Investigation of the predictive value of a novel algorithm based on coronary CT angiography regarding fractional flow reserve and revascularization in patients with stable coronary artery disease.

Fractional flow reserve (FFR) has been established as a gold standard for functional coronary ischemia. At present, the FFR can be calculated from coronary computed tomography angiography (CCTA) images (CT-FFR). Previous studies have suggested that CT-FFR outperforms CCTA and invasive coronary angiography (ICA) in determining hemodynamic significance of stenoses. Recently, a novel automatical algorithm of CT-FFR called RuiXin-FFR has been developed. The present study is designed to investigate the predictive value of this algorithm and its value in therapeutic decision making. The present study retrospectively included 166 patients with stable coronary artery disease (CAD) who underwent CCTA screening and diagnostic ICA examination at Peking University People's Hospital, in 73 of whom wire-derived FFR was also measured. CT-FFR analyses were performed with a dedicated software. All patients were followed up for at least 1 year. We validated the accuracy of RuiXin-FFR with invasive FFR as the standard of reference, and investigated the role of RuiXin-FFR in predicting treatment strategy and long-term prognosis. The mean age of the patients was 63.3 years with 63.9% male. The CT-FFR showed a moderate correlation with wire-derived FFR (r = 0.542, p < 0.0001) and diagnostic accuracy of 87.6% to predict myocardial ischemia (AUC: 0.839, 95% CI 0.728-0.950), which was significantly higher than CCTA and ICA. In the multivariate logistic regression analysis, CT-FFR ≤ 0.80 was an independent predictor of undergoing coronary revascularization (OR: 45.54, 95% CI 12.03-172.38, p < 0.0001), whereas CT-FFR > 0.80 was an independent predictor of non-obstructive CAD (OR: 14.67, 95% CI 5.42-39.72, p < 0.0001). Reserving ICA and revascularization for vessels with positive CT-FFR could have reduced the rate of ICA by 29.6%, lowered the rate of ICA in vessels without stenosis > 50% by 11.7%, and increased the rate of revascularization in patients receiving ICA by 21.2%. The average follow-up was 23.7 months, and major adverse cardiovascular events (MACE) occurred in 11 patients. The rate of MACE was significantly lower in patients with CT-FFR > 0.80. The new algorithm of CT-FFR can be used to predict the invasive FFR. The RuiXin-FFR can also provide useful information for the screening of patients in whom further ICA is indeed needed and prognosis evaluation.

Self-Assembly and Photothermal Conversion of MetallaRussian Doll and Metalla[2]catenanes Induced via Multiple Stacking Interactions.

A variety of organometallic supramolecular architectures have been constructed over the past decades and their properties were also explored via different strategies. However, the synthesis of metalla-Russian doll is still a fascinating challenge. Herein, a series of new coordination supramolecular complexes, including a metalla-Russian doll, metalla[2]catenanes, and metallarectangles, were synthesized by using meticulously selected Cp*Rh (Cp* = η5-C5Me5) building units (E1, E2, and E3) and three rigid anthracylpyridine ligands (L1, L2, and L3) via a self-assembly strategy. While the combination of the short ligand L1 and E1 or E2 generated two metallarectangles, the longer ligand L2 containing an alkynyl group resulted in two new [2]catenanes, most likely due to which the strong electron-donating effect of alkynyl groups causes self-accumulation. Interestingly, an unusual Russian doll assembly was obtained through the reaction of L3 and E3 based on sextuple π···π stacking interactions. Furthermore, the dynamic structural conversion between [2]catenanes and the corresponding metallarectangles could be observed through concentration-, solvent-, and guest-induced effects. The [2]catenane complexes 4b displayed efficient photothermal conversion efficiency in solution (20.2%), in comparison with other organometallic macrocycles. We believe that π···π stacking interactions generate active nonradiative pathways and promote radiative photodeactivation pathways. This study proves the versatility of half-sandwich building units, not only to build complicated supramolecular topologies but also in effective functional materials for various appealing applications.

DGCL: Distance-wise and Graph Contrastive Learning for medication recommendation.

