Search

BackgroundDespite the essential role of ectopic osteogenic calcium-phosphate metabolism in the development of calcific aortic valve disease (CAVD), the implications of high serum phosphate levels in CAVD development are not fully understood.MethodsAsymptomatic individuals who underwent health screening using serial cardiac computed tomography (CT) and echocardiography were selected from a multicenter registry. CAVD was identified and quantified on CT images using the aortic valve calcification (AVC) score. The associations between initial serum phosphate levels and the presence of baseline CAVD, development of new CAVD, and the AVC score progression rate were investigated using multivariable regression models.ResultsA total of 736 individuals were selected for analysis, and the median interscan duration was 36.4 months. On initial CT, 83 (13.7%) participants had baseline CAVD, while 52 (7.0%) individuals developed new CAVD during follow-up. Serum phosphate levels were not associated with a higher probability of baseline CAVD but were predictive of newly developed CAVD (odds ratio per 1 mg/dl, 1.05; 95% confidence interval, 1.01–1.10; p = 0.02). Higher phosphate levels were also associated with a faster AVC score progression in those with baseline CAVD (regression coefficient per 1 mg/dl, 15.55 Agatston units/year; 95% confidence interval, 6.02–25.07; p

Radiomics has gained popularity as a quantitative analysis method for medical images. However, computed tomography (CT) scans are performed using various parameters, such as X-ray dose and reconstruction kernels, which is a fundamental reason for the lack of reproducibility of radiomic features. This study evaluated whether the proposed network improves the reproducibility of radiomic features across various CT protocols and reconstruction kernels. We set five CT scan protocols and two reconstruction kernels to create various noise settings for the obtained CT images with an abdominal phantom. We developed an enhanced hierarchical feature synthesis (EHFS) network to improve the reproducibility of radiomic features across various CT protocols and reconstruction kernels. Eight hundred and nineteen radiomic features were extracted, including first-order, second-order, and wavelet features. Reproducibility was assessed using Lin’s concordance correlation coefficient (CCC) on internal and external testing. We considered a radiomic feature with CCC $\ge0.85$ as a high-agreement feature. As a result, the average number of reproducible features increased in all protocols, from 241 ± 38 to 565 ± 11 in internal testing. In external testing, consisting of a new phantom and unseen protocol, 239 ± 74 reproducible features were in source images and 324 ± 16 were in generated images. The EHFS network is a novel approach to improving the reproducibility of radiomic features. It outperforms existing methods in reproducibility and generalization, as demonstrated by comprehensive experiments on both internal and external datasets. Our deep-learning-based CT image conversion could be a solution for standardization in ongoing radiomics research.

Abstract Background This study aimed to develop a prediction model for transferring patients to an inappropriate hospital for suspected cardiovascular emergency diseases at the pre-hospital stage, using variables obtained from an integrated nationwide dataset, and to assess the performance of this model. Methods We integrated three nationwide datasets and developed a two-step prediction model utilizing a machine learning algorithm. Ninety-eight clinical characteristics of patients identified at the pre-hospital stage and 13 hospital components were used as input data for the model. The primary endpoint of the model was the prediction of transfer to an inappropriate hospital. Results A total of 94,256 transferred patients in the public pre-hospital care system matched the National Emergency Department Information System data of patients with a pre-hospital cardiovascular registry created in South Korea between July 2017 and December 2018. Of these, 1,770 (6.26%) patients failed to be transferred to a capable hospital. The area under the receiver operating characteristic curve of the final predictive model was 0.813 (0.800–0.825), and the area under the receiver precision-recall curve was 0.286 (0.265–0.308). Conclusions Our prediction model used machine learning to show favorable performance in transferring patients with suspected cardiovascular disease to a capable hospital. For our results to lead to changes in the pre-hospital care system, a digital platform for sharing real-time information should be developed.

In coronary computed tomography angiography (CCTA), motion artifacts due to heartbeats can obscure coronary artery diagnoses. In this study, we introduce a cycle-consistent adversarial-network-based method for motion artifact correction in CCTA. Our methodology involves extracting image patches and using style transfer for synthetic ground truth creation, followed by CycleGAN network training for motion compensation. We employ Dynamic Time Warping (DTW) to align extracted image patches along the artery centerline with their corresponding motion-free phase patches, ensuring matched pixel correspondences and similar anatomical features for accuracy in subsequent processing steps. Our quantitative analysis, using metrics like the Dice Similarity Coefficient (DSC) and Hausdorff Distance (HD), demonstrates CycleGAN’s superior performance in reducing motion artifacts, with improvements in image quality and clarity. An observer study using a 5-point Likert scale further validates the reduction of motion artifacts and improved visibility of coronary arteries. Additionally, we present a quantitative analysis on clinical data, affirming the correction of motion artifacts through metric-based evaluations.

Coronary artery procedures are primarily performed based on X-ray angiography images. However, coronary arteries in X-ray images are often partially broken, complicating diagnoses and procedures owing to lack of visibility. In this paper, we propose a fully automatic method to restore locally broken parts of coronary arteries in X-ray images without using any external information, such as computed tomography images. To this end, we design a new multi-scale generative adversarial network and a vesselness-loss function. The proposed method is optimized for focus on elongated structures and can be utilized in various clinical applications. The proposed method is evaluated and compared with four other existing methods using the performance metrics, PSNR, MSE, and SSIM, and the result shows 34.3, 0.18, and 0.91 averages, respectively for each metric. Based on the performance result, the blocked regions are plausibly reconstructed into such original shapes of blood vessels, which can aid in image-based guiding catheter manipulation during coronary artery procedures. Eventually, the proposed method can be utilized in various clinical applications, e.g., image-based planning and guidance of coronary procedures and prior simulation of results.

Abstract In this retrospective observational study, we aimed to develop a machine-learning model using data obtained at the prehospital stage to predict in-hospital cardiac arrest in the emergency department (ED) of patients transferred via emergency medical services. The dataset was constructed by attaching the prehospital information from the National Fire Agency and hospital factors to data from the National Emergency Department Information System. Machine-learning models were developed using patient variables, with and without hospital factors. We validated model performance and used the SHapley Additive exPlanation model interpretation. In-hospital cardiac arrest occurred in 5431 of the 1,350,693 patients (0.4%). The extreme gradient boosting model showed the best performance with area under receiver operating curve of 0.9267 when incorporating the hospital factor. Oxygen supply, age, oxygen saturation, systolic blood pressure, the number of ED beds, ED occupancy, and pulse rate were the most influential variables, in that order. ED occupancy and in-hospital cardiac arrest occurrence were positively correlated, and the impact of ED occupancy appeared greater in small hospitals. The machine-learning predictive model using the integrated information acquired in the prehospital stage effectively predicted in-hospital cardiac arrest in the ED and can contribute to the efficient operation of emergency medical systems.

Abstract Background and Objective Macitentan is approved for treating pulmonary arterial hypertension. However, the real-world evidence of macitentan use is limited. Therefore, we evaluated the safety and clinical outcomes of macitentan use in clinical practice under a post-marketing surveillance. Methods Patients with pulmonary arterial hypertension receiving macitentan treatment were prospectively and consecutively enrolled from 2014 to 2020 at 50 medical centers in Korea. Safety and clinical outcomes were monitored from baseline to the nearest timepoint of 24 weeks after macitentan initiation. The adverse events and adverse drug reactions were identified. Changes in the World Health Organization functional class were assessed as the primary clinical outcome, which was used to estimate the final effectiveness (both improved and maintained). Factors associated with safety and final effectiveness were identified. Results Among 474 patients enrolled in the study, 467 and 440 were included in the safety and clinical outcome analyses, respectively. Dyspnea, nasopharyngitis, and worsening pulmonary arterial hypertension were the most frequent adverse events with incidences of 5%, 3%, and 3%, respectively. The final effectiveness rate was 93%. Older age (adjusted odds ratio [aOR] = 1.021, p = 0.003) and higher level (III vs II) of baseline World Health Organization functional class (aOR = 1.784; p = 0.022) were significantly associated with a higher adverse event occurrence. Younger age (aOR = 0.947; p = 0.001) and shorter disease duration (aOR = 0.991; p = 0.010) were significantly associated with positive final effectiveness. Conclusions This real-world study demonstrated the safety and clinical outcomes of macitentan use in Korean patients with pulmonary arterial hypertension. Macitentan was well tolerated and significantly effective with no new safety concerns during the 24 weeks.

