Robotic exoskeletons and total knee arthroplasty: The future of knee rehabilitation and replacement - A meta-analysis.

OBJECTIVE: Exoskeletons can play a crucial role in post-TKA rehabilitation by accelerating recovery, improving mobility, and reducing further injury risk. This meta-analysis evaluated the effectiveness of exoskeletons in post-total knee replacement (TKR) rehabilitation. DESIGN: Comprehensive searches were conducted on PubMed, OVID Medline, Cochrane Collaboration Library, and Embase (period: database inception to March 2023). Randomized controlled trials enrolling patients who underwent TKR and studies examining the effect of robot-assisted rehabilitation on physical function and pain outcomes were eligible for inclusion. Eight studies (302 patients) were thus included. RESULTS: Exoskeletons significantly improved active range of motion (ROM) (SMD: 10.98, 95% confidence interval (CI): 7.81-14.16, P < .001), passive ROM (SMD: 4.11, 95% CI: 1.02-7.20, P = .009), Hospital for Special Surgery scores (SMD: 7.78, 95% CI: 5.87-9.68, P < .00001), and hospital stay length (SMD: -3.19, 95% CI: -4 to -2.38, P < .00001) compared with conventional rehabilitation. Active and passive ROM improvements suggest that exoskeletons aid knee function restoration and mobility post-TKR, whereas Hospital for Special Surgery score improvements support exoskeleton use in TKR rehabilitation. A shorter hospital stay was an important finding which could potentially reduce healthcare costs and improve outcomes. CONCLUSION: Despite the inclusion of a limited number of studies, our findings suggest that exoskeletons can enhance post-TKR rehabilitation outcomes and improve quality of life. Robot-assisted rehabilitation may be effective following TKR. Further research should confirm these findings.

Utilizing Nearest-Neighbor Clustering for Addressing Imbalanced Datasets in Bioengineering.

Imbalance classification is common in scenarios like fault diagnosis, intrusion detection, and medical diagnosis, where obtaining abnormal data is difficult. This article addresses a one-class problem, implementing and refining the One-Class Nearest-Neighbor (OCNN) algorithm. The original inter-quartile range mechanism is replaced with the K-means with outlier removal (KMOR) algorithm for efficient outlier identification in the target class. Parameters are optimized by treating these outliers as non-target-class samples. A new algorithm, the Location-based Nearest-Neighbor (LBNN) algorithm, clusters one-class training data using KMOR and calculates the farthest distance and percentile for each test data point to determine if it belongs to the target class. Experiments cover parameter studies, validation on eight standard imbalanced datasets from KEEL, and three applications on real medical imbalanced datasets. Results show superior performance in precision, recall, and G-means compared to traditional classification models, making it effective for handling imbalanced data challenges.

Membrane lipid remodeling eradicates Helicobacter pylori by manipulating the cholesteryl 6'-acylglucoside biosynthesis.

BACKGROUND: Helicobacter pylori, the main cause of various gastric diseases, infects approximately half of the human population. This pathogen is auxotrophic for cholesterol which it converts to various cholesteryl α-glucoside derivatives, including cholesteryl 6'-acyl α-glucoside (CAG). Since the related biosynthetic enzymes can be translocated to the host cells, the acyl chain of CAG likely comes from its precursor phosphatidylethanolamine (PE) in the host membranes. This work aims at examining how the acyl chain of CAG and PE inhibits the membrane functions, especially bacterial adhesion. METHODS: Eleven CAGs that differ in acyl chains were used to study the membrane properties of human gastric adenocarcinoma cells (AGS cells), including lipid rafts clustering (monitored by immunofluorescence with confocal microscopy) and lateral membrane fluidity (by the fluorescence recovery after photobleaching). Cell-based and mouse models were employed to study the degree of bacterial adhesion, the analyses of which were conducted by using flow cytometry and immunofluorescence staining, respectively. The lipidomes of H. pylori, AGS cells and H. pylori-AGS co-cultures were analyzed by Ultraperformance Liquid Chromatography-Tandem Mass Spectroscopy (UPLC-MS/MS) to examine the effect of PE(10:0)2, PE(18:0)2, PE(18:3)2, or PE(22:6)2 treatments. RESULTS: CAG10:0, CAG18:3 and CAG22:6 were found to cause the most adverse effect on the bacterial adhesion. Further LC-MS analysis indicated that the treatment of PE(10:0)2 resulted in dual effects to inhibit the bacterial adhesion, including the generation of CAG10:0 and significant changes in the membrane compositions. The initial (1 h) lipidome changes involved in the incorporation of 10:0 acyl chains into dihydro- and phytosphingosine derivatives and ceramides. In contrast, after 16 h, glycerophospholipids displayed obvious increase in their very long chain fatty acids, monounsaturated and polyunsaturated fatty acids that are considered to enhance membrane fluidity. CONCLUSIONS: The PE(10:0)2 treatment significantly reduced bacterial adhesion in both AGS cells and mouse models. Our approach of membrane remodeling has thus shown great promise as a new anti-H. pylori therapy.

