Risk factors based vessel-specific prediction for stages of coronary artery disease using Bayesian quantile regression machine learning method: Results from the PARADIGM registry.

BACKGROUND AND HYPOTHESIS: The recently introduced Bayesian quantile regression (BQR) machine-learning method enables comprehensive analyzing the relationship among complex clinical variables. We analyzed the relationship between multiple cardiovascular (CV) risk factors and different stages of coronary artery disease (CAD) using the BQR model in a vessel-specific manner. METHODS: From the data of 1,463 patients obtained from the PARADIGM (NCT02803411) registry, we analyzed the lumen diameter stenosis (DS) of the three vessels: left anterior descending (LAD), left circumflex (LCx), and right coronary artery (RCA). Two models for predicting DS and DS changes were developed. Baseline CV risk factors, symptoms, and laboratory test results were used as the inputs. The conditional 10%, 25%, 50%, 75%, and 90% quantile functions of the maximum DS and DS change of the three vessels were estimated using the BQR model. RESULTS: The 90th percentiles of the DS of the three vessels and their maximum DS change were 41%-50% and 5.6%-7.3%, respectively. Typical anginal symptoms were associated with the highest quantile (90%) of DS in the LAD; diabetes with higher quantiles (75% and 90%) of DS in the LCx; dyslipidemia with the highest quantile (90%) of DS in the RCA; and shortness of breath showed some association with the LCx and RCA. Interestingly, High-density lipoprotein cholesterol showed a dynamic association along DS change in the per-patient analysis. CONCLUSIONS: This study demonstrates the clinical utility of the BQR model for evaluating the comprehensive relationship between risk factors and baseline-grade CAD and its progression.

Effects of chronic kidney disease and declining renal function on coronary atherosclerotic plaque progression: a PARADIGM substudy.

AIMS: To investigate the change in atherosclerotic plaque volume in patients with chronic kidney disease (CKD) and declining renal function, using coronary computed tomography angiography (CCTA). METHODS AND RESULTS: In total, 891 participants with analysable serial CCTA and available glomerular filtration rate (GFR, derived using Cockcroft-Gault formulae) at baseline (CCTA 1) and follow-up (CCTA 2) were included. CKD was defined as GFR <60 mL/min/1.73 m2. Declining renal function was defined as ≥10% drop in GFR from the baseline. Quantitative assessment of plaque volume and composition were performed on both scans. There were 203 participants with CKD and 688 without CKD. CKD was associated with higher baseline total plaque volume, but similar plaque progression, measured by crude (57.5 ± 3.4 vs. 65.9 ± 7.7 mm3/year, P = 0.28) or annualized (17.3 ± 1.0 vs. 19.9 ± 2.0 mm3/year, P = 0.25) change in total plaque volume. There were 709 participants with stable GFR and 182 with declining GFR. Declining renal function was independently associated with plaque progression, with higher crude (54.1 ± 3.2 vs. 80.2 ± 9.0 mm3/year, P < 0.01) or annualized (16.4 ± 0.9 vs. 23.9 ± 2.6 mm3/year, P < 0.01) increase in total plaque volume. In CKD, plaque progression was driven by calcified plaques whereas in patients with declining renal function, it was driven by non-calcified plaques. CONCLUSION: Decline in renal function was associated with more rapid plaque progression, whereas the presence of CKD was not.

Per-lesion versus per-patient analysis of coronary artery disease in predicting the development of obstructive lesions: the Progression of AtheRosclerotic PlAque DetermIned by Computed TmoGraphic Angiography Imaging (PARADIGM) study.

To determine whether the assessment of individual plaques is superior in predicting the progression to obstructive coronary artery disease (CAD) on serial coronary computed tomography angiography (CCTA) than per-patient assessment. From a multinational registry of 2252 patients who underwent serial CCTA at a ≥ 2-year inter-scan interval, patients with only non-obstructive lesions at baseline were enrolled. CCTA was quantitatively analyzed at both the per-patient and per-lesion level. Models predicting the development of an obstructive lesion at follow up using either the per-patient or per-lesion level CCTA measures were constructed and compared. From 1297 patients (mean age 60 ± 9 years, 43% men) enrolled, a total of 3218 non-obstructive lesions were identified at baseline. At follow-up (inter-scan interval: 3.8 ± 1.6 years), 76 lesions (2.4%, 60 patients) became obstructive, defined as > 50% diameter stenosis. The C-statistics of Model 1, adjusted only by clinical risk factors, was 0.684. The addition of per-patient level total plaque volume (PV) and the presence of high-risk plaque (HRP) features to Model 1 improved the C-statistics to 0.825 [95% confidence interval (CI) 0.823-0.827]. When per-lesion level PV and the presence of HRP were added to Model 1, the predictive value of the model improved the C-statistics to 0.895 [95% CI 0.893-0.897]. The model utilizing per-lesion level CCTA measures was superior to the model utilizing per-patient level CCTA measures in predicting the development of an obstructive lesion (p < 0.001). Lesion-level analysis of coronary atherosclerotic plaques with CCTA yielded better predictive power for the development of obstructive CAD than the simple quantification of total coronary atherosclerotic burden at a per-patient level.Clinical Trial Registration: ClinicalTrials.gov NCT0280341.

Impact of Non-obstructive left main disease on the progression of coronary artery disease: A PARADIGM substudy.

BACKGROUND: The aim of the study is examine the impact of non-obstructive (<50%stenosis) left main (LM) disease on the natural history of coronary artery disease using serial coronary computed tomography angiography (CTA). METHODS: CTAs from the PARADIGM (Progression of atherosclerotic plaque determined by computed tomographic angiography imaging) study, a prospective multinational registry of patients who underwent serial CTA at a ≥2 year interval were analyzed. Those without evidence of CAD on their baseline scan were excluded, as were those with obstructive left main disease. Coronary artery vessels and their branches underwent quantification of: plaque volume and composition; diameter stenosis; presence of high-risk plaque. RESULTS: Of 944 (62 ±â€¯9 years, 60% male) who had evidence of CAD at baseline, 444 (47%) had LM disease. Those with LM disease had a higher baseline plaque volume (194.8 ±â€¯221mm3 versus 72.9 ±â€¯84.3mm3, p < 0.001) and a higher prevalence of high-risk plaque (17.5% versus 13%, p < 0.001) than those without LM disease. On multivariable general linear model, patients with LM disease had greater annual rates of progression of total (26.5 ±â€¯31.4mm3/yr versus 14.9 ±â€¯20.1mm3/yr, p < 0.001) and calcified plaque volume (17 ±â€¯24mm3/yr versus 7 ±â€¯11mm3/yr, p < 0.001), with no difference in fibrous, fibrofatty or necrotic core plaque components. CONCLUSION: The presence of non-obstructive LM disease is associated with greater rates of plaque progression and a higher prevalence of high-risk plaque throughout the entire coronary artery tree compared to CAD without LM involvement. Our data suggests that non-obstructive LM disease may be a marker for an aggressive phenotype of CAD that may benefit from more intensive treatment strategies.