Abstract Background The baseline coronary plaque burden is the most important factor for rapid plaque progression (RPP) in the coronary artery. However, data on the independent predictors of RPP in the absence of a baseline coronary plaque burden are limited. Thus, this study aimed to investigate the predictors for RPP in patients without coronary plaques on baseline coronary computed tomography angiography (CCTA) images. Methods A total of 402 patients (mean age: 57.6 ± 10.0 years, 49.3% men) without coronary plaques at baseline who underwent serial coronary CCTA were identified from the Progression of Atherosclerotic Plaque Determined by Computed Tomographic Angiography Imaging (PARADIGM) registry and included in this retrospective study. RPP was defined as an annual change of ≥ 1.0%/year in the percentage atheroma volume (PAV). Results During a median inter-scan period of 3.6 years (interquartile range: 2.7–5.0 years), newly developed coronary plaques and RPP were observed in 35.6% and 4.2% of the patients, respectively. The baseline traditional risk factors, i.e., advanced age (≥ 60 years), male sex, hypertension, diabetes mellitus, hyperlipidemia, obesity, and current smoking status, were not significantly associated with the risk of RPP. Multivariate linear regression analysis showed that the serum hemoglobin A1c level (per 1% increase) measured at follow-up CCTA was independently associated with the annual change in the PAV (β: 0.098, 95% confidence interval [CI]: 0.048–0.149; P

Abstract Background The triglyceride glucose (TyG) index has been suggested as a reliable surrogate marker of insulin resistance which is a substantial risk factor for atherosclerotic cardiovascular disease (ASCVD). Several recent studies have shown the relationship between the TyG index and cardiovascular disease; however, the role of the TyG index in coronary artery calcification (CAC) progression has not been extensively assessed especially in low-risk population. Methods We enrolled 5775 Korean adults who had at least two CAC evaluations. We determined the TyG index using ln (fasting triglycerides [mg/dL] x fasting glucose [mg/dL]/2). The CAC progression was defined as either incident CAC in a CAC-free population at baseline or an increase of ≥ 2.5 units between the square roots of the baseline and follow-up coronary artery calcium scores (CACSs) of subjects with detectable CAC at baseline. Results CAC progression was seen in 1,382 subjects (23.9%) during mean 3.5 years follow-up. Based on the TyG index, subjects were stratified into four groups. Follow-up CACS and incidence of CAC progression were markedly elevated with rising TyG index quartiles (group I [lowest]:17.6% vs. group II:22.2% vs. group III:24.6% vs. group IV [highest]: 31.3%, p

BackgroundDespite the importance of attaining optimal lipid levels from a young age to secure long-term cardiovascular health, the detailed impact of non-optimal lipid levels in young adults on coronary artery calcification (CAC) is not fully explored. We sought to investigate the risk of CAC progression as per lipid profiles and to demonstrate lipid optimality in young adults.MethodsFrom the KOrea Initiative on Coronary Artery calcification (KOICA) registry that was established in six large volume healthcare centers in Korea, 2,940 statin-naïve participants aged 20–45 years who underwent serial coronary calcium scans for routine health check-ups between 2002 and 2017 were included. The study outcome was CAC progression, which was assessed by the square root method. The risk of CAC progression was analyzed according to the lipid optimality and each lipid parameter.ResultsIn this retrospective cohort (mean age, 41.3 years; men 82.4%), 477 participants (16.2%) had an optimal lipid profile, defined as triglycerides 60 mg/dl. During follow-up (median, 39.7 months), CAC progression was observed in 434 participants (14.8%), and more frequent in the non-optimal lipid group (16.5% vs. 5.7%; p

Abstract Background We evaluated the accuracy of the prehospital Field Triage Decision Scheme, which has recently been applied in the Korean trauma system, and the factors associated with severe injury and prognosis at a regional trauma center in Korea. Methods From 2016 to 2018, prehospital data of injured patients were obtained from the emergency medical services of the national fire agency and matched with trauma outcomes at our institution. Severe injury (Injury Severity Score > 15), overtriage/undertriage rate, positive predictive value, negative predictive value, and accuracy were reviewed according to the triage protocol steps. A multivariate logistic regression analysis was performed to identify influencing factors in the field triage. Results Of the 2438 patients reviewed, 853 (35.0%) were severely injured. The protocol accuracy was as follows: step 1, 72.3%; step 2, 65.0%; step 3, 66.2%; step 1 or 2, 70.2%; and step 1, 2, or 3, 66.4%. Odds ratios (OR) (95% confidence interval [CIfor systolic blood pressure

Pulmonary hypertension (PH) is a condition of increased blood pressure in the pulmonary arteries and is diagnosed with an increased a mean pulmonary artery pressure ≥25 mm Hg. This condition may be associated with multiple clinical situations. Based on pathophysiological mechanisms, clinical presentation, hemodynamic profiles, and treatment strategies, the patients were classified into five clinical groups. Although there have been major advances in the management of PH, it is still associated with significant morbidity and mortality. The diagnosis and treatment of PH have been performed mainly by following European guidelines, even in Korea because the country lacks localized PH guidelines. European treatment guidelines do not reflect the actual status of Korea. Therefore, the European diagnosis and treatment of PH have not been tailored well to suit the needs of Korean patients with PH. To address this issue, we developed this guideline to facilitate the diagnosis and treatment of PH appropriately in Korea, a country where the consensus for the diagnosis and treatment of PH remains insufficient. This is the first edition of the guidelines for the diagnosis and treatment of PH in Korea, and it is primarily based on the ‘2015 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension.’ with the acceptance and adaptation of recent publications of PH.

Abstract Patient-specific phenotyping of coronary atherosclerosis would facilitate personalized risk assessment and preventive treatment. We explored whether unsupervised cluster analysis can categorize patients with coronary atherosclerosis according to their plaque composition, and determined how these differing plaque composition profiles impact plaque progression. Patients with coronary atherosclerotic plaque (n = 947; median age, 62 years; 59% male) were enrolled from a prospective multi-national registry of consecutive patients who underwent serial coronary computed tomography angiography (median inter-scan duration, 3.3 years). K-means clustering applied to the percent volume of each plaque component and identified 4 clusters of patients with distinct plaque composition. Cluster 1 (n = 52), which comprised mainly fibro-fatty plaque with a significant necrotic core (median, 55.7% and 16.0% of the total plaque volume, respectively), showed the least total plaque volume (PV) progression (+ 23.3 mm3), with necrotic core and fibro-fatty PV regression (− 5.7 mm3 and − 5.6 mm3, respectively). Cluster 2 (n = 219), which contained largely fibro-fatty (39.2%) and fibrous plaque (46.8%), showed fibro-fatty PV regression (− 2.4 mm3). Cluster 3 (n = 376), which comprised mostly fibrous (62.7%) and calcified plaque (23.6%), showed increasingly prominent calcified PV progression (+ 21.4 mm3). Cluster 4 (n = 300), which comprised mostly calcified plaque (58.7%), demonstrated the greatest total PV increase (+ 50.7mm3), predominantly increasing in calcified PV (+ 35.9 mm3). Multivariable analysis showed higher risk for plaque progression in Clusters 3 and 4, and higher risk for adverse cardiac events in Clusters 2, 3, and 4 compared to that in Cluster 1. Unsupervised clustering algorithms may uniquely characterize patient phenotypes with varied atherosclerotic plaque profiles, yielding distinct patterns of progressive disease and outcome.

