM with mean disease duration of 5.0 ± 9.7 years.47/197 patients (23.9%) used insulin, and a history of diabetic complications was present in 11 (5.6%). Eighty-five patients (43.1%) had hyperlipidemia, and 113 (57.4%) had hypertension. Sixty patients (30.4%) had a smoking history. Statins were used by 89 patients (45.2%). The mean BMI was 25.7 ± 3.3 kg/m². Based on the pretest probability using updated Diamond-forrester methods, there was no difference in the distribution of pretest likelihood of CAD for both groups. Based on the CAD-RADS reporting guideline, the two groups showed insignificant difference in severity of coronary stenosis. Further patient demographics and baseline characteristics are presented in Table 1.

The median time interval between the first and second CCTA was 18.9 months (Q1-Q3, 16.4–32.5 months). The change of plaque characteristics in both patient groups between serial CCTA acquisitions is shown in Table 2. Patients with CACS≤10 showed a total mean plaque volume increase of 20.6 (±63.5) mm³ annually between CCTA scans, resulting in an increase of 2.2%. The CACS> 10 group demonstrated an average total plaque volume increase of 30.1 (±102.6) mm³ annually between acquisitions, corresponding to an increase of 3%. Comparing the increase of total plaque volume between each group showed a significantly larger increase in the CACS> 10 group (p = 0.002). There was no statistically significant difference regarding the percent change of atheroma volume between the two groups (p = 0.676), with calculated increases of 5.7% (±3.9) and 4.2% (±3.6) in CACS≤10 group and CACS> 10 group, respectively.

The CACS> 10 group showed a mean increase in dense calcium volume of 5.7 (±28.5) mm³ (23%) over baseline, whereas the dense calcium mean volume increased by 0.3 (± 13.6) mm³ in the group with CACS≤10, corresponding to an increase of 4.1%. The larger increase in dense calcium volume in the CACS> 10 group was statistically significant when compared to the CACS≤10 group (p = 0.029). Between acquisitions, the average volume of plaque with low “soft”, “lipid-rich” attenuation values (i.e., -100HU-30HU), increased by 14.3 (±34.9) mm³ in the CACS≤10 group, corresponding to an increase of 3.6% (±7.9) by percent atheroma volume. In patients with CACS> 10, the mean low attenuation plaque volume increased by 10.9 (±34.1) mm³, corresponding to an increase of 1.8% (±7.4) by percent atheroma volume. Significant differences between the groups regarding the change of mean low attenuation plaque volumes were observed (p = 0.018); however, there was no significant difference regarding the percent change of atheroma volume (p > 0.05).

The comparison of both groups showed a significant difference in the change of average total intermediate attenuation (i.e., >30HU - 350HU) “fibrotic” volume (p = 0.018), but no significant difference in percent change of atheroma volume (p > 0.05).

Our results showed that in our population of diabetics the change in calcium volume was inversely related to adverse cardiac outcomes. Additionally, the study results indicated that the low attenuation “lipid-rich”, soft plaque volume serves as an independent biomarker for predicting future cardiac events, progression of which was positively associated with adverse outcomes during follow-up. It has been argued that the plaque component exhibiting low-attenuation “lipid-rich” features as determined by CT to some extent represents the necrotic core of “vulnerable” atheromatous lesions [21]. Lastly, patients with a CACS> 10 included in the study demonstrated a higher cumulative event-free survival rate compared to patients with a CACS≤10.

The relationship between atherosclerotic plaque progression and cardiovascular risk in asymptomatic diabetic patients is insufficiently studied. Baseline CCTA demonstrated a moderate prevalence of coronary plaque burden (65%) in the study population, while patients with obstructive CAD at baseline were excluded from further analysis. During the follow-up CT scan, plaque quantification yielded different plaque progression patterns in the study subgroups. Specifically, using a median interval time of 18.9 months between CCTA acquisitions, total plaque volume increased by 20.6mm³ in the CACS≤10 group, compared to an increase of 30.1mm³ in the CACS> 10 group (p = 0.002). This statistically significant difference can be mostly attributed to the increase in total dense calcium volume (0.3mm³ vs. 5.7mm³, p = 0.029). Furthermore, low-attenuation plaque volume increased by 14.3mm³ in the CACS≤10 group, compared to an increase of 10.9 mm³ in the CACS> 10 group (p = 0.018). This significant increase in low-attenuation plaque volume in participants with CACS≤10 corresponded with a higher rate of adverse events in this patient group compared to subjects with CACS> 10. The more pronounced increase in low-attenuation plaque volume showed independent discriminatory power for the prediction of cardiac events on Cox regression analysis.