Medicine recommendation aims to provide a combination of medicine based on the patient's electronic health record (EHR), which is an essential task in healthcare. Existing methods either base recommendations on EHRs or provide models with knowledge of drug-drug interactions (DDIs) to achieve DDI reduction. However, the former models the patient's health history but ignores undesirable DDIs, while the latter lacks mining of patient health records and gets low recommendation accuracy. Therefore, this study contributes to research on personalized medication recommendations that consider drug interaction effects and models the patient's past medical history. In this paper, the Distance-wise and Graph Contrastive Learning (DGCL) framework is proposed. Specifically, we develop a two-stage neural network module for clinical record learning. We propose the distance detection loss to model the difference between the output distribution of current cases and historical records. In the DDI recognition and control task, DGCL proposes a graph contrastive learning method to jointly train the DDI knowledge graph and the electronic record graph, thereby effectively controlling the level of DDI for recommended medications. By comparing the performance on the MIMIC-III dataset with several baselines, DGCL outperforms other models in terms of efficacy and safety.

The Coronary Angiography-Derived Index of Microcirculatory Resistance Predicts Left Ventricular Performance Recovery in Patients with ST-Segment Elevation Myocardial Infarction.

Objectives: The present study is designed to investigate the impact of coronary angiography-derived index of microcirculatory resistance (caIMR) on left ventricular performance recovery. Background: IMR has been established as a gold standard for coronary microvascular assessment and a predictor of left ventricular recovery after ST-segment elevation myocardial infarction (STEMI). CaIMR is a novel and accurate alternative of IMR. Methods: The present study retrospectively included 80 patients with STEMI who underwent primary percutaneous coronary intervention (PCI). We offline performed the post-PCI caIMR analysis of the culprit vessel. Echocardiography was performed within the first 24 hours and at 3 months after the index procedure. Left ventricular recovery was defined as the change in left ventricular ejection fraction (LVEF) more than zero. Results: The mean age of the patients was 58.0 years with 80.0% male. The average post-PCI caIMR was 43.2. Overall left ventricular recovery was seen in 41 patients. Post-PCI caIMR (OR: 0.948, 95% CI: 0.916-0.981, p = 0.002), left anterior descending as the culprit vessel (OR: 3.605, 95% CI: 1.23-10.567, p = 0.019), and male (OR: 0.254, 95% CI: 0.066-0.979, p = 0.047) were independent predictors of left ventricular recovery at 3 months follow-up. A predictive model was established with the best cutoff value for the prediction of left ventricular recovery 2.33 (sensitivity 0.610, specificity 0.897, and area under the curve 0.765). In patients with a predictive model score less than 2.33, the LVEF increased significantly at 3 months. Conclusions: The post-PCI caIMR can accurately predict left ventricular functional recovery at 3 months follow-up in patients with STEMI treated by primary PCI, supporting its use in clinical practice.

Left ventricular function and coronary microcirculation in patients with mild reduced ejection fraction after STEMI.

BACKGROUND: The characteristics of heart failure (HF) with mildly reduced ejection fraction (EF) (HFmrEF) overlap with those of HF with reduced EF (HFrEF) and HF with preserved EF (HFpEF) and need to be further explored. This study aimed to evaluate left ventricular (LV) function and coronary microcirculation in patients with mildly reduced ejection fraction after acute ST-segment elevation myocardial infarction (STEMI). METHODS: We enrolled 119 patients with STEMI who had undergone speckle tracking imaging and myocardial contrast echocardiography during hospitalization from June 2016 to June 2021. They were classified into normal, HFmrEF, and HFrEF groups according to their left ventricular EF (LVEF): ≥ 50%, 40-50%, and ≤ 40%, respectively. The data of the HFmrEF group were analyzed and compared with those of the normal and HFrEF groups. RESULTS: HFmrEF was observed in 32 patients (26.9%), HFrEF in 17 (14.3%), and normal LVEF in 70 patients (58.8%). The mean global longitudinal strain (GLS) of all patients was - 11.9 ± 3.8%. The GLS of HFmrEF patients was not significantly different from that of the HFrEF group (- 9.9 ± 2.5% and - 8.0 ± 2.3%, respectively, P = 0.052), but they were both lower than that of the normal group (- 13.8% ± 3.5%, P < 0.001). The HFmrEF group exhibited significantly poorer myocardial perfusion index (1.24 ± 0.33) than the normal group (1.08 ± 0.14, P = 0.005) but displayed no significant difference from the HFrEF group (1.18 ± 0.19, P = 0.486). Moreover, a significant difference in the incidence of regional wall motion (WM) abnormalities in the three groups was observed (P = 0.009), and the WM score index of patients with HFmrEF was 1.76 ± 0.30, similar to that of patients with HFrEF (1.81 ± 0.43, P = 0.618), but poorer than that in the normal group (1.33 ± 0.25, P < 0.001). CONCLUSIONS: GLS is a more sensitive tool than LVEF for detecting LV systolic dysfunction. The LV systolic function, coronary microcirculation, and WM in patients with HFmrEF was poorer than that of patients with normal LVEF, but comparable to that in patients with HFrEF. Patients with HFmrEF after STEMI require more attention and appropriate management.

