PAGE: Prototype-Based Model-Level Explanations for Graph Neural Networks.

Aside from graph neural networks (GNNs) attracting significant attention as a powerful framework revolutionizing graph representation learning, there has been an increasing demand for explaining GNN models. Although various explanation methods for GNNs have been developed, most studies have focused on instance-level explanations, which produce explanations tailored to a given graph instance. In our study, we propose Prototype-bAsed GNN-Explainer ([Formula: see text]), a novel model-level GNN explanation method that explains what the underlying GNN model has learned for graph classification by discovering human-interpretable prototype graphs. Our method produces explanations for a given class, thus being capable of offering more concise and comprehensive explanations than those of instance-level explanations. First, [Formula: see text] selects embeddings of class-discriminative input graphs on the graph-level embedding space after clustering them. Then, [Formula: see text] discovers a common subgraph pattern by iteratively searching for high matching node tuples using node-level embeddings via a prototype scoring function, thereby yielding a prototype graph as our explanation. Using six graph classification datasets, we demonstrate that [Formula: see text] qualitatively and quantitatively outperforms the state-of-the-art model-level explanation method. We also carry out ystematic experimental studies by demonstrating the relationship between [Formula: see text] and instance-level explanation methods, the robustness of [Formula: see text] to input data scarce environments, and the computational efficiency of the proposed prototype scoring function in [Formula: see text].

On the Power of Gradual Network Alignment Using Dual-Perception Similarities.

Network alignment (NA) is the task of finding the correspondence of nodes between two networks based on the network structure and node attributes. Our study is motivated by the fact that, since most of existing NA methods have attempted to discover all node pairs at once, they do not harness information enriched through interim discovery of node correspondences to more accurately find the next correspondences during the node matching. To tackle this challenge, we propose [Formula: see text], a new NA method that gradually discovers node pairs by making full use of node pairs exhibiting strong consistency, which are easy to be discovered in the early stage of gradual matching. Specifically, [Formula: see text] first generates node embeddings of the two networks based on graph neural networks along with our layer-wise reconstruction loss, a loss built upon capturing the first-order and higher-order neighborhood structures. Then, nodes are gradually aligned by computing dual-perception similarity measures including the multi-layer embedding similarity as well as the Tversky similarity, an asymmetric set similarity using the Tversky index applicable to networks with different scales. Additionally, we incorporate an edge augmentation module into [Formula: see text] to reinforce the structural consistency. Through comprehensive experiments using real-world and synthetic datasets, we empirically demonstrate that [Formula: see text] consistently outperforms state-of-the-art NA methods.

Continuous cuffless blood pressure monitoring using photoplethysmography-based PPG2BP-net for high intrasubject blood pressure variations.

Continuous, comfortable, convenient (C3), and accurate blood pressure (BP) measurement and monitoring are needed for early diagnosis of various cardiovascular diseases. To supplement the limited C3 BP measurement of existing cuff-based BP technologies, though they may achieve reliable accuracy, cuffless BP measurement technologies, such as pulse transit/arrival time, pulse wave analysis, and image processing, have been studied to obtain C3 BP measurement. One of the recent cuffless BP measurement technologies, innovative machine-learning and artificial intelligence-based technologies that can estimate BP by extracting BP-related features from photoplethysmography (PPG)-based waveforms have attracted interdisciplinary attention of the medical and computer scientists owing to their handiness and effectiveness for both C3 and accurate, i.e., C3A, BP measurement. However, C3A BP measurement remains still unattainable because the accuracy of the existing PPG-based BP methods was not sufficiently justified for subject-independent and highly varying BP, which is a typical case in practice. To circumvent this issue, a novel convolutional neural network(CNN)- and calibration-based model (PPG2BP-Net) was designed by using a comparative paired one-dimensional CNN structure to estimate highly varying intrasubject BP. To this end, approximately [Formula: see text], [Formula: see text], and [Formula: see text] of 4185 cleaned, independent subjects from 25,779 surgical cases were used for training, validating, and testing the proposed PPG2BP-Net, respectively and exclusively (i.e., subject-independent modelling). For quantifying the intrasubject BP variation from an initial calibration BP, a novel 'standard deviation of subject-calibration centring (SDS)' metric is proposed wherein high SDS represents high intrasubject BP variation from the calibration BP and vice versa. PPG2BP-Net achieved accurately estimated systolic and diastolic BP values despite high intrasubject variability. In 629-subject data acquired after 20 minutes following the A-line (arterial line) insertion, low error mean and standard deviation of [Formula: see text] and [Formula: see text] for highly varying A-line systolic and diastolic BP values, respectively, where their SDSs are 15.375 and 8.745. This study moves one step forward in developing the C3A cuffless BP estimation devices that enable the push and agile pull services.