Several studies have evaluated the prognostic value of coronary artery calcium, stenosis severity, and cardiac events in diabetic patients. Results from these studies demonstrated an incremental increase in event rates corresponding with increasing calcium Agatston score and stenosis severity [9, 10, 22]. However, these studies failed to evaluate the association between calcification and plaque progression in the early disease stages of asymptomatic diabetics with non-obstructive CAD. Additionally, very few studies using serial CCTA for the evaluation of plaque progression in diabetic patients are available. Kim et al. demonstrated a 2.5-fold greater progression of total plaque volume during a mean scan interval of 3.8 years in diabetic vs. non-diabetic patient populations. More importantly, the progression of non-calcified plaque volume was significantly greater than that of calcified plaque volume [23]. The significant increase in total plaque volume and the increase of low-attenuation plaque material as a possible marker of necrotic core volume in the present study confirm these previous results.

A recent study focusing on the coronary plaque progression in diabetic patients showed that the baseline percentile of plaque volume and male sex were predictors of progression, but not the coronary calcification [23]. However, in present study the greater increase in low-attenuation plaque in the CACS≤10 group corresponded with a higher incidence of high-risk plaque features and eventual adverse cardiac events. The CACS> 10 group showed a significantly greater progression in total plaque volume, which was mainly due to the increase in total dense calcium. The discrepancy between higher calcium progression and lower cardiac event rates may be explained by plaque stabilization under statin therapy [24]. On a histo-pathologic level, asymptomatic patients with diabetes usually exhibit chronic endothelial damage and local factors that promote progressive atherosclerosis. These factors in conjunction with glycated protein oxidation could accelerate the calcification process, leading to spotty calcification associated with plaque vulnerability [25]. The present study illustrated a greater progression of spotty calcifications in patients with CACS≤10 than in patients with CACS> 10 while it appears conceivable that dense calcification may stabilize the plaque with a decrease in necrotic core burden and frequency of cardiac events.

Although CCTA can be a useful tool for risk stratification in diabetic patients with at least 5 years of asymptomatic CAD [26], it requires further investigation to determine whether the progression of plaque burden can be used for personalized treatment guidance and surveillance, especially in patients with high cardiovascular risk. Using serial CCTA acquisitions, our investigation demonstrated the feasibility of quantifying plaque progression and identified several imaging biomarkers that may have a positive or inverse association with future cardiac events in asymptomatic diabetics. Conceivably, repeat CCTA at an intermediate time-point after the initiation of optimized medical therapy may inform granular custom-tailoring of the treatment regimen based on the individual progression pattern of the coronary plaque burden. Although current Appropriate Use Criteria for Cardiac Computed Tomography do not incorporate prior CCTA results [27], the value of serial CCTA may warrant more consideration, especially for asymptomatic diabetic patients at high risk who may benefit from personalized risk stratification and modification of optimized medical therapy [28].

Several limitations to the present study deserve consideration. This is a single-center retrospective investigation with a patient population representing a single ethnicity, with potential bias due to selection and test power. It is possible that the prevalence and disease course of CAD could be influenced by the patient population’s ethnicity or other geographical factors. Furthermore, we did not systematically correlate our findings on coronary CTA with an invasive reference standard for intracoronary plaque assessment. Although the plaque software prototype has been validated in several studies [13, 29], there might be variation among different automated plaque quantification software. Thus, the same software should be used for serial assessment. Likewise, plaque quantification might be affected by technical influences, e.g. by the attenuation level of vascular contrast enhancement [30]. Moreover, the present study lacked of medication details in treating diabetes. New advanced treatment like incretin (GLP-1) therapy may play an important role in preventing plaque progression based on its pluripotent cardiovascular protective effects [31]. Also, the potential mechanism, for example the effect of genes’ polymorphism in secondary affecting the prognosis in the study population [32], needs to be further investigated. Therefore, these results may be applicable to very specific population, and prospective studies on larger study cohorts will be necessary to validate our findings.

This investigation demonstrates that quantification of plaque progression derived from non-invasive serial CCTA portends incremental prognostic value, with low-attenuation “lipid-rich” atheromatous plaque volume showing the highest predictive value. Meanwhile, an increase of dense coronary artery calcification is inversely correlated with the progression of low-attenuation plaque burden, arguably reflecting the role of the pathophysiology of calcification on plaque stabilization during coronary atherosclerosis in diabetes. Granular serial evaluation of atheromatous plaque progression patterns may benefit asymptomatic diabetic patients via personalized risk re-stratification and subsequent adjustment of optimized medical therapy.