White thrombi on optical coherence tomography after rotational atherectomy of severely calcified coronary lesions.

PURPOSE: Rotational atherectomy (RA) has improved percutaneous treatment of severely calcified coronary lesions, but the "no-reflow" phenomenon remains a serious complication. Platelet activation by RA may contribute to no-reflow, and the use of optical coherence tomography (OCT) to test the effect of RA on white thrombus could confirm platelet activation indirectly. METHODS: We analyzed 53 consecutive patients with severely calcified lesions on coronary angiography. All patients were examined with OCT. In total, 20 patients who received RA and for whom OCT imaging was performed before and after RA and stent implantation comprised the RA group. The remaining 33 patients formed the control group, for whom OCT imaging was performed before balloon dilatation and after stent implantation. RESULTS: The patients in the RA group were older and had a higher incidence of diabetes mellitus. In the control group, there was no thrombogenesis during the procedure, whereas in the RA group, all the target vessels had white thrombi on OCT after RA. The average number of white thrombi per lesion after RA was 7.23 ± 4.4, and the average length of white thrombus was 0.51 ± 0.33 mm. Statistical analysis with Pearson's correlation coefficient showed that thrombus load was related to burr size (r = 0.575, p = 0.040) and number of rotations (r = 0.599, p = 0.031). CONCLUSION: White thrombi during RA can be verified by performing OCT. Treating calcified lesions with RA may enhance thrombogenesis. These data suggest using appropriate therapy to avoid no-reflow during RA.

On the spurious resonance modes of time domain integral equations for analyzing acoustic scattering from penetrable objects.

The interior resonance problem of time domain integral equations (TDIEs) formulated to analyze acoustic field interactions on penetrable objects is investigated. Two types of TDIEs are considered: The first equation, which is termed the time domain potential integral equation (TDPIE), suffers from the interior resonance problem, i.e., its solution is replete with spurious modes that are excited at the resonance frequencies of the acoustic cavity in the shape of the scatterer. Numerical experiments demonstrate that, unlike the frequency-domain integral equations, the amplitude of these modes in the time domain could be suppressed to a level that does not significantly affect the solution. This is achieved by increasing the numerical solution accuracy through the use of a higher-order discretization in space and the band limited approximate prolate spheroidal wave function with high interpolation accuracy as basis function in time. The second equation is obtained by linearly combining TDPIE with its normal derivative. The solution of this equation, which is termed the time domain combined potential integral equation (TDCPIE), does not involve any spurious interior resonance modes but it is not as accurate as the TDPIE solution at non-resonance frequencies. In addition, TDCPIE's discretization calls for treatment of hypersingular integrals.

JLAN: medical code prediction via joint learning attention networks and denoising mechanism.