Ultrasensitive Upconversion Nanoparticle Immunoassay for Human Serum Cardiac Troponin I Detection Achieved with Resonant Waveguide Grating.

Selective detection of biomarkers at low concentrations in blood is crucial for the clinical diagnosis of many diseases but remains challenging. In this work, we aimed to develop an ultrasensitive immunoassay that can detect biomarkers in serum with an attomolar limit of detection (LOD). We proposed a sandwich-type heterogeneous immunosensor in a 3 × 3 well array format by integrating a resonant waveguide grating (RWG) substrate with upconversion nanoparticles (UCNPs). UCNPs were used to label a target biomarker captured by capture antibody molecules immobilized on the surface of the RWG substrate, and the RWG substrate was used to enhance the upconversion luminescence (UCL) of UCNPs through excitation resonance. The LOD of the immunosensor was greatly reduced due to the increased UCL of UCNPs and the reduction of nonspecific adsorption of detection antibody-conjugated UCNPs on the RWG substrate surface by coating the RWG substrate surface with a carboxymethyl dextran layer. The immunosensor exhibited an extremely low LOD [0.24 fg/mL (9.1 aM)] and wide detection range (1 fg/mL to 100 pg/mL) in the detection of cardiac troponin I (cTnI). The cTnI concentrations in human serum samples collected at different times during cyclophosphamide, epirubicin, and 5-fluorouracil (CEF) chemotherapy in a breast cancer patient were measured by an immunosensor, and the results showed that the CEF chemotherapy did cause cardiotoxicity in the patient. Having a higher number of wells in such an array-based biosensor, the sensor can be developed as a high-throughput diagnostic tool for clinically important biomarkers.

Characterization and catalytic investigation of fungal single-module nonribosomal peptide synthetase in terpene-amino acid meroterpenoid biosynthesis.

Hybrid natural products are compounds that originate from diverse biosynthetic pathways and undergo a conjugation process, which enables them to expand their chemical diversity and biological functionality. Terpene-amino acid meroterpenoids have garnered increasing attention in recent years, driven by the discovery of noteworthy examples such as the anthelmintic CJ-12662, the insecticidal paeciloxazine, and aculene A (1). In the biosynthesis of terpene-amino acid natural products, single-module nonribosomal peptide synthetases (NRPSs) have been identified to be involved in the esterification step, catalyzing the fusion of modified terpene and amino acid components. Despite prior investigations into these NRPSs through gene deletion or in vivo experiments, the enzymatic basis and mechanistic insights underlying this family of single-module NRPSs remain unclear. In this study, we performed biochemical characterization of AneB by in vitro characterization, molecular docking, and site-directed mutagenesis. The enzyme reaction analyses, performed with L-proline and daucane/nordaucane sesquiterpene substrates, revealed that AneB specifically esterifies the C10-OH of aculenes with L-proline. Notably, in contrast to ThmA in CJ-12662 biosynthesis, which exclusively recognizes oxygenated amorpha-4,11-diene sesquiterpenes for L-tryptophan transfer, AneB demonstrates broad substrate selectivity, including oxygenated amorpha-4,11-diene and 2-phenylethanol, resulting in the production of diverse unnatural prolyl compounds. Furthermore, site-directed mutagenesis experiments indicated the involvement of H794 and D798 in the esterification catalyzed by AneB. Lastly, domain swapping between AneB and ThmA unveiled that the A‒T domains of ThmA can be effectively harnessed by the C domain of AneB for L-tryptophan transfer, thus highlighting the potential of the C domain of AneB for generating various terpene-amino acid meroterpenoid derivatives. ONE-SENTENCE SUMMARY: The enzymatic basis and mechanistic insights into AneB, a single-module NRPS, highlight its capacity to generate various terpene-amino acid meroterpenoid derivatives.