Background Previously developed prediction models for type 2 diabetes mellitus (T2DM) have limited performance. We developed a deep learning (DL) based model using a cohort representative of the Korean population. Methods This study was conducted on the basis of the National Health Insurance Service-Health Screening (NHIS-HEALS) cohort of Korea. Overall, 335,302 subjects without T2DM at baseline were included. We developed the model based on 80% of the subjects, and verified the power in the remainder. Predictive models for T2DM were constructed using the recurrent neural network long short-term memory (RNN-LSTM) network and the Cox longitudinal summary model. The performance of both models over a 10-year period was compared using a time dependent area under the curve. Results During a mean follow-up of 10.4±1.7 years, the mean frequency of periodic health check-ups was 2.9±1.0 per subject. During the observation period, T2DM was newly observed in 8.7% of the subjects. The annual performance of the model created using the RNN-LSTM network was superior to that of the Cox model, and the risk factors for T2DM, derived using the two models were similar; however, certain results differed. Conclusion The DL-based T2DM prediction model, constructed using a cohort representative of the population, performs better than the conventional model. After pilot tests, this model will be provided to all Korean national health screening recipients in the future.

Background: The prognostic value of whole vessel plaque quantification has not been fully understood. Objectives: We aimed to investigate the clinical relevance of whole vessel plaque quantification on coronary computed tomography angiography. Methods: In a total of 1,013 vessels with fractional flow reserve (FFR) measurement and available coronary computed tomography angiography, high-risk plaque characteristics (HRPC) included minimum lumen area 0.80, the number of HRVC was significantly associated with the risk of VOCO (HR: 2.54; 95% CI: 1.77-3.64) and enhanced the predictability for VOCO of % diameter stenosis and the number of HRPC (P < 0.001). In a landmark analysis at 2 years, the number of HRVC showed sustained prognostic implications beyond 2 years, but the number of HRPC did not. Conclusions: Whole vessel plaque quantification can provide incremental predictability for low FFR and additive prognostic value in deferred vessels with high FFR over anatomical severity and lesion plaque characteristics. (CCTA-FFR Registry for Risk Prediction; NCT04037163)

Abstract Low-density-lipoprotein cholesterol (LDL-C) is the main target in atherosclerotic cardiovascular disease (ASCVD). We aimed to validate and compare a new LDL-C estimation equation with other well-known equations. 177,111 samples were analysed from two contemporary population-based cohorts comprising asymptomatic Korean adults who underwent medical examinations. Performances of the Friedewald (FLDL), Martin (MLDL), and Sampson (SLDL) equations in estimating direct LDL-C by homogenous assay were assessed by measures of concordance (R2, RMSE, and mean absolute difference). Analyses were performed according to various triglyceride (TG) and/or LDL-C strata. Secondary analyses were conducted within dyslipidaemia populations of each database. MLDL was superior or at least similar to other equations regardless of TG/LDL-C, in both the general and dyslipidaemia populations (RMSE = 11.45/9.20 mg/dL; R2 = 0.88/0.91; vs FLDL: RMSE = 13.66/10.42 mg/dL; R2 = 0.82/0.89; vs SLDL: RMSE = 12.36/9.39 mg/dL; R2 = 0.85/0.91, per Gangnam Severance Hospital Check-up/Korea Initiatives on Coronary Artery Calcification data). MLDL had a slight advantage over SLDL with the lowest MADs across the full spectrum of TG levels, whether divided into severe hyper/non-hyper to moderate hypertriglyceridaemia samples or stratified by 100-mg/dL TG intervals, even up to TG values of 500–600 mg/dL. MLDL may be a readily adoptable and cost-effective alternative to direct LDL-C measurement, irrespective of dyslipidaemia status. In populations with relatively high prevalence of mild-to-moderate hypertriglyceridaemia, Martin’s equation may be optimal for LDL-C and ASCVD risk estimation.

Abstract Background The triglyceride glucose (TyG) index is a noninsulin-based marker for insulin resistance (IR) in general practice. Although smoking and heavy drinking have been regarded as major risk factors for various chronic diseases, there is limited evidence regarding the combined effects of smoking and alcohol consumption on IR. This study aimed to investigate the relationship between the TyG index and smoking and alcohol consumption using two Korean population-based datasets. Methods This study included 10,568 adults in the Korean National Health and Nutrition Examination Survey (KNHANES) and 9586 adults in the Korean Initiatives on Coronary Artery Calcification (KOICA) registry datasets. Multivariate logistic analysis was conducted to explore the relationship between smoking and alcohol consumption and the TyG index. To assess the predictive value of smoking and alcohol consumption on high TyG index, the area under the curve (AUC) were compared and net reclassification improvement (NRI) and integrated discrimination improvement (IDI) analyses were derived. Results The combined effect of smoking and alcohol consumption was an independent risk factor of a higher TyG index in the KNHANES (adjusted odds ratio: 4.33, P

BackgroundCoronary computed tomography angiography (CTA) and optical coherence tomography (OCT) provide additional functional information beyond the anatomy by applying computational fluid dynamics (CFD). This study sought to evaluate a novel approach for estimating computational fractional flow reserve (FFR) from coronary CTA-OCT fusion images.MethodsAmong patients who underwent coronary CTA, 148 patients who underwent both pressure wire-based FFR measurement and OCT during angiography to evaluate intermediate stenosis in the left anterior descending artery were included from the prospective registry. Coronary CTA-OCT fusion images were created, and CFD was applied to estimate computational FFR. Based on pressure wire-based FFR as a reference, the diagnostic performance of Fusion-FFR was compared with that of CT-FFR and OCT-FFR.ResultsFusion-FFR was strongly correlated with FFR (r = 0.836, P < 0.001). Correlation between FFR and Fusion-FFR was stronger than that between FFR and CT-FFR (r = 0.682, P < 0.001; z statistic, 5.42, P < 0.001) and between FFR and OCT-FFR (r = 0.705, P < 0.001; z statistic, 4.38, P < 0.001). Area under the receiver operating characteristics curve to assess functionally significant stenosis was higher for Fusion-FFR than for CT-FFR (0.90 vs. 0.83, P = 0.024) and OCT-FFR (0.90 vs. 0.83, P = 0.043). Fusion-FFR exhibited 84.5% accuracy, 84.6% sensitivity, 84.3% specificity, 80.9% positive predictive value, and 87.5% negative predictive value. Especially accuracy, specificity, and positive predictive value were superior for Fusion-FFR than for CT-FFR (73.0%, P = 0.007; 61.4%, P < 0.001; 64.0%, P < 0.001) and OCT-FFR (75.7%, P = 0.021; 73.5%, P = 0.020; 69.9%, P = 0.012).ConclusionCFD-based computational FFR from coronary CTA-OCT fusion images provided more accurate functional information than coronary CTA or OCT alone.Clinical Trial Registration[www.ClinicalTrials.gov], identifier [NCT03298282].