BACKGROUND: Clinical notes are documents that contain detailed information about the health status of patients. Medical codes generally accompany them. However, the manual diagnosis is costly and error-prone. Moreover, large datasets in clinical diagnosis are susceptible to noise labels because of erroneous manual annotation. Therefore, machine learning has been utilized to perform automatic diagnoses. Previous state-of-the-art (SOTA) models used convolutional neural networks to build document representations for predicting medical codes. However, the clinical notes are usually long-tailed. Moreover, most models fail to deal with the noise during code allocation. Therefore, denoising mechanism and long-tailed classification are the keys to automated coding at scale. RESULTS: In this paper, a new joint learning model is proposed to extend our attention model for predicting medical codes from clinical notes. On the MIMIC-III-50 dataset, our model outperforms all the baselines and SOTA models in all quantitative metrics. On the MIMIC-III-full dataset, our model outperforms in the macro-F1, micro-F1, macro-AUC, and precision at eight compared to the most advanced models. In addition, after introducing the denoising mechanism, the convergence speed of the model becomes faster, and the loss of the model is reduced overall. CONCLUSIONS: The innovations of our model are threefold: firstly, the code-specific representation can be identified by adopted the self-attention mechanism and the label attention mechanism. Secondly, the performance of the long-tailed distributions can be boosted by introducing the joint learning mechanism. Thirdly, the denoising mechanism is suitable for reducing the noise effects in medical code prediction. Finally, we evaluate the effectiveness of our model on the widely-used MIMIC-III datasets and achieve new SOTA results.

Comparison of optical coherence tomography-guided and intravascular ultrasound-guided rotational atherectomy for calcified coronary lesions.

BACKGROUND: To compare the effect and outcomes of optical coherence tomography (OCT)-guided rotational atherectomy (RA) with intravascular ultrasound (IVUS)-guided RA in the treatment of calcified coronary lesions. METHODS: Data of calcified coronary lesions treated with RA that underwent OCT-guided or IVUS-guided from January 2016 to December 2019 at a single-center registry were retrospectively analyzed. The effect and outcomes between underwent OCT-guided RA and IVUS-guided RA were compared. RESULTS: A total of 33 lesions in 32 patients received OCT-guided RA and 51 lesions in 47 patients received IVUS-guided RA. There was no significant difference between OCT-guided RA group and IVUS-guided RA group in clinical baselines characteristics. Comparing the procedural and lesions characteristics of the two groups, the contrast volume was larger [(348.8 ± 110.6) ml vs. (275.2 ± 76.8) ml, P = 0.002] and the scoring balloon was more frequently performed (33.3% vs. 3.9%, P = 0.001) after RA and before stenting in the OCT-guided RA group. Comparing the intravascular imaging findings of the two groups, stent expansion was significantly larger in the OCT-guided RA group ([82 ± 8]% vs. [75 ± 9]%, P = 0.001). Both groups achieved procedural success immediately. There were no significantly differences in the incidence of complications. Although there was no statistical difference in the occurrence of MACE at 1 year between OCT-guided RA group and IVUS-guided RA group (3.1% vs. 6.4%, P = 0.517), no cardiovascular death, TVR and stent thrombosis occurred in OCT-guided RA group. CONCLUSIONS: OCT-guided RA compared to IVUS-guided RA for treating calcified coronary lesions resulted in better stent expansion and may have improved prognosis.

Influence of seasonal temperature change on autotrophic nitrogen removal for mature landfill leachate treatment with high-ammonia by partial nitrification-Anammox process.

In this study, a denitrification (DN)-partial nitritation (PN)-anaerobic ammonia oxidation (Anammox) system for the efficient nitrogen removal of mature landfill leachate was built with a zone-partitioning self-reflux biological reactor as the core device, and the effects of changes in seasonal temperature on the nitrogen removal in non-temperature-control environment were explored. The results showed that as the seasonal temperature decreased from 34°C to 11.3°C, the total nitrogen removal rate of the DN-PN-Anammox system gradually decreased from the peak value of 1.42 kg/(m3•day) to 0.49 kg/(m3•day). At low temperatures (<20°C), when the nitrogen load (NLR) of the system is not appropriate, the fluctuation of high NH4+-N concentration in the landfill leachate greatly influenced the stability of the nitrogen removal. At temperatures of 11°C-15°C, the NLR of the system is controlled below 0.5 kg/(m3•day), which can achieve stable nitrogen removal and the nitrogen removal efficiency can reach above 96%. The abundance of Candidatus Brocadia gradually increased with the decrease of temperature. Nitrosomonas, Candidatus Brocadia and Candidatus Kuenenia as the main functional microorganisms in the low temperature.