Molecular subtypes of breast cancer predicting clinical benefits of radiotherapy after breast-conserving surgery: a propensity-score-matched cohort study.

BACKGROUND: Based on the molecular expression of cancer cells, molecular subtypes of breast cancer have been applied to classify patients for predicting clinical outcomes and prognosis. However, further evidence is needed regarding the influence of molecular subtypes on the efficacy of radiotherapy (RT) after breast-conserving surgery (BCS), particularly in a population-based context. Hence, the present study employed a propensity-score-matched cohort design to investigate the potential role of molecular subtypes in stratifying patient outcomes for post-BCS RT and to identify the specific clinical benefits that may emerge. METHODS: From 2006 to 2019, the present study included 59,502 breast cancer patients who underwent BCS from the Taiwan National Health Insurance Research Database. Propensity scores were utilized to match confounding variables between patients with and without RT within each subtype of breast cancer, namely luminal A, luminal B/HER2-negative, luminal B/HER2-positive, basal-like, and HER2-enriched ones. Several clinical outcomes were assessed, in terms of local recurrence (LR), regional recurrence (RR), distant metastasis (DM), disease-free survival (DFS), and overall survival (OS). RESULTS: After post-BCS RT, patients with luminal A and luminal B/HER2-positive breast cancers exhibited a decrease in LR (adjusted hazard ratio [aHR] = 0.18, p < 0.0001; and, 0.24, p = 0.0049, respectively). Furthermore, reduced RR and improved DFS were observed in patients with luminal A (aHR = 0.15, p = 0.0004; and 0.29, p < 0.0001), luminal B/HER2-negative (aHR = 0.06, p = 0.0093; and, 0.46, p = 0.028), and luminal B/HER2-positive (aHR = 0.14, p = 0.01; and, 0.38, p < 0.0001) breast cancers. Notably, OS benefits were found in patients with luminal A (aHR = 0.62, p = 0.002), luminal B/HER2-negative (aHR = 0.30, p < 0.0001), basal-like (aHR = 0.40, p < 0.0001), and HER2-enriched (aHR = 0.50, p = 0.03), but not luminal B/HER2-positive diseases. Remarkably, when considering DM, luminal A patients who received RT demonstrated a lower cumulative incidence of DM than those without RT (p = 0.02). CONCLUSION: In patients with luminal A breast cancer who undergo BCS, RT could decrease the likelihood of tumor metastasis. After RT, the tumor's hormone receptor status may predict tumor control regarding LR, RR, and DFS. Besides, the HER2 status of luminal breast cancer patients may serve as an additional predictor of OS after post-BCS RT. However, further prospective studies are required to validate these findings.

Sulfoglycolipids and Related Analogues of Mycobacterium tuberculosis: Chemical Synthesis and Immunological Studies.

Mycobacterium tuberculosis (Mtb) causes tuberculosis as one major threat to human health, which has been deteriorated owing to the emerging multidrug resistance. Mtb contains a complex lipophilic cell wall structure that is important for bacterial persistence. Among the lipid components, sulfoglycolipids (SGLs), known to induce immune cell responses, are composed of a trehalose core attached with a conserved sulfate group and 1-4 fatty acyl chains in an asymmetric pattern. At least one of these acyl chains is polymethylated with 3-12 methyl branches. Although Mtb SGL can be isolated from bacterial culture, resulting SGL is still a homologous mixture, impeding accurate research studies. This up-to-date review covers the chemical synthesis and immunological studies of Mtb SGLs and structural analogues, with an emphasis on the development of new glycosylation methods and the asymmetric synthesis of polymethylated scaffolds. Both are critical to advance further research on biological functions of these complicated SGLs.

Development and validation of a deep learning pipeline to measure pericardial effusion in echocardiography.