Abstract Background Hypertension (HTN), diabetes mellitus (DM), and dyslipidemia (DL) are well-known risk factors of cardiovascular disease (CVD), but not all patients develop CVDs. Studies have been limited investigating genetic risk of CVDs specific to individuals with metabolic diseases. This study aimed to identify disease-specific and/or common genetic loci associated with CVD susceptibility in chronic metabolic disease patients. Methods We conducted a genome-wide association study (GWAS) of a multiple case–control design with data from the City Cohort within Health EXAminees subcohort of the Korean Genome and Epidemiology Study (KoGES_HEXA). KoGES_HEXA is a population-based prospective cohort of 173,357 urban Korean adults that had health examinations at medical centers. 42,393 participants (16,309 HTN; 5,314 DM; 20,770 DL) were analyzed, and each metabolic disease group was divided into three CVD case-controls: coronary artery disease (CAD), ischemic stroke (IS), and cardio-cerebrovascular disease (CCD). GWASs were conducted for each case–control group with 7,975,321 imputed single nucleotide polymorphisms using the Phase 3 Asian panel from 1000 Genomes Project, by logistic regression and controlled for confounding variables. Genome-wide significant levels were implemented to identify important susceptibility loci. Results Totaling 42,393 individuals, this study included 16,309 HTN (mean age [SD], 57.28 [7.45]; 816 CAD, 398 IS, and 1,185 CCD cases), 5,314 DM (57.79 [7.39]; 361 CAD, 153 IS, and 497 CCD cases), and 20,770 DL patients (55.34 [7.63]; 768 CAD, 295 IS, and 1,039 CCD cases). Six genome-wide significant CVD risk loci were identified, with relatively large effect sizes: 1 locus in HTN (HTN-CAD: 17q25.3/CBX8-CBX4 [OR, 2.607; P = 6.37 × 10−9]), 2 in DM (DM-IS: 4q32.3/MARCH1-LINC01207 [OR, 5.587; P = 1.34 × 10−8], and DM-CCD: 17q25.3/RPTOR [OR, 3.511; P = 1.99 × 10−8]), and 3 in DL (DL-CAD: 9q22.2/UNQ6494-LOC101927847 [OR, 2.282; P = 7.78 × 10−9], DL-IS: 3p22.1/ULK4 [OR, 2.162; P = 2.97 × 10−8], and DL-CCD: 2p22.2/CYP1B1-CYP1B1-AS1 [OR, 2.027; P = 4.24 × 10−8]). Conclusions This study identified 6 susceptibility loci and positional candidate genes for CVDs in HTN, DM, and DL patients using an unprecedented study design. 1 locus (17q25.3) was commonly associated with CAD. These associations warrant validation in additional studies for potential therapeutic applications.

Abstract We investigated the optimal combinations of systolic blood pressure (SBP) and diastolic blood pressure (DBP) levels for lowest mortality in participants not taking hypertensive medication at the study baseline using nationwide representative databases. Survival rates and hazard ratios (HRs) were calculated using Kaplan‐Meier curves and multivariable Cox regression analyses. The discriminatory ability for clinical outcomes was assessed by Harrell's C‐index analysis. A survival spline curve was presented, and Classification and Regression Tree (CART) analysis was performed. SBP ≥ 140 group and DBP ≥ 90 group had the highest risk of mortality. Within SBP

Automatic detection and classification of thoracic diseases using deep learning algorithms have many applications supporting radiologists’ diagnosis and prognosis. However, in the medical field, the class-imbalanced problem is extremely common due to the differences in prevalence among diseases, making it difficult to develop these applications. Many GAN-based methods have been proposed to solve the class-imbalance problem on chest X-ray (CXR) data. However, these models have not been trained well for small-sized diseases because it is challenging to extract sufficient information with only a few pixels. In this paper, we propose a novel deep generative model called a class activation region influence maximization conditional generative adversarial network (CARIM-cGAN). The proposed network can control the target disease’s presence, location, and size with a controllable conditional mask. We newly introduced class activation region influence maximization (CARIM) loss to maximize the probability of disease occurrence in the bounded region represented by a conditional mask. To demonstrate an enhanced generative performance, we conducted numerous qualitative and quantitative evaluations with the samples generated using a CARIM-cGAN. The results showed that our method has a better performance than other methods. In conclusion, because the CARIM-cGAN can generate high-quality samples based on information on the location and size of the disease, we can contribute to solving problems such as disease classification, -detection, and -localization, requiring a higher annotation cost.

Abstract This study aimed to evaluate the association between the atherogenic index of plasma (AIP), which has been suggested as a novel marker for atherosclerosis, and coronary artery calcification (CAC) progression according to the baseline coronary artery calcium score (CACS). We included 12,326 asymptomatic Korean adults who underwent at least two CAC evaluations from December 2012 to August 2016. Participants were stratified into four groups according to AIP quartiles, which were determined by the log of (triglyceride/high-density lipoprotein cholesterol). Baseline CACSs were divided into three groups: 0, 1 − 100, and > 100. CAC progression was defined as a difference ≥ 2.5 between the square roots (√) of the baseline and follow-up CACSs (Δ√transformed CACS). Annualized Δ√transformed CACS was defined as Δ√transformed CACS divided by the inter-scan period. During a mean 3.3-year follow-up period, the overall incidence of CAC progression was 30.6%. The incidences of CAC progression and annualized Δ√transformed CACS were markedly elevated with increasing AIP quartile in participants with baseline CACSs of 0 and 1 − 100, but not in those with a baseline CACS > 100. The AIP level was associated with the annualized Δ√transformed CACS in participants with baseline CACSs of 0 (β = 0.016; P 100 (β = 0.032; P = 0.385). After adjusting for traditional risk factors, the AIP was significantly associated with CAC progression in those with baseline CACS ≤ 100. The AIP has value for predicting CAC progression in asymptomatic adults without heavy baseline CAC.

BackgroundLDL-C is the primary target of lipid-lowering therapy and used to classify patients by cardiovascular disease risk. We aimed to develop a deep neural network (DNN) model to estimate LDL-C levels and compare its performance with that of previous LDL-C estimation equations using two large independent datasets of Korean populations.MethodsThe final analysis included participants from two independent population-based cohorts: 129,930 from the Gangnam Severance Health Check-up (GSHC) and 46,470 participants from the Korean Initiatives on Coronary Artery Calcification registry (KOICA). The DNN model was derived from the GSHC dataset and validated in the KOICA dataset. We measured our proposed model's performance according to bias, root mean-square error (RMSE), proportion (P)10–P20, and concordance. P was defined as the percentage of patients whose LDL was within ±10–20% of the measured LDL. We further determined the RMSE scores of each LDL equation according to Pooled cohort equation intervals.ResultsOur DNN method has lower bias and root mean-square error than Friedewald's, Martin's, and NIH equations, showing a high agreement with LDL-C measured by homogenous assay. The DNN method offers more precise LDL estimation in all pooled cohort equation strata.ConclusionThis method may be particularly helpful for managing a patient's cholesterol levels based on their atherosclerotic cardiovascular disease risk.

Abstract Background Accurate measurement of the maximum aortic diameter (Dmax) is crucial for patients with abdominal aortic aneurysm (AAA). Aortic computed tomography (CT) provides accurate Dmax values by three-dimensional (3-D) reconstruction but may cause nephrotoxicity because of contrast use and radiation hazard. We aimed to evaluate the accuracy of a novel semi-automated 3-D ultrasonography (3-D US) system compared with that of CT as a reference. Methods Patients with AAA (n = 59) or individuals with normal aorta (n = 18) were prospectively recruited in an outpatient setting. Two-dimensional ultrasonography (2-D US) and 3-D US images were acquired with a single-sweep volumetric transducer. The analysis was performed offline with a software. Dmax and the vessel area of the Dmax slice were measured with 2-D US, 3-D US, and CT. The lumen and thrombus areas of the Dmax slice were also measured in 40 patients with intraluminal thrombus. Vessel and thrombus volumes were measured using 3-D US and CT. Results The Dmax values from 3-D US demonstrated better agreement (R 2 = 0.984) with the CT values than with the 2-D US values (R 2 = 0.938). Overall, 2-D US underestimated Dmax compared with 3-D US (32.3 ± 12.1 mm vs. 35.1 ± 12.0 mm). The Bland-Altman analysis of the 3-D US values, revealed better agreement with the CT values (2 standard deviations [SD], 2.9 mm) than with the 2-D US values (2 SD, 5.4 mm). The vessel, lumen, and thrombus areas all demonstrated better agreement with CT than with 2-D US (R 2 = 0.986 vs. 0.960 for the vessel, R 2 = 0.891 vs. 0.837 for the lumen, and R 2 = 0.977 vs. 0.872 for the thrombus). The thrombus volume assessed with 3-D US showed good correlation with the CT value (R 2 = 0.981 and 2 SD in the Bland-Altman analysis: 13.6 cm3). Conclusions Our novel semi-automated 3-D US analysis system provides more accurate Dmax values than 2-D US and provides precise volumetric data, which were not evaluable with 2-D US. The application of the semi-automated 3-D US analysis system in abdominal aorta assessment is easy and accurate.