Efficacy and safety of moderate-intensity statin with ezetimibe combination therapy in patients after percutaneous coronary intervention: a post-hoc analysis of the RACING trial.

Background: Moderate-intensity statin role with ezetimibe combination therapy following percutaneous coronary intervention (PCI) has not been thoroughly investigated, particularly compared to high-intensity statin monotherapy. We aimed to investigate the effect of ezetimibe combination with moderate-intensity statin in patients with atherosclerotic cardiovascular disease following PCI. Methods: This was a post-hoc analysis of a subset of patients who underwent PCI in the RACING trial. At 26 centres in South Korea, patients with atherosclerotic cardiovascular disease (ASCVD) were randomly assigned to receive either moderate-intensity statin with ezetimibe combination therapy (rosuvastatin 10 mg with ezetimibe 10 mg) or high-intensity statin monotherapy (rosuvastatin 20 mg). The prespecified endpoints of the RACING trial were used. The primary endpoint was the 3-year composite of cardiovascular death, major cardiovascular events, and nonfatal stroke. Event rates between the two groups were compared using log-rank tests, and hazard ratios (HR) with 95% confidence intervals (CI) were estimated using Cox regression analysis. Consistent with the RACING trial, the primary and secondary efficacy endpoints were evaluated using an intention-to-treatment approach, and the safety endpoints were assessed in the safety population. The RACING trial was registered at ClinicalTrials.gov (NCT03044665). Findings: Between Feb 14, 2017, and Dec 18, 2018, 3780 participants were enrolled in the RACING trial. Prior history of PCI was found in 2497 patients (67%, median 64 years, 79% male), and was associated with higher rates of the primary endpoint (hazard ratio [HR], 1.34; 95% confidence interval [CI], 1.06-1.69; p = 0.014). Among patients with prior PCI, moderate-intensity statin therapy with ezetimibe combination versus high-intensity statin therapy did not increase the risk of the primary endpoint (HR, 0.95; 95% CI, 0.74-1.24; p = 0.781). The proportion of patients with low-density lipoprotein cholesterol (LDL-C) <70 mg/dL at 1, 2, and 3 years was 74%, 76%, and 73%, respectively, in the combination therapy group, and was significantly higher than that in the high-intensity statin monotherapy group (57%, 62%, and 59%, respectively, all p < 0.001). Discontinuation of lipid-lowering drugs occurred less frequently in the combination group (4.2% vs. 7.6%, p = 0.001). Interpretation: The effects of ezetimibe combination therapy observed in the RACING trial were consistently preserved among patients with ASCVD following PCI. Ezetimibe combination could be considered as a suitable therapeutic strategy to achieve strict control of LDL-C and reduce drug intolerance in patients who underwent PCI. Funding: Hanmi Pharmaceutical, Seoul, South Korea.

Effectiveness and Safety of a Fixed-Dose Combination of Valsartan and Rosuvastatin (Rovatitan® Tablet) in Patients with Concomitant Hypertension and Hyperlipidemia: An Observational Study.

Purpose: This study aimed to assess the effectiveness and safety of a fixed-dose combination of rosuvastatin and valsartan (Rovatitan®) in Korean patients with concomitant hypertension and hyperlipidemia. Patients and Methods: A total of 1008 eligible patients with concomitant hypertension and hyperlipidemia were enrolled and treated for 12 weeks. Both upward and downward drug dose titrations were allowed based on the investigator's discretion. This study evaluated the effectiveness of the study drug, defined by the percentage of patients achieving the blood pressure (BP) and low-density lipoprotein cholesterol (LDL-C) treatment targets. Additionally, regression analyses were conducted to evaluate the factors associated with the effectiveness and safety of the study drug. Of the 1008 patients enrolled in the study, 911 were analyzed for clinical effectiveness. Results: At 12 weeks, 84.6% and 75.9% of patients treated with the study drug achieved their BP and LDL-C targets, respectively, and 64.8% of patients achieved both targets simultaneously. Furthermore, the percentage of patients who achieved their BP and LDL-C treatment targets demonstrated a trend across the respective risk groups; the higher the risk group, the lower the success of attaining the respective target. This trend was also observed regardless of the prior antihypertensive and/or lipid-lowering treatments. According to regression analysis, poor metabolic profiles, including a higher body mass index (BMI) and higher BP and LDL-C levels at baseline, were significantly associated with treatment failure for BP. Among the 1005 patients included in the safety analysis, 17 patients (1.7%) experienced serious adverse events; however, none were considered related to the study drug. Conclusion: The study drug used for the treatment of concomitant hypertension and hyperlipidemia in a real-world setting was effective and was well tolerated. Therefore, the study drug is suggested as a good alternative to increase patient convenience and compliance, particularly in those taking multiple medications.