Objectives: The aim of this study was to develop a deep-learning pipeline for the measurement of pericardial effusion (PE) based on raw echocardiography clips, as current methods for PE measurement can be operator-dependent and present challenges in certain situations. Methods: The proposed pipeline consisted of three distinct steps: moving window view selection (MWVS), automated segmentation, and width calculation from a segmented mask. The MWVS model utilized the ResNet architecture to classify each frame of the extracted raw echocardiography files into selected view types. The automated segmentation step then generated a mask for the PE area from the extracted echocardiography clip, and a computer vision technique was used to calculate the largest width of the PE from the segmented mask. The pipeline was applied to a total of 995 echocardiographic examinations. Results: The proposed deep-learning pipeline exhibited high performance, as evidenced by intraclass correlation coefficient (ICC) values of 0.867 for internal validation and 0.801 for external validation. The pipeline demonstrated a high level of accuracy in detecting PE, with an area under the receiving operating characteristic curve (AUC) of 0.926 (95% CI: 0.902-0.951) for internal validation and 0.842 (95% CI: 0.794-0.889) for external validation. Conclusion: The machine-learning pipeline developed in this study can automatically calculate the width of PE from raw ultrasound clips. The novel concepts of moving window view selection for image quality control and computer vision techniques for maximal PE width calculation seem useful in the field of ultrasound. This pipeline could potentially provide a standardized and objective approach to the measurement of PE, reducing operator-dependency and improving accuracy.

Synthesis of Metal/SU-8 Nanocomposites through Photoreduction on SU-8 Substrates.

The paper presents a simple, fast, and cost-effective method for creating metal/SU-8 nanocomposites by applying a metal precursor drop onto the surface or nanostructure of SU-8 and exposing it to UV light. No pre-mixing of the metal precursor with the SU-8 polymer or pre-synthesis of metal nanoparticles is required. A TEM analysis was conducted to confirm the composition and depth distribution of the silver nanoparticles, which penetrate the SU-8 film and uniformly form the Ag/SU-8 nanocomposites. The antibacterial properties of the nanocomposites were evaluated. Moreover, a composite surface with a top layer of gold nanodisks and a bottom layer of Ag/SU-8 nanocomposites was produced using the same photoreduction process with gold and silver precursors, respectively. The reduction parameters can be manipulated to customize the color and spectrum of various composite surfaces.

Joint Data Transmission and Energy Harvesting for MISO Downlink Transmission Coordination in Wireless IoT Networks.

The advent of simultaneous wireless information and power (SWIPT) has been regarded as a promising technique to provide power supplies for an energy sustainable Internet of Things (IoT), which is of paramount importance due to the proliferation of high data communication demands of low-power network devices. In such networks, a multi-antenna base station (BS) in each cell can be utilized to concurrently transmit messages and energies to its intended IoT user equipment (IoT-UE) with a single antenna under a common broadcast frequency band, resulting in a multi-cell multi-input single-output (MISO) interference channel (IC). In this work, we aim to find the trade-off between the spectrum efficiency (SE) and energy harvesting (EH) in SWIPT-enabled networks with MISO ICs. For this, we derive a multi-objective optimization (MOO) formulation to obtain the optimal beamforming pattern (BP) and power splitting ratio (PR), and we propose a fractional programming (FP) model to find the solution. To tackle the nonconvexity of FP, an evolutionary algorithm (EA)-aided quadratic transform technique is proposed, which recasts the nonconvex problem as a sequence of convex problems to be solved iteratively. To further reduce the communication overhead and computational complexity, a distributed multi-agent learning-based approach is proposed that requires only partial observations of the channel state information (CSI). In this approach, each BS is equipped with a double deep Q network (DDQN) to determine the BP and PR for its UE with lower computational complexity based on the observations through a limited information exchange process. Finally, with the simulation experiments, we verify the trade-off between SE and EH, and we demonstrate that, apart from the FP algorithm introduced to provide superior solutions, the proposed DDQN algorithm also shows its performance gain in terms of utility to be up to 1.23-, 1.87-, and 3.45-times larger than the Advantage Actor Critic (A2C), greedy, and random algorithms, respectively, in comparison in the simulated environment.

Use of Machine Learning to Differentiate Children With Kawasaki Disease From Other Febrile Children in a Pediatric Emergency Department.