Abstract Background The association between triglyceride glucose (TyG) index and coronary atherosclerotic change remains unclear. We aimed to evaluate the association between TyG index and coronary plaque progression (PP) using serial coronary computed tomography angiography (CCTA). Methods A total of 1143 subjects (aged 60.7 ± 9.3 years, 54.6% male) who underwent serial CCTA with available data on TyG index and diabetic status were analyzed from The Progression of AtheRosclerotic PlAque DetermIned by Computed TomoGraphic Angiography IMaging (PARADIGM) registry. PP was defined as plaque volume (PV) (mm3) at follow-up minus PV at index > 0. Annual change of PV (mm3/year) was defined as PV change divided by inter-scan period. Rapid PP was defined as the progression of percent atheroma volume (PV divided by vessel volume multiplied by 100) ≥ 1.0%/year. Results The median inter-scan period was 3.2 (range 2.6–4.4) years. All participants were stratified into three groups based on TyG index tertiles. The overall incidence of PP was 77.3%. Baseline total PV (group I [lowest]: 30.8 (0.0–117.7), group II: 47.2 (6.2–160.4), and group III [highest]: 57.5 (8.4–154.3); P

Abstract Background Data on the relationship between the triglyceride glucose (TyG) index and coronary artery calcification (CAC) progression is limited. This longitudinal study evaluated the association of TyG index with CAC progression in asymptomatic adults. Methods We enrolled 12,326 asymptomatic Korean adults who had at least two CAC evaluations. The TyG index was determined using ln (fasting triglycerides [mg/dL] × fasting glucose [mg/dL]/2). CAC progression was defined as a difference ≥ 2.5 between the square roots (√) of the baseline and follow-up coronary artery calcium score (CACS) (Δ√transformed CACS). Annualized Δ√transformed CACS was defined as Δ√transformed CACS divided by the inter-scan period. Results During a mean 3.3 years, the overall incidence of CAC progression was 30.6%. The incidence of CAC progression (group I [lowest]: 22.7% versus [vs.] group II: 31.7% vs. group III [highest]: 37.5%, P

Motion artifacts may occur in coronary computed tomography angiography (CCTA) due to the heartbeat and impede the clinician's diagnosis of coronary arterial diseases. Thus, motion artifact correction of the coronary artery is required to quantify the risk of disease more accurately. We present a novel method based on deep learning for motion artifact correction in CCTA. Because the image of the coronary artery without motion (the ground-truth data required in supervised deep learning) is medically unattainable, we apply a style transfer method to 2D image patches cropped from full-phase 4D computed tomography (CT) to synthesize these images. We then train a convolutional neural network (CNN) for motion artifact correction using this synthetic ground-truth (SynGT). During testing, the output motion-corrected 2D image patches of the trained network are reinserted into the 3D CT volume with volumetric interpolation. The proposed method is evaluated using both phantom and clinical data. A phantom study demonstrates comparable results to other methods in quantitative performance and outperforms those methods in computation time. For clinical data, a quantitative analysis based on metric measurements is presented that confirms the correction of motion artifacts. Moreover, an observer study finds that by applying the proposed method, motion artifacts are markedly reduced, and boundaries of the coronary artery are much sharper, with a strong inter-observer agreement (κ = 0.78). Finally, evaluations using commercial software on the original and resulting CT volumes of the proposed method reveal a considerable increase in tracked coronary artery length.

Models for predicting acute myocardial infarction (AMI) at the prehospital stage were developed and their efficacy compared, based on variables identified from a nationwide systematic emergency medical service (EMS) registry using conventional statistical methods and machine learning algorithms. Patients in the EMS cardiovascular registry aged >15 years who were transferred from the public EMS to emergency departments in Korea from January 2016 to December 2018 were enrolled. Two datasets were constructed according to the hierarchical structure of the registry. A total of 184,577 patients (Dataset 1) were included in the final analysis. Among them, 72,439 patients (Dataset 2) were suspected to have AMI at prehospital stage. Between the models derived using the conventional logistic regression method, the B-type model incorporated AMI-specific variables from the A-type model and exhibited a superior discriminative ability (p = 0.02). The models that used extreme gradient boosting and a multilayer perceptron yielded a higher predictive performance than the conventional logistic regression-based models for analyses that used both datasets. Each machine learning algorithm yielded different classification lists of the 10 most important features. Therefore, prediction models that use nationwide prehospital data and are developed with appropriate structures can improve the identification of patients who require timely AMI management.

Abstract Background Insulin resistance (IR) is an important risk factor for subclinical atherosclerosis. This study evaluated the relationship between the triglyceride glucose (TyG) index, which is a simple and reliable surrogate marker for IR, and arterial stiffness. Methods This study included 2560 Korean subjects without a previous history of coronary artery disease, stroke, and malignancies who participated in a community-based cohort study. Arterial stiffness was measured using the brachial-ankle pulse wave velocity (baPWV). Results All participants were stratified into four groups based on the quartile of the TyG index. The prevalence of metabolic syndrome and diabetes significantly increased with increasing TyG index quartile. The mean baPWV was significantly different among all groups (group I [lowest]: 1421 ± 242 vs. group II: 1480 ± 244 vs. group III: 1534 ± 260 vs. group IV [highest]: 1575 ± 279 cm/s; p

Even with increasing awareness of sex-related differences in atherosclerotic cardiovascular disease (ASCVD), it remains unclear whether the progression of coronary atherosclerosis differs between women and men. We sought to compare coronary artery calcium (CAC) progression between women and men. From a retrospective, multicentre registry of consecutive asymptomatic individuals who underwent CAC scoring, we identified 9,675 men and 1,709 women with follow-up CAC scoring. At baseline, men were more likely to have a CAC score >0 than were women (47.8% vs. 28.6%). The probability of CAC progression at 5 years, defined as [√CAC score (follow-up)-√CAC score (baseline)] ≥2.5, was 47.4% in men and 29.7% in women (p

Background Rapid coronary plaque progression (RPP) is associated with incident cardiovascular events. To date, no method exists for the identification of individuals at risk of RPP at a single point in time. This study integrated coronary computed tomography angiography–determined qualitative and quantitative plaque features within a machine learning (ML) framework to determine its performance for predicting RPP. Methods and Results Qualitative and quantitative coronary computed tomography angiography plaque characterization was performed in 1083 patients who underwent serial coronary computed tomography angiography from the PARADIGM (Progression of Atherosclerotic Plaque Determined by Computed Tomographic Angiography Imaging) registry. RPP was defined as an annual progression of percentage atheroma volume ≥1.0%. We employed the following ML models: model 1, clinical variables; model 2, model 1 plus qualitative plaque features; model 3, model 2 plus quantitative plaque features. ML models were compared with the atherosclerotic cardiovascular disease risk score, Duke coronary artery disease score, and a logistic regression statistical model. 224 patients (21%) were identified as RPP. Feature selection in ML identifies that quantitative computed tomography variables were higher‐ranking features, followed by qualitative computed tomography variables and clinical/laboratory variables. ML model 3 exhibited the highest discriminatory performance to identify individuals who would experience RPP when compared with atherosclerotic cardiovascular disease risk score, the other ML models, and the statistical model (area under the receiver operating characteristic curve in ML model 3, 0.83 [95% CI 0.78–0.89], versus atherosclerotic cardiovascular disease risk score, 0.60 [0.52–0.67]; Duke coronary artery disease score, 0.74 [0.68–0.79]; ML model 1, 0.62 [0.55–0.69]; ML model 2, 0.73 [0.67–0.80]; all P