Aspirin vs Clopidogrel for Long-term Maintenance After Coronary Stenting in Patients With Diabetes: A Post Hoc Analysis of the HOST-EXAM Trial.

Importance: Selecting the optimal antiplatelet agent in patients who have received percutaneous coronary intervention is especially important in those with diabetes due to the heightened risk of ischemic events in this population. Studies on the efficacy and safety of clopidogrel vs aspirin for long-term maintenance after percutaneous coronary intervention in patients with diabetes are lacking. Objective: To investigate cardiovascular outcomes with clopidogrel vs aspirin in patients with and without diabetes. Design, Setting, and Participants: This was a post hoc analysis of the HOST-EXAM randomized clinical trial, an investigator-initiated, prospective, randomized, open-label, multicenter trial performed at 37 centers in Korea. Patients who received dual antiplatelet therapy without clinical events for 6 to 18 months after percutaneous coronary intervention with drug-eluting stents were enrolled from March 2014 to May 2018 with follow-up at 6, 12, 18, and 24 months. All 5438 patients in the original trial were included in this analysis, which was conducted from June to October 2021. Interventions and Exposures: Enrolled patients were randomized 1:1 to clopidogrel or aspirin monotherapy. Subgroup analyses were performed by the presence of diabetes. Main Outcomes and Measures: The main outcome was primary composite end point of all-cause death, nonfatal myocardial infarction, stroke, readmission due to acute coronary syndrome, and major bleeding (Bleeding Academic Research Consortium type 3 or 5) at 24-month follow-up. Results: Of 5438 patients (mean [SD] age, 63.5 [10.7] years; 1384 [25.5%] female), 1860 (34.2%) had diabetes (925 in the clopidogrel arm and 935 in the aspirin arm), and 5338 (98.2%) completed follow-up. The rate of the primary composite end point was significantly lower in the clopidogrel group compared to the aspirin group in patients with diabetes (6.3% vs 9.2%; hazard ratio [HR], 0.69; 95% CI, 0.49-0.96; P = .03; absolute risk difference [ARD], 2.7%; number needed to treat [NNT], 37) and without diabetes (5.3% vs 7.0%; HR, 0.76; 95% CI, 0.58-1.00; P = .046; ARD, 1.6%, NNT, 63; P for interaction = .65). The presence of diabetes was not associated with a difference in benefit observed with clopidogrel monotherapy over aspirin for the thrombotic composite end point (HR, 0.68; 95% CI, 0.45-1.04 for patients with diabetes vs HR, 0.68; 95% CI, 0.49-0.93 for those without; P for interaction = .99) and any bleeding with Bleeding Academic Research Consortium 2, 3, or 5 (HR, 0.65; 95% CI, 0.39-1.09 for patients with diabetes vs HR, 0.74; 95% CI, 0.48-1.13 for those without; P for interaction = .71). Conclusion and Relevance: In this study, clopidogrel monotherapy was associated with a lower rate of the primary composite end point compared to aspirin monotherapy as long-term maintenance therapy after dual antiplatelet therapy for coronary stenting in both patients with and without diabetes. Clopidogrel might thus be considered rather than aspirin in patients who have undergone coronary stenting and successfully completed dual antiplatelet therapy, regardless of diabetes status. Trial Registration: ClinicalTrials.gov Identifier: NCT02044250.

Assessment of fractional flow reserve in intermediate coronary stenosis using optical coherence tomography-based machine learning.