Comparison of bailout and planned rotational atherectomy for severe coronary calcified lesions.

BACKGROUND: To compare outcomes of bailout and planned rotational atherectomy (RA) in the treatment of severe calcified coronary lesions. METHODS: Data of patients treated with RA from 2017 to 2018 at a single-center registry were retrospectively analyzed. All patients were divided into planned RA and bailout RA groups, data between two groups were compared. RESULTS: A total of 190 patients were included in this study, 138 patients received planned RA and 52 patients received bailout RA. Baseline clinical characteristics had no significant differences between groups. The number of implanted stents and total stents length were similar. But the number of balloon (1.6 ± 0.8 vs. 2.7 ± 1.3, P < 0.001), procedure time (83.5 ± 26.2 vs. 100.8 ± 36.4 min, P = 0.007), fluoroscopy volume (941 ± 482 vs. 1227 ± 872 mGy, P = 0.012] and contrast amount (237 ± 62 vs. 275 ± 90 ml, P = 0.003) were all lower in planned RA group. Planned RA had a higher procedural success rate (99.3% vs. 92.3%, P = 0.007) and a lower complication incidence (4.3% vs. 17.3%, P = 0.009). But the primary outcomes at 3 years (9.2 and 16.6%, log rank p = 0.24) had no difference between groups. CONCLUSIONS: For severe coronary artery calcification, although planned RA did not improved the long term prognosis compared with bailout RA, but it can improve the immediate procedural success rate, reduce the incidence of complications, the procedure time and the volume of contrast.

SADR: Self-Supervised Graph Learning With Adaptive Denoising for Drug Repositioning.

Traditional drug development is often high-risk and time-consuming. A promising alternative is to reuse or relocate approved drugs. Recently, some methods based on graph representation learning have started to be used for drug repositioning. These models learn the low dimensional embeddings of drug and disease nodes from the drug-disease interaction network to predict the potential association between drugs and diseases. However, these methods have strict requirements for the dataset, and if the dataset is sparse, the performance of these methods will be severely affected. At the same time, these methods have poor robustness to noise in the dataset. In response to the above challenges, we propose a drug repositioning model based on self-supervised graph learning with adptive denoising, called SADR. SADR uses data augmentation and contrastive learning strategies to learn feature representations of nodes, which can effectively solve the problems caused by sparse datasets. SADR includes an adaptive denoising training (ADT) component that can effectively identify noisy data during the training process and remove the impact of noise on the model. We have conducted comprehensive experiments on three datasets and have achieved better prediction accuracy compared to multiple baseline models. At the same time, we propose the top 10 new predictive approved drugs for treating two diseases. This demonstrates the ability of our model to identify potential drug candidates for disease indications.

From Simple to Complex Scenes: Learning Robust Feature Representations for Accurate Human Parsing.

Human parsing has attracted considerable research interest due to its broad potential applications in the computer vision community. In this paper, we explore several useful properties, including high-resolution representation, auxiliary guidance, and model robustness, which collectively contribute to a novel method for accurate human parsing in both simple and complex scenes. Starting from simple scenes: we propose the boundary-aware hybrid resolution network (BHRN), an advanced human parsing network. BHRN utilizes deconvolutional layers and multi-scale supervision to generate rich high-resolution representations. Additionally, it includes an edge perceiving branch designed to enhance the fineness of part boundaries. Building on BHRN, we construct a dual-task mutual learning (DTML) framework. It not only provides implicit guidance to assist the parser by incorporating boundary features, but also explicitly maintains the high-order consistency between the parsing prediction and the ground truth. Toward complex scenes: we develop a domain transform method to enhance the model robustness. By transforming the input space from the spatial domain to the polar harmonic Fourier moment domain, the mapping relationship to the output semantic space is highly stable. This transformation yields robust representations for both clean and corrupted data. When evaluated on standard benchmark datasets, our method achieves superior performance compared to state-of-the-art human parsing methods. Furthermore, our domain transform strategy significantly improves the robustness of DTML dramatically in most complex scenes.