OBJECTIVES:To develop, demonstrate and evaluate an automated deep learning method for multiple cardiovascular structure segmentation. BACKGROUND:Segmentation of cardiovascular images is resource-intensive. We design an automated deep learning method for the segmentation of multiple structures from Coronary Computed Tomography Angiography (CCTA) images. METHODS:Images from a multicenter registry of patients that underwent clinically-indicated CCTA were used. The proximal ascending and descending aorta (PAA, DA), superior and inferior vena cavae (SVC, IVC), pulmonary artery (PA), coronary sinus (CS), right ventricular wall (RVW) and left atrial wall (LAW) were annotated as ground truth. The U-net-derived deep learning model was trained, validated and tested in a 70:20:10 split. RESULTS:The dataset comprised 206 patients, with 5.130 billion pixels. Mean age was 59.9 ± 9.4 yrs., and was 42.7% female. An overall median Dice score of 0.820 (0.782, 0.843) was achieved. Median Dice scores for PAA, DA, SVC, IVC, PA, CS, RVW and LAW were 0.969 (0.979, 0.988), 0.953 (0.955, 0.983), 0.937 (0.934, 0.965), 0.903 (0.897, 0.948), 0.775 (0.724, 0.925), 0.720 (0.642, 0.809), 0.685 (0.631, 0.761) and 0.625 (0.596, 0.749) respectively. Apart from the CS, there were no significant differences in performance between sexes or age groups. CONCLUSIONS:An automated deep learning model demonstrated segmentation of multiple cardiovascular structures from CCTA images with reasonable overall accuracy when evaluated on a pixel level.

BackgroundPulmonary arterial hypertension (PAH) is a progressive chronic disease with poor outcomes. One reason for poor prognosis is the lack of understanding regarding individual variability in response to treatment. Idiopathic PAH (IPAH) patients with bone morphogenetic protein receptor type 2 (BMPR2) mutations have distinct phenotypes that are crucial for individualized therapy but evidence regarding their prevalence and clinical features in the Korean population is lacking. Therefore, the present study aimed to screen Korean IPAH patients for BMPR2 mutations and analyze their clinical phenotypes.MethodsWe enrolled 73 unrelated IPAH patients for BMPR2 mutation screening between March 2010 to November 2015 from 11 hospitals in Korea. Thirty-three lineal family members from 6 families of BMPR2 mutation carriers were also screened.ResultsAmong 73 patients, 16 (22%) had BMPR2 mutations. Mutation carriers were younger (27 vs. 47 years; p = 0.02) and had a higher mean pulmonary arterial pressure (mPAP) than non-carriers (64 vs. 51 mmHg; pConclusionThe prevalence of BMPR2 mutations in Korean IPAH patients was 22%. Mutation carriers were younger and had a poorer hemodynamic profile compared with the non-carriers.Clinical trial registrationClinicaltrials.gov NCT01054105.

BackgroundMachine learning (ML) is able to extract patterns and develop algorithms to construct data-driven models. We use ML models to gain insight into the relative importance of variables to predict obstructive coronary artery disease (CAD) using the Coronary Computed Tomographic Angiography for Selective Cardiac Catheterization (CONSERVE) study, as well as to compare prediction of obstructive CAD to the CAD consortium clinical score (CAD2). We further perform ML analysis to gain insight into the role of imaging and clinical variables for revascularization.MethodsFor prediction of obstructive CAD, the entire ICA arm of the study, comprising 719 patients was used. For revascularization, 1,028 patients were randomized to invasive coronary angiography (ICA) or coronary computed tomographic angiography (CCTA). Data was randomly split into 80% training 20% test sets for building and validation. Models used extreme gradient boosting (XGBoost).ResultsMean age was 60.6 ± 11.5 years and 64.3% were female. For the prediction of obstructive CAD, the AUC was significantly higher for ML at 0.779 (95% CI: 0.672-0.886) than for CAD2 (0.696 [95% CI: 0.594-0.798]) (P = 0.01). BMI, age, and angina severity were the most important variables. For revascularization, the model obtained an overall area under the receiver-operation curve (AUC) of 0.958 (95% CI = 0.933-0.983). Performance did not differ whether the imaging parameters used were from ICA (AUC 0.947, 95% CI = 0.903-0.990) or CCTA (AUC 0.941, 95% CI = 0.895-0.988) (P = 0.90). The ML model obtained sensitivity and specificity of 89.2% and 92.9%, respectively. Number of vessels with ≥70% stenosis, maximum segment stenosis severity (SSS) and body mass index (BMI) were the most important variables. Exclusion of imaging variables resulted in performance deterioration, with an AUC of 0.705 (95% CI 0.614-0.795) (P ConclusionsFor obstructive CAD, the ML model outperformed CAD2. BMI is an important variable, although currently not included in most scores. In this ML model, imaging variables were most associated with revascularization. Imaging modality did not influence model performance. Removal of imaging variables reduced model performance.

Emergency medical service (EMS) occurs in a high-pressure and error-prone environment, where paramedics must provide prompt decisions in care while recording information with limited time, incomplete data, restricted resources, and competing priorities. The EMS requires cooperative workflows between patients or caregivers, paramedics and medical centers in the community. In a conventional EMS, they have difficulties in obtaining causes of emergencies and personal medical histories, which are important for a rapid and proper response. We analyzed the requirement of a smart EMS (SEMS) system and derived the key components in connected care environments leveraging information and communication technology. A survey of paramedics (n=113) revealed that a SEMS system using IoT technology should integrate personal lifelogs, electronic medical records, and patient monitoring in ambulances into pre-hospital care recording systems. It also addressed context-awareness in the EMS accelerates first responder's activities, while supporting personalized care not only at the scene of the emergency but also during the entire hospital stay. Based on requirement analysis, we designed and implemented SEMS using health information standards to provide interoperability between devices and systems. As an application of SEMS, an example service is introduced: lifelog-connected EMS for stroke patients with a real-time location service for managing timeline of treatment.