Objectives: This study aimed to evaluate and compare the diagnostic accuracy of machine learning (ML)- fractional flow reserve (FFR) based on optical coherence tomography (OCT) with wire-based FFR irrespective of the coronary territory. Background: ML techniques for assessing hemodynamics features including FFR in coronary artery disease have been developed based on various imaging modalities. However, there is no study using OCT-based ML models for all coronary artery territories. Methods: OCT and FFR data were obtained for 356 individual coronary lesions in 130 patients. The training and testing groups were divided in a ratio of 4:1. The ML-FFR was derived for the testing group and compared with the wire-based FFR in terms of the diagnosis of ischemia (FFR ≤ 0.80). Results: The mean age of the subjects was 62.6 years. The numbers of the left anterior descending, left circumflex, and right coronary arteries were 130 (36.5%), 110 (30.9%), and 116 (32.6%), respectively. Using seven major features, the ML-FFR showed strong correlation (r = 0.8782, P < 0.001) with the wire-based FFR. The ML-FFR predicted wire-based FFR ≤ 0.80 in the test set with sensitivity of 98.3%, specificity of 61.5%, and overall accuracy of 91.7% (area under the curve: 0.948). External validation showed good correlation (r = 0.7884, P < 0.001) and accuracy of 83.2% (area under the curve: 0.912). Conclusion: OCT-based ML-FFR showed good diagnostic performance in predicting FFR irrespective of the coronary territory. Because the study was a small-size study, the results should be warranted the performance in further large-scale research.

Prasugrel dose de-escalation in diabetic patients with acute coronary syndrome receiving percutaneous coronary intervention: results from the HOST-REDUCE-POLYTECH-ACS trial.

AIMS: The aim of this study was to evaluate the efficacy and safety of prasugrel dose de-escalation therapy in patients with diabetes mellitus (DM)-acute coronary syndrome (ACS) who underwent percutaneous coronary intervention (PCI). METHODS AND RESULTS: This was a post-hoc analysis of the HOST-REDUCE-POLYTECH-ACS (Harmonizing Optimal Strategy for Treatment of Coronary Artery Diseases-Comparison of Reduction of Prasugrel Dose or Polymer Technology in ACS Patients) randomized trial. The efficacy and safety of prasugrel dose de-escalation therapy (prasugrel 5 mg daily) were compared with conventional therapy (prasugrel 10 mg daily) in patients with DM. The primary endpoint was net adverse clinical events (NACE), defined as a composite of all-cause death, non-fatal myocardial infarction (MI), stent thrombosis (ST), clinically driven revascularization, stroke, and Bleeding Academic Research Consortium (BARC) class ≥2 bleeding events. The secondary ischaemic outcome was major adverse cardiovascular and cerebrovascular events, defined as the composite of cardiac death, non-fatal MI, ST, or ischaemic stroke. Of 2338 patients randomized, 990 had DM. The primary endpoint of NACE occurred in 38 patients (7.6%) receiving prasugrel dose de-escalation and in 53 patients (11.3%) receiving conventional therapy among patients with DM [hazard ratio (HR) 0.66; 95% confidence interval (CI) 0.43-0.99; P = 0.049]. Prasugrel dose de-escalation as compared with conventional therapy did not increase the risk of ischaemic events (HR 1.03; 95% CI 0.56-1.88; P = 0.927) but decreased BARC class ≥2 bleeding in patients with DM (HR 0.44; 95% CI 0.23-0.84; P = 0.012). CONCLUSION: Prasugrel dose de-escalation compared with conventional therapy may reduce the risk of net clinical outcomes, mostly driven by a reduction in bleeding without an increase in ischaemic events in patients with DM. Trial Registration: HOST-REDUCE-POLYTECH-ACS, NCT02193971, https://clinicaltrials.gov/ct2/show/NCT02193971.

Moderate-intensity statin with ezetimibe vs. high-intensity statin in patients with diabetes and atherosclerotic cardiovascular disease in the RACING trial.

AIMS: This study evaluated the effect of moderate-intensity statin with ezetimibe combination therapy vs. high-intensity statin monotherapy among patients with diabetes mellitus (DM) and atherosclerotic cardiovascular disease (ASCVD). METHODS AND RESULTS: This was a pre-specified, stratified subgroup analysis of the DM cohort in the RACING trial. The primary outcome was a 3-year composite of cardiovascular death, major cardiovascular events, or non-fatal stroke. Among total patients, 1398 (37.0%) had DM at baseline. The incidence of the primary outcome was 10.0% and 11.3% among patients with DM randomized to ezetimibe combination therapy vs. high-intensity statin monotherapy (hazard ratio: 0.89; 95% confidence interval: 0.64-1.22; P = 0.460). Intolerance-related discontinuation or dose reduction of the study drug was observed in 5.2% and 8.7% of patients in each group, respectively (P = 0.014). LDL cholesterol levels <70 mg/dL at 1, 2, and 3 years were observed in 81.0%, 83.1%, and 79.9% of patients in the ezetimibe combination therapy group, and 64.1%, 70.2%, and 66.8% of patients in the high-intensity statin monotherapy group (all P < 0.001). In the total population, no significant interactions were found between DM status and therapy regarding primary outcome, intolerance-related discontinuation or dose reduction, and the proportion of patients with LDL cholesterol levels <70 mg/dL. CONCLUSION: Ezetimibe combination therapy effects observed in the RACING trial population are preserved among patients with DM. This study supports moderate-intensity statin with ezetimibe combination therapy as a suitable alternative to high-intensity statins if the latter cannot be tolerated, or further reduction in LDL cholesterol is required among patients with DM and ASCVD. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov, Identifier:NCT03044665.