TransFOL: A Logical Query Model for Complex Relational Reasoning in Drug-Drug Interaction.

Predicting drug-drug interaction (DDI) plays a crucial role in drug recommendation and discovery. However, wet lab methods are prohibitively expensive and time-consuming due to drug interactions. In recent years, deep learning methods have gained widespread use in drug reasoning. Although these methods have demonstrated effectiveness, they can only predict the interaction between a drug pair and do not contain any other information. However, DDI is greatly affected by various other biomedical factors (such as the dose of the drug). As a result, it is challenging to apply them to more complex and meaningful reasoning tasks. Therefore, this study regards DDI as a link prediction problem on knowledge graphs and proposes a DDI prediction model based on Cross-Transformer and Graph Convolutional Networks (GCN) in first-order logical query form, TransFOL. In the model, a biomedical query graph is first built to learn the embedding representation. Subsequently, an enhancement module is designed to aggregate the semantics of entities and relations. Cross-Transformer is used for encoding to obtain semantic information between nodes, and GCN is used to gather neighbour information further and predict inference results. To evaluate the performance of TransFOL on common DDI tasks, we conduct experiments on two benchmark datasets. The experimental results indicate that our model outperforms state-of-the-art methods on traditional DDI tasks. Additionally, we introduce different biomedical information in the other two experiments to make the settings more realistic. Experimental results verify the strong drug reasoning ability and generalization of TransFOL in complex settings. Data and code are available at https://github.com/Cheng0829/TransFOL.

Knowledge enhanced attention aggregation network for medicine recommendation.

The combination of deep learning and the medical field has recently achieved great success, particularly in recommending medicine for patients. However, patients' clinical records often contain repeated medical information that can significantly impact their health condition. Most existing methods for modeling longitudinal patient information overlook the impact of individual diagnoses and procedures on the patient's health, resulting in insufficient patient representation and limited accuracy of medicine recommendations. Therefore, we propose a medicine recommendation model called KEAN, which is based on an attention aggregation network and enhanced graph convolution. Specifically, KEAN can aggregate individual diagnoses and procedures in patient visits to capture significant features that affect patients' diseases. We further incorporate medicine knowledge from complex medicine combinations, reduce drug-drug interactions (DDIs), and recommend medicines that are beneficial to patients' health. The experimental results on the MIMIC-III dataset demonstrate that our model outperforms existing advanced methods, which highlights the effectiveness of the proposed method.

High-Performance All-Textile Triboelectric Nanogenerator toward Intelligent Sports Sensing and Biomechanical Energy Harvesting.

Merging textiles with advanced energy harvesting technology via triboelectric effects brings novel insights into self-powered wearable textile electronics. However, fabrication of a comfortable textile-based triboelectric nanogenerator (TENG) with high outputs remains challenging. Herein, we propose a highly flexible, tailorable, single-electrode all-textile TENG (t-TENG) with both wear comfort and high outputs. A dielectric modulated porous composite coating containing poly(vinylidene fluoride)-hexafluoropropylene copolymer and barium titanate nanoparticles is constructed on conductive fabric to counterpart with highly positive glass fiber fabric through knotted yarn bonding, maintaining the superiority of textiles and strong triboelectricity. Through the synergistic optimization of charge storage via dielectric modulation and charge dissipation offset by electrical poling, remarkable outputs (261 V, 1.5 µA, and 12.7 nC) are obtained from a miniaturized, lightweight t-TENG (2 × 2 cm2, 130 mg) with an instantaneous power density of 654.48 mW·m-2, as well as excellent electrical robustness and device durability over 20,000 cycles. The t-TENG also exhibits a high sensitivity of 3.438 V·kPa-1 in the force region (1-10 N), demonstrating great potential in TENG-based intelligent sports sensing applications for monitoring and correcting the basketball shooting hand and foot arch posture. Furthermore, over 110 light-emitting diode arrays can be lightened up by gently tapping this miniaturized t-TENG. It also offers a wearable power source scheme through integrating the single-electrode device into clothing and utilizing the skin as the grounded electrode, revealing its ease of integration and biomechanical energy harvesting capability. This work provides an attractive paradigm for next-generation textile electronics with well-balanced device performance and wear comfort.