Abstract Background Data on the influence of glycemic status on the progression of coronary calcification, an important marker for future adverse cardiovascular events, are limited. Methods Data from the Korea Initiatives on Coronary Artery Calcification (KOICA) registry on 12,441 asymptomatic Korean adults (52 ± 9 years, 84.2% males) without previous history of coronary artery disease and stroke, who underwent serial coronary artery calcification (CAC) screening examinations, were included in this study. The median inter-scan period was 3.0 (2.0–4.8) years. All participants were categorized into three groups based on their glycemic status: normal (n = 6578), pre-diabetes (n = 4146), and diabetes (n = 1717). CAC progression was defined as a difference ≥ 2.5 between the square roots (√) of the baseline and follow-up CAC scores. Results The incidence of CAC progression was significantly different between the three groups (normal, 26.3%; pre-diabetes, 30.9%; and diabetes, 46.9%; p

Abstract Background Arterial stiffness increases pressure load to the left ventricle (LV), leading to LV hypertrophy and subendocardial ischemia. Neurohormones stimulate myocardial fibrosis and LV dysfunction. We aimed to explore the associations of arterial stiffness and plasma aldosterone with multi-directional, layer-specific LV, and left atrial (LA) mechanical function in hypertensive patients. Methods Layer-specific LV global longitudinal strain (GLS-trans, GLS-endo, GLS-epi), global circumferential strain (GCS-trans, GCS-endo, GCS-epi), LV torsional parameters, and LA global longitudinal strain (LA GLS) were analyzed by two-dimensional speckle tracking echocardiography in 195 hypertensive patients (110 men, mean age 55 years). Pulse wave velocity (PWV) was analyzed as a measure of arterial stiffness, and plasma aldosterone was measured for evaluation of neurohormonal activation. Results In a simple correlation, PWV significantly correlated with LV GLS-endo and LA GLS. Log aldosterone correlated with both LV GCS-endo and LV GCS-trans. Multiple regression analysis revealed that LV GLS-endo (β = 0.223, p = 0.031) and LA GLS (β = −0.311, p = 0.002) were independently correlated with PWV even after controlling for confounding factors. Conclusions In hypertensive patients without clinically apparent target organ damage, LV GLS, especially endocardium, and LA GLS were more dominantly affected by arterial stiffness because, among the three myocardial layers, the endocardium is most susceptible to pressure overload. Two-dimensional layer-specific speckle-tracking echocardiography sensitively detects LV mechanical dysfunction and provides pathophysiologic insights into LV mechanical adaptations in hypertension.

Abstract Background Accurate evaluation of diplopia during bedside physical examination is challenging. We developed a new computerized red glass test (CRT) to detect, localize, and quantify diplopia and investigated whether the CRT is useful and feasible. Methods During the CRT, a white dot randomly appears on a monitor. Because a red glass is applied on the right eye, a patient can see one white dot and one red dot when diplopia is present. We defined the degree of diplopia as the direct distance of the two points with the largest deviation and compared the degree with the Hess score and Hess area ratio. Results We prospectively enrolled 14 patients with binocular diplopia. Test–retest reliability of the CRT was excellent (overall intraclass correlation coefficient = 0.948, 95% CI 0.939–0.956). The degree of diplopia in the CRT was well correlated with both the Hess score (r = 0.719, p = 0.005) and the Hess area ratio (r = −0.620, p = 0.018). Conclusions The CRT can easily detect the presence of diplopia and provided the quantitative values of the degree of diplopia. The CRT was useful and feasible for improving routine bedside examination.

[This corrects the article DOI: 10.1371/journal.pone.0211957.].

The early identification and prevention of dementia is important for reducing its worldwide burden and increasing individuals' quality of life. Although several dementia prediction models have been developed, there remains a need for a practical and precise model targeted to middle-aged and Asian populations. Here, we used national Korean health examination data from adults (331,126 individuals, 40-69 years of age, mean age: 52 years) from 2002-2003 to predict the incidence of dementia after 10 years. We divided the dataset into two cohorts to develop and validate of our prediction model. Cox proportional hazards models were used to construct dementia prediction models for the total group and sex-specific subgroups. Receiver operating characteristics curves, C-statistics, calibration plots, and cumulative hazards were used to validate model performance. Discriminative accuracy as measured by C-statistics was 0.81 in the total group (95% confidence interval (CI) = 0.81 to 0.82), 0.81 in the male subgroup (CI = 0.80 to 0.82), and 0.81 in the female subgroup (CI = 0.80 to 0.82). Significant risk factors for dementia in the total group were age; female sex; underweight; current hypertension; comorbid psychiatric or neurological disorder; past medical history of cardiovascular disease, diabetes mellitus, or hypertension; current smoking; and no exercise. All identified risk factors were statistically significant in the sex-specific subgroups except for low body weight and current hypertension in the female subgroup. These results suggest that public health examination data can be effectively used to predict dementia and facilitate the early identification of dementia within a middle-aged Asian population.

ObjectivesCardiovascular disease (CVD) is one of the major causes of death worldwide. For improved accuracy of CVD prediction, risk classification was performed using national time-series health examination data. The data offers an opportunity to access deep learning (RNN-LSTM), which is widely known as an outstanding algorithm for analyzing time-series datasets. The objective of this study was to show the improved accuracy of deep learning by comparing the performance of a Cox hazard regression and RNN-LSTM based on survival analysis.Methods and findingsWe selected 361,239 subjects (age 40 to 79 years) with more than two health examination records from 2002-2006 using the National Health Insurance System-National Health Screening Cohort (NHIS-HEALS). The average number of health screenings (from 2002-2013) used in the analysis was 2.9 ± 1.0. Two CVD prediction models were developed from the NHIS-HEALS data: a Cox hazard regression model and a deep learning model. In an internal validation of the NHIS-HEALS dataset, the Cox regression model showed a highest time-dependent area under the curve (AUC) of 0.79 (95% CI 0.70 to 0.87) for in females and 0.75 (95% CI 0.70 to 0.80) in males at 2 years. The deep learning model showed a highest time-dependent AUC of 0.94 (95% CI 0.91 to 0.97) for in females and 0.96 (95% CI 0.95 to 0.97) in males at 2 years. Layer-wise Relevance Propagation (LRP) revealed that age was the variable that had the greatest effect on CVD, followed by systolic blood pressure (SBP) and diastolic blood pressure (DBP), in that order.ConclusionThe performance of the deep learning model for predicting CVD occurrences was better than that of the Cox regression model. In addition, it was confirmed that the known risk factors shown to be important by previous clinical studies were extracted from the study results using LRP.

Conventional scoring and identification methods for coronary artery calcium (CAC) and aortic calcium (AC) result in information loss from the original image and can be time-consuming. In this study, we sought to demonstrate an end-to-end deep learning model as an alternative to the conventional methods. Scans of 377 patients with no history of coronary artery disease (CAD) were obtained and annotated. A deep learning model was trained, tested and validated in a 60:20:20 split. Within the cohort, mean age was 64.2 ± 9.8 years, and 33% were female. Left anterior descending, right coronary artery, left circumflex, triple vessel, and aortic calcifications were present in 74.87%, 55.82%, 57.41%, 46.03%, and 85.41% of patients respectively. An overall Dice score of 0.952 (interquartile range 0.921, 0.981) was achieved. Stratified by subgroups, there was no difference between male (0.948, interquartile range 0.920, 0.981) and female (0.965, interquartile range 0.933, 0.980) patients (p = 0.350), or, between age p = 0.742). There was good correlation and agreement for CAC prediction (rho = 0.876, p < 0.001), with a mean difference of 11.2% (p = 0.100). AC correlated well (rho = 0.947, p < 0.001), with a mean difference of 9% (p = 0.070). Automated segmentation took approximately 4 s per patient. Taken together, the deep-end learning model was able to robustly identify vessel-specific CAC and AC with high accuracy, and predict Agatston scores that correlated well with manual annotation, facilitating application into areas of research and clinical importance.

We propose a robust method to simultaneously localize multiple objects in cardiac computed tomography angiography (CTA) images. The relative prior distributions of the multiple objects in the three-dimensional (3D) space can be obtained through integrating the geometric morphological relationship of each target object to some reference objects. In cardiac CTA images, the cross-sections of ascending and descending aorta can play the role of the reference objects. We employed the maximum a posteriori (MAP) estimator that utilizes anatomic prior knowledge to address this problem of localizing multiple objects. We propose a new feature for each pixel using the relative distances, which can define any objects that have unclear boundaries. Our experimental results targeting four pulmonary veins (PVs) and the left atrial appendage (LAA) in cardiac CTA images demonstrate the robustness of the proposed method. The method could also be extended to localize other multiple objects in different applications.