Durable polymer versus biodegradable polymer drug-eluting stents in patients with acute coronary syndrome undergoing complex percutaneous coronary intervention: a post hoc analysis of the HOST-REDUCE-POLYTECH-ACS trial.

BACKGROUND: Comparative data of durable polymer (DP) versus biodegradable polymer (BP) drug-eluting stents (DES) are limited in patients presenting with acute coronary syndrome (ACS) undergoing complex percutaneous coronary intervention (PCI). AIMS: We sought to evaluate the efficacy and safety of DP-DES and BP-DES in ACS patients receiving complex PCI. METHODS: This study was a post hoc analysis of the HOST-REDUCE-POLYTECH-ACS trial. ACS patients were randomly assigned 1:1 to DP-DES or BP-DES in the HOST-REDUCE-POLYTECH-ACS trial. Complex PCI was defined as having at least 1 of the following features: ≥3 stents implanted, ≥3 lesions treated, total stent length ≥60 mm, bifurcation PCI with 2 stents, left main PCI, or heavy calcification. Patient-oriented (POCO, a composite of all-cause death, non-fatal myocardial infarction, and any repeat revascularisation) and device-oriented composite outcomes (DOCO, a composite of cardiac death, target vessel myocardial infarction, or target lesion revascularisation) were evaluated at 12 months. RESULTS: Among 3,301 patients for whom full procedural data were available, 1,140 patients received complex PCI. Complex PCI was associated with higher risks of POCO and DOCO. The risks of POCO were comparable between DP-DES and BP-DES in both the complex (HR 0.87, 95% confidence interval [CI]: 0.57-1.33; p=0.522) and non-complex (HR 0.83, 95% CI: 0.56-1.24; p=0.368; p for interaction=0.884) PCI groups. DOCO was also not significantly different between DP-DES and BP-DES in both the complex (HR 0.74, 95% CI: 0.43-1.27; p=0.278) and non-complex (HR 0.67, 95% CI: 0.38-1.19; p=0.175; p for interaction=0.814) PCI groups. CONCLUSIONS: In ACS patients, DP-DES and BP-DES showed similar clinical outcomes irrespective of PCI complexity.

SiReN: Sign-Aware Recommendation Using Graph Neural Networks.

In recent years, many recommender systems using network embedding (NE) such as graph neural networks (GNNs) have been extensively studied in the sense of improving recommendation accuracy. However, such attempts have focused mostly on utilizing only the information of positive user-item interactions with high ratings. Thus, there is a challenge on how to make use of low rating scores for representing users' preferences since low ratings can be still informative in designing NE-based recommender systems. In this study, we present SiReN, a new Sign-aware Recommender system based on GNN models. Specifically, SiReN has three key components: 1) constructing a signed bipartite graph for more precisely representing users' preferences, which is split into two edge-disjoint graphs with positive and negative edges each; 2) generating two embeddings for the partitioned graphs with positive and negative edges via a GNN model and a multilayer perceptron (MLP), respectively, and then using an attention model to obtain the final embeddings; and 3) establishing a sign-aware Bayesian personalized ranking (BPR) loss function in the process of optimization. Through comprehensive experiments, we empirically demonstrate that SiReN consistently outperforms state-of-the-art NE-aided recommendation methods.

Machine learning approaches to predicting no-shows in pediatric medical appointment.