[Safety and efficacy of rotational atherectomy followed by drug-eluting stenting for treating patients with heavily calcified coronary lesions].

OBJECTIVE: To investigate the safety and efficacy of rotational atherectomy followed by drug-eluting stent implantation for treating patients with heavily calcified coronary lesions. METHODS: From March 1, 2010 to September 1, 2012, 65 cases with 78 heavily calcified coronary lesions which were treated with rotational atherectomy followed by drug-eluting stent implantation in Peking University People's Hospital were included, and 36 cases also underwent intravascular ultrasound to guide the rotational atherectomy procedure and drug-eluting stent implantation.All patients were followed up in hospital and post discharge. Procedure parameters, complications and major adverse cardiovascular events (cardiac death, non-fatal myocardial infarction, percutaneous coronary intervention related myocardial infarction, target vessel revascularization, recurrent angina, intra-stent restenosis and stent thrombosis) were analyzed. RESULTS: Direct rotational atherectomy was performed in 64.6%(42/65) patients, rescued rotational atherectomy in 35.4%(23/65) patients, drug-eluting stents implantation was applied to all cases after rotational atherectomy. The immediate procedural success rate was 100% (78/78). The average burr/artery ratio was 0.50 ± 0.04, the average number of burr used per case was 1.15 ± 0.36. The average burr/artery ratio was 0.52 ± 0.03 and the average number of burr used per cases was 1.19 ± 0.40 in 36 cases guided with intravascular ultrasound. Five cases (7.7%) developed complications and were treated accordingly during procedure with satisfactory results. The incidence of major adverse cardiovascular events was 13.8% (9/65) during (17.6 ± 8.5) months follow-up. CONCLUSION: Rotational atherectomy followed by drug-eluting stent implantation is a safe and efficient technique for treating heavily calcified coronary lesions.

Semantic-enhanced Graph Contrastive Learning with Adaptive Denoising for Drug Repositioning.

The traditional drug development process requires a significant investment in workforce and financial resources. Drug repositioning as an efficient alternative has attracted much attention during the last few years. Despite the wide application and success of the method, there are still many shortcomings in the existing model. For example, sparse datasets will seriously affect the existing methods' performance. Additionally, these methods do not pay attention to the noise in datasets. In response to the above defects, we propose a semantic-enriched augmented graph contrastive learning with an adaptive denoising method, called SGCD. This method enhances data from the perspective of the embedding layer, deeply mines potential neighborhood relation-ships in semantic space, and combines similar drugs in the semantic neighborhoods into prototype comparison targets, thus effectively mitigating the impact of data sparsity on the model. Moreover, to enhance the model's robustness to noisy data, we use the adaptive denoising method, which can effectively identify noisy data in the training process. Exhaustive experiments on multiple real datasets show the effectiveness of the proposed model. The code implementation is available at https://github.com/yuhuimin11/SGCD-master.

Automatic International Classification of Diseases Coding via Note-Code Interaction Network with Denoising Mechanism.

Clinical notes are comprehensive files containing explicit information about a patient's visit. However, accurately assigning medical codes from clinical documents can be a persistent challenge due to the complexity of clinical data and the vast range of medical codes. Moreover, the large volume of medical records, the noisy medical records, and the uneven quality of coders all negatively impact the quality of the final codes. Deep learning technology has recently been integrated into automatic International Classification of Diseases (ICD) coding tasks to improve accuracy. Nevertheless, the imbalanced class problem, the complexness of code associations, and the noise in lengthy records still restrict the advancement of ICD coding tasks in deep learning. Thus, we present the Note-code Interaction Denoising Network (NIDN) that employs the self-attention mechanism to pull critical semantic features in electronic medical records (EMRs). Our model utilizes the label attention mechanism for retaining code-specific text expression. We introduce Clinical Classifications Software coding for multitask learning, capturing the functional relationships of medical coding to oblige in model prediction. To minimize the impact of noise on model prediction and improve the label distribution imbalance, a denoising module is introduced to filter noise. Our practical consequences indicate that the model NIDN exceeds competitive models on a third version of Medical Information Mart for Intensive Care data set.