Background and Purpose Cerebral small vessel disease (SVDs) are related with large artery atherosclerosis. However, the association between aortic atheroma (AA) and cerebral small vessel disease has rarely been reported. This study evaluated the relationship between presence and burden of AAs and those of SVDs in patients with acute ischemic stroke. Methods We included 737 consecutive patients who underwent transesophageal echocardiography (TEE) and brain magnetic resonance imaging (MRI) for evaluation of acute stroke. AA subtypes were classified as complex aortic plaque (CAP) and simple aortic plaque (SAP). Presence and burden of SVDs including cerebral microbleeds (CMBs), white matter hyperintensities (WMHs), perivascular spaces (PVSs), asymptomatic lacunar infarctions (ALIs), and total SVD score, were investigated. Results AA was found by TEE in 360 (48.8%) patients including 11.6% with CAP and 37.2% with SAP. One or more types of SVDs was found in 269 (36.4%) patients. In multivariable analysis, presence of CMBs (odds ratio [OR] 4.68), high-grade WMHs (OR 3.13), high-grade PVSs (OR 3.35), and ALIs (OR 4.24) were frequent in patients with AA than those without AA. Each 1-point increase in total SVD score increased the odds of presence of CAP (OR 1.94, 95% confidence interval (CI) 1.44-1.85) and SAP (OR 1.54, 95% CI 1.35-1.75). Conclusions In this study, patients with AA frequently had cerebral SVDs. Larger burden of AA was associated with advanced cerebral SVDs. Our findings give an additional information for positive relationship with systemic atherosclerosis and coexisting cerebral SVDs in acute ischemic stroke patients.

Bayesian network is an increasingly popular method in modeling uncertain and complex problems, because its interpretability is often more useful than plain prediction. To satisfy the core requirement in medical research to obtain interpretable prediction with high accuracy, we constructed an inference engine for post-stroke outcomes based on Bayesian network classifiers. The prediction system that was trained on data of 3,605 patients with acute stroke forecasts the functional independence at 3 months and the mortality 1 year after stroke. Feature selection methods were applied to eliminate less relevant and redundant features from 76 risk variables. The Bayesian network classifiers were trained with a hill-climbing searching for the qualified network structure and parameters measured by maximum description length. We evaluated and optimized the proposed system to increase the area under the receiver operating characteristic curve (AUC) while ensuring acceptable sensitivity for the class-imbalanced data. The performance evaluation demonstrated that the Bayesian network with selected features by wrapper-type feature selection can predict 3-month functional independence with an AUC of 0.889 using only 19 risk variables and 1-year mortality with an AUC of 0.893 using 24 variables. The Bayesian network with 50 features filtered by information gain can predict 3-month functional independence with an AUC of 0.875 and 1-year mortality with an AUC of 0.895. We also built an online prediction service, Yonsei Stroke Outcome Inference System, to substantialize the proposed solution for patients with stroke.

Epidemiologic data regarding pulmonary arterial hypertension (PAH) have relied on registries from Western countries. We assessed the current status of PAH in the Korean population. The Health Insurance Review and Assessment Service (HIRA) claim database, which comprises nationwide medical insurance data of Koreans from 2008-2016, was assessed to determine the current status of PAH. Overall, 1,307 patients were newly diagnosed with PAH from 2008-2016 (0.0005%, annual incidence: 4.84 patients/1 million people/year). The mean age at diagnosis was 44±13 years (range 18-65) and patients were mostly women (n = 906, 69.3%). Cases of idiopathic PAH (51.6%) accounted for the largest proportion, followed by acquired PAH (APAH) associated with congenital heart disease (25.8%) and APAH with connective tissue disease (17.2%). Overall, 807 (61.7%) patients received a single PAH-specific treatment based on their last prescription, of which bosentan (50.6%) was the most frequently used. Only 240 (18.4%) patients received combination therapy, with the bosentan-beraprost combination (32.9%) being the most common. During the mean follow-up of 1.9 years, the 1-, 2-, 3-, and 5-year estimated survival rates were 85%, 62%, 54%, and 46%, respectively. The prevalence and incidence of PAH in the Korean population is currently comparable with that in previous registries. The 5-year survival rate was slightly higher in the Korean population than previously reported.

BackgroundThe extent to which the presence and extent of subclinical atherosclerosis by coronary computed tomography angiography influences a potential mortality benefit of statin is unknown. We evaluated the relationship between statin therapy, mortality, and subclinical atherosclerosis.MethodsIn the CONFIRM study, patients with normal or non-obstructive plaque (Results1.2% of patients experienced all-cause mortality. Patients not on baseline statin therapy had a stepwise increased risk of all-cause mortality by CAC (relative to CAC = 0; CAC 1-99: hazard ratio [HR] 1.65, CAC 100-299: HR 2.19, and CAC≥300: HR 2.98) or SIS (relative to SIS = 0; SIS 1: HR 1.62, SIS 2-3: 2.48 and SIS≥4: 2.95). Conversely, in patients on baseline statin therapy, there was no significant increase in mortality risk with increasing CAC (p value for interaction = 0.049) or SIS (p value for interaction = 0.007). The incidence of MACE was 2.1%. Similar to the all-cause mortality, the risk of MACE was increased with CAC or SIS strata in patient not on baseline statin therapy. However, this relation was not observed in patient on baseline statin therapy.ConclusionIn individuals with non-obstructive coronary artery disease, increased risk of adverse events occurs with increasing CAC or SIS who are not on baseline statin therapy. Statin therapy is associated with a mitigation of risk of cardiac events in the presence of increasing atherosclerosis, with no particular threshold of disease burden.

OBJECTIVE:We sought to investigate determinants and prognosis of sinus node dysfunction (SND) after surgical ablation of atrial fibrillation (AF) with concomitant mitral valve (MV) surgery. A total of 202 patients who underwent surgical AF ablation with concomitant MV surgery were studied. STUDY DESIGN AND SETTING:SND was defined as electrocardiographic manifestations, such as junctional bradycardia, symptomatic sick sinus syndrome, or symptomatic sinus bradycardia, 7 days after surgery. Baseline clinical and echocardiographic characteristics, rhythm outcomes [AF recurrence or permanent pacemaker (PM) implantation] at 6 and 12 months, and clinical outcomes were compared between patients without SND (n = 165) and those with SND (n = 37) after surgery. RESULTS:Patients with SND showed a significantly larger left atrial volume index (LAVI) and a higher right ventricular systolic pressure than those without SND. In addition, there was a higher likelihood for AF recurrence and PM implantation in patients with SND than in those without SND. Although clinical outcomes did not differ between the two groups, patients with SND had a significantly longer length of hospital stay (p

The minimally invasive transcatheter aortic valve implantation (TAVI) is the most prevalent method to treat aortic valve stenosis. For pre-operative surgical planning, contrast-enhanced coronary CT angiography (CCTA) is used as the imaging technique to acquire 3-D measurements of the valve. Accurate localization of the eight aortic valve landmarks in CT images plays a vital role in the TAVI workflow because a small error risks blocking the coronary circulation. In order to examine the valve and mark the landmarks, physicians prefer a view parallel to the hinge plane, instead of using the conventional axial, coronal or sagittal view. However, customizing the view is a difficult and time-consuming task because of unclear aorta pose and different artifacts of CCTA. Therefore, automatic localization of landmarks can serve as a useful guide to the physicians customizing the viewpoint. In this paper, we present an automatic method to localize the aortic valve landmarks using colonial walk, a regression tree-based machine-learning algorithm. For efficient learning from the training set, we propose a two-phase optimized search space learning model in which a representative point inside the valvular area is first learned from the whole CT volume. All eight landmarks are then learned from a smaller area around that point. Experiment with preprocedural CCTA images of TAVI undergoing patients showed that our method is robust under high stenotic variation and notably efficient, as it requires only 12 milliseconds to localize all eight landmarks, as tested on a 3.60 GHz single-core CPU.