Patients' no-shows, scheduled but unattended medical appointments, have a direct negative impact on patients' health, due to discontinuity of treatment and late presentation to care. They also lead to inefficient use of medical resources in hospitals and clinics. The ability to predict a likely no-show in advance could enable the design and implementation of interventions to reduce the risk of it happening, thus improving patients' care and clinical resource allocation. In this study, we develop a new interpretable deep learning-based approach for predicting the risk of no-shows at the time when a medical appointment is first scheduled. The retrospective study was conducted in an academic pediatric teaching hospital with a 20% no-show rate. Our approach tackles several challenges in the design of a predictive model by (1) adopting a data imputation method for patients with missing information in their records (77% of the population), (2) exploiting local weather information to improve predictive accuracy, and (3) developing an interpretable approach that explains how a prediction is made for each individual patient. Our proposed neural network-based and logistic regression-based methods outperformed persistence baselines. In an unobserved set of patients, our method correctly identified 83% of no-shows at the time of scheduling and led to a false alert rate less than 17%. Our method is capable of producing meaningful predictions even when some information in a patient's records is missing. We find that patients' past no-show record is the strongest predictor. Finally, we discuss several potential interventions to reduce no-shows, such as scheduling appointments of high-risk patients at off-peak times, which can serve as starting point for further studies on no-show interventions.

A Paradigm Shift in Dyslipidemia Management in Primary Care: A 12-Month Cohort Study.

PURPOSE: LDL-lowering therapy is beneficial to reduce the risk of cardiovascular disease (CVD). Higher statin doses lower LDL-C levels and prevent CVD; however, they increase adverse events, such as muscle-related adverse events and new-onset diabetes mellitus (DM). Ezetimibe combined with statin therapy improves LDL-C-lowering levels and tolerability in patients with established CVD. We aimed to analyze the efficacy and safety of a fixed-dose rosuvastatin and ezetimibe (R+E) combination therapy in intermediate-risk patients with hypercholesterolemia and no DM after 12 months of visiting a primary physician. METHODS: This multicenter, open-label, single-arm, prospective observational study involved 5717 patients from 258 primary health care centers in Korea enrolled between 2016 and 2018. Patients had no DM or previous CVD but had cardiovascular risk factors and were taking a statin or a fixed-dose combination of E (10 mg) + R (5, 10, or 20 mg). We analyzed 700 patients using propensity score matching. FINDINGS: A fixed-dose R+E combination therapy significantly reduced LDL-C in 5/10 mg R+E (29.35%), 10/10 mg R+E (36.19%), and 20/10 mg R+E (41.83%) compared with statin monotherapy (19.09%) at 12-month follow-up (P = 0.017). Compared with statin monotherapy, HDL-C levels increased in 5/10 mg R+E (mean change at 12 months; P = 0.004), and triglyceride levels decreased in 10/10 mg R+E (mean change at 12 months; P = 0.033). The fixed-dose R+E combination therapy was associated with fewer adverse events and a neutral effect on glucose deterioration compared with statin monotherapy at 12 months of follow-up. IMPLICATIONS: In a possible paradigm shift, a fixed-dose R+E combination therapy may be beneficial for primary cardiovascular prevention with potent LDL-lowering efficacy and tolerability; however, further large prospective studies are needed.

Explainable gait recognition with prototyping encoder-decoder.

Human gait is a unique behavioral characteristic that can be used to recognize individuals. Collecting gait information widely by the means of wearable devices and recognizing people by the data has become a topic of research. While most prior studies collected gait information using inertial measurement units, we gather the data from 40 people using insoles, including pressure sensors, and precisely identify the gait phases from the long time series using the pressure data. In terms of recognizing people, there have been a few recent studies on neural network-based approaches for solving the open set gait recognition problem using wearable devices. Typically, these approaches determine decision boundaries in the latent space with a limited number of samples. Motivated by the fact that such methods are sensitive to the values of hyper-parameters, as our first contribution, we propose a new network model that is less sensitive to changes in the values using a new prototyping encoder-decoder network architecture. As our second contribution, to overcome the inherent limitations due to the lack of transparency and interpretability of neural networks, we propose a new module that enables us to analyze which part of the input is relevant to the overall recognition performance using explainable tools such as sensitivity analysis (SA) and layer-wise relevance propagation (LRP).

[Formula: see text]: Deep Generative Network Completion.

Most network data are collected from partially observable networks with both missing nodes and missing edges, for example, due to limited resources and privacy settings specified by users on social media. Thus, it stands to reason that inferring the missing parts of the networks by performing network completion should precede downstream applications. However, despite this need, the recovery of missing nodes and edges in such incomplete networks is an insufficiently explored problem due to the modeling difficulty, which is much more challenging than link prediction that only infers missing edges. In this paper, we present DeepNC, a novel method for inferring the missing parts of a network based on a deep generative model of graphs. Specifically, our method first learns a likelihood over edges via an autoregressive generative model, and then identifies the graph that maximizes the learned likelihood conditioned on the observable graph topology. Moreover, we propose a computationally efficient [Formula: see text] algorithm that consecutively finds individual nodes that maximize the probability in each node generation step, as well as an enhanced version using the expectation-maximization algorithm. The runtime complexities of both algorithms are shown to be almost linear in the number of nodes in the network. We empirically demonstrate the superiority of DeepNC over state-of-the-art network completion approaches.

Durable Polymer Versus Biodegradable Polymer Drug-Eluting Stents After Percutaneous Coronary Intervention in Patients with Acute Coronary Syndrome: The HOST-REDUCE-POLYTECH-ACS Trial.

BACKGROUND: Large-scale randomized comparison of drug-eluting stents (DES) based on durable polymer versus biodegradable polymer technology is currently insufficient in patients with acute coronary syndrome (ACS). The present study aimed to prove the noninferiority of the durable polymer DES (DP-DES) compared with the biodegradable polymer DES (BP-DES) in such patients. METHODS: The HOST-REDUCE-POLYTECH-ACS (Harmonizing Optimal Strategy for Treatment of Coronary Artery Diseases-Comparison of Reduction of Prasugrel Dose or Polymer Technology in ACS Patients) trial is an investigator-initiated, randomized, open-label, adjudicator-blinded, multicenter, noninferiority trial comparing the efficacy and safety of DP-DES and BP-DES in patients with ACS. The primary end point was a patient-oriented composite outcome (a composite of all-cause death, nonfatal myocardial infarction, and any repeat revascularization) at 12 months. The key secondary end point was device-oriented composite outcome (a composite of cardiac death, target-vessel myocardial infarction, or target lesion revascularization) at 12 months. RESULTS: A total of 3413 patients were randomized to receive the DP-DES (1713 patients) and BP-DES (1700 patients). At 12 months, patient-oriented composite outcome occurred in 5.2% in the DP-DES group and 6.4% in the BP-DES group (absolute risk difference, -1.2%; Pnoninferiority<0.001). The key secondary end point, device-oriented composite outcome, occurred less frequently in the DP-DES group (DP-DES vs BP-DES, 2.6% vs 3.9%; hazard ratio, 0.67 [95% CI, 0.46-0.98]; P=0.038), mostly because of a reduction in target lesion revascularization. The rate of spontaneous nonfatal myocardial infarction and stent thrombosis were extremely low, with no significant difference between the 2 groups (0.6% versus 0.8%; P=0.513 and 0.1% versus 0.4%; P=0.174, respectively). CONCLUSIONS: In ACS patients receiving percutaneous coronary intervention, DP-DES was noninferior to BP-DES with regard to patient-oriented composite outcomes at 12 months after index percutaneous coronary intervention. Registration: URL: https://wwwclinicaltrials.gov; Unique identifier: NCT02193971.

Revisiting Information Detection and Energy Harvesting: A Power Splitting-Based Approach.

Wireless sensors are becoming essential in machine-type communications and Internet of Things. As the key performance metrics, the spectral efficiency as well as the energy efficiency have been considered while determining the effectiveness of sensor networks. In this paper, we present several power-splitting solutions to maximize the average harvested energy under a rate constraint when both the information and power are transmitted through the same wireless channel to a sensor (i.e., a receiver). More specifically, we first designed the optimal dynamic power-splitting policy, which decides the optimal fractional power of the received signal used for energy harvesting at the receiver. As effective solutions, we proposed two types of single-threshold-based power-splitting policies, namely, Policies I and II, which decide to switch between energy harvesting and information decoding by comparing the received signal power with some given thresholds. Additionally, we performed asymptotic analysis for a large number of packets along with practical statistics-based policies. Consequently, we demonstrated the effectiveness of the proposed power-splitting solutions in terms of the rate-energy trade-off.

Efficacy and Safety of Omega-3 Fatty Acids in Patients Treated with Statins for Residual Hypertriglyceridemia: A Randomized, Double-Blind, Placebo-Controlled Clinical Trial.

BACKGROUND: Cardiovascular risk remains increased despite optimal low density lipoprotein cholesterol (LDL-C) level induced by intensive statin therapy. Therefore, recent guidelines recommend non-high density lipoprotein cholesterol (non-HDL-C) as a secondary target for preventing cardiovascular events. The aim of this study was to assess the efficacy and tolerability of omega-3 fatty acids (OM3-FAs) in combination with atorvastatin compared to atorvastatin alone in patients with mixed dyslipidemia. METHODS: This randomized, double-blind, placebo-controlled, parallel-group, and phase III multicenter study included adults with fasting triglyceride (TG) levels ≥200 and <500 mg/dL and LDL-C levels <110 mg/dL. Eligible subjects were randomized to ATOMEGA (OM3-FAs 4,000 mg plus atorvastatin calcium 20 mg) or atorvastatin 20 mg plus placebo groups. The primary efficacy endpoints were the percent changes in TG and non-HDL-C levels from baseline at the end of treatment. RESULTS: After 8 weeks of treatment, the percent changes from baseline in TG (-29.8% vs. 3.6%, P<0.001) and non-HDL-C (-10.1% vs. 4.9%, P<0.001) levels were significantly greater in the ATOMEGA group (n=97) than in the atorvastatin group (n=103). Moreover, the proportion of total subjects reaching TG target of <200 mg/dL in the ATOMEGA group was significantly higher than that in the atorvastatin group (62.9% vs. 22.3%, P<0.001). The incidence of adverse events did not differ between the two groups. CONCLUSION: The addition of OM3-FAs to atorvastatin improved TG and non-HDL-C levels to a significant extent compared to atorvastatin alone in subjects with residual hypertriglyceridemia.

In-hospital outcome differences between transradial and transfemoral coronary approaches: Data from the Korean percutaneous coronary intervention registry.

OBJECTIVES: We aimed to investigate specific subgroups in which the benefit of transradial coronary interventions (TRIs) would be enhanced. BACKGROUND: The advantage of TRIs over transfemoral coronary interventions (TFIs) might differ according to a given clinical condition, urgency of the procedure, and operator volume pattern. METHODS: Using a cohort from the 2014 Korean Percutaneous Coronary Intervention Registry, in-hospital outcomes of the TRI group (n = 22,993) were matched to those of the TFI group (n = 15,581). After propensity score matching, the composite endpoints between the groups and subgroups for all-cause death, nonfatal myocardial infarctions (MIs), or transfusions were analyzed. RESULTS: The composite endpoints occurred less frequently in the TRI group than the TFI group [2.1% vs. 5.5%, OR 0.63, 95% CI 0.55-0.72]. The TRI group had a lower rate of death (OR 0.44, 95% CI 0.33-0.60) and nonfatal MI (OR 0.66, 95% CI 0.54-0.81) than the TFI group. The TRI group required fewer transfusions than the TFI group (OR 0.72, 95% CI 0.59-0.88). TRI benefits were consistent across subgroups except patients with chronic kidney disease and those treated in low tertile PCI volume centers. The favorable outcome of TRI was greater in the elderly (≥75 years), patients with ST-elevation MI, those who underwent emergent PCI, and those treated in high tertile PCI volume hospitals (P for the interaction <0.001 for all). CONCLUSIONS: Compared to TFI, TRI had favorable composite in-hospital outcomes. TRI benefits were pronounced in high-risk clinical settings and in high PCI volume centers.

Support Vector Machine-Based Transmit Antenna Allocation for Multiuser Communication Systems.

In this paper, a support vector machine (SVM) technique has been applied to an antenna allocation system with multiple antennas in multiuser downlink communications. Here, only the channel magnitude information is available at the transmitter. Thus, a subset of transmit antennas that can reduce multiuser interference is selected based on such partial channel state information to support multiple users. For training, we generate the feature vectors by fully utilizing the characteristics of the interference-limited setup in the multiuser downlink system and determine the corresponding class label by evaluating a key performance indicator, i.e., sum rate in multiuser communications. Using test channels, we evaluate the performance of our antenna allocation system invoking the SVM-based allocation and optimization-based allocation, in terms of sum-rate performance and computational complexity. Rigorous testing allowed for a comparison of a SVM algorithm design between one-vs-one (OVO) and one-vs-all (OVA) strategies and a kernel function: (i) OVA is preferable to OVO since OVA can achieve almost the same sum rate as OVO with significantly reduced computational complexity, (ii) a Gaussian function is a good choice as the kernel function for the SVM, and (iii) the variance (kernel scale) and penalty parameter (box constraint) of an SVM kernel function are determined by 21.56 and 7.67, respectively. Further simulation results revealed that the designed SVM-based approach can remarkably reduce the time complexity compared to a traditional optimization-based approach, at the cost of marginal sum rate degradation. Our proposed framework offers some important insights for intelligently combining machine learning techniques and multiuser wireless communications.