Importance: Early awareness of Kawasaki disease (KD) helps physicians administer appropriate therapy to prevent acquired heart disease in children. However, diagnosing KD is challenging and relies largely on subjective diagnosis criteria. Objective: To develop a prediction model using machine learning with objective parameters to differentiate children with KD from other febrile children. Design, Setting, and Participants: This diagnostic study included 74 641 febrile children younger than 5 years who were recruited from 4 hospitals, including 2 medical centers and 2 regional hospitals, between January 1, 2010, and December 31, 2019. Statistical analysis was performed from October 2021 to February 2023. Main Outcomes and Measures: Demographic data and laboratory values from electronic medical records, including complete blood cell count with differential, urinalysis, and biochemistry, were collected as possible parameters. The primary outcome was whether the febrile children fulfilled the diagnostic criteria of KD. The supervised eXtreme Gradient Boosting (XGBoost) machine learning method was applied to establish a prediction model. The confusion matrix and likelihood ratio were used to evaluate the performance of the prediction model. Results: This study included a total of 1142 patients with KD (mean [SD] age, 1.1 [0.8] years; 687 male patients [60.2%]) and 73 499 febrile children (mean [SD] age, 1.6 [1.4] years; 41 465 male patients [56.4%]) comprising the control group. The KD group was predominantly male (odds ratio, 1.79; 95% CI, 1.55-2.06) with younger age (mean difference, -0.6 years [95% CI, -0.6 to -0.5 years]) compared with the control group. The prediction model's best performance in the testing set was able to achieve 92.5% sensitivity, 97.3% specificity, 34.5% positive predictive value, 99.9% negative predictive value, and a positive likelihood ratio of 34.0, which indicates outstanding performance. The area under the receiver operating characteristic curve of the prediction model was 0.980 (95% CI, 0.974-0.987). Conclusions and Relevance: This diagnostic study suggests that results of objective laboratory tests had the potential to be predictors of KD. Furthermore, these findings suggested that machine learning with XGBoost can help physicians differentiate children with KD from other febrile children in pediatric emergency departments with excellent sensitivity, specificity, and accuracy.

Using Machine Learning for the Risk Factors Classification of Glycemic Control in Type 2 Diabetes Mellitus.

Several risk factors are related to glycemic control in patients with type 2 diabetes mellitus (T2DM), including demographics, medical conditions, negative emotions, lipid profiles, and heart rate variability (HRV; to present cardiac autonomic activity). The interactions between these risk factors remain unclear. This study aimed to use machine learning methods of artificial intelligence to explore the relationships between various risk factors and glycemic control in T2DM patients. The study utilized a database from Lin et al. (2022) that included 647 T2DM patients. Regression tree analysis was conducted to identify the interactions among risk factors that contribute to glycated hemoglobin (HbA1c) values, and various machine learning methods were compared for their accuracy in classifying T2DM patients. The results of the regression tree analysis revealed that high depression scores may be a risk factor in one subgroup but not in others. When comparing different machine learning classification methods, the random forest algorithm emerged as the best-performing method with a small set of features. Specifically, the random forest algorithm achieved 84% accuracy, 95% area under the curve (AUC), 77% sensitivity, and 91% specificity. Using machine learning methods can provide significant value in accurately classifying patients with T2DM when considering depression as a risk factor.

Use of a Deep-Learning Algorithm to Guide Novices in Performing Focused Assessment With Sonography in Trauma.

This quality improvement study compares the diagnostic quality and completion time between ultrasonography operators guided by artificial intelligence vs those without such assistance.

Explainable deep learning model to predict invasive bacterial infection in febrile young infants: A retrospective study.

BACKGROUND: Machine learning models have demonstrated superior performance in predicting invasive bacterial infection (IBI) in febrile infants compared to commonly used risk stratification criteria in recent studies. However, the black-box nature of these models can make them difficult to apply in clinical practice. In this study, we developed and validated an explainable deep learning model that can predict IBI in febrile infants ≤ 60 days of age visiting the emergency department. METHODS: We conducted a retrospective study of febrile infants aged ≤ 60 days who presented to the pediatric emergency department of a medical center in Taiwan between January 1, 2011 and December 31, 2019. Patients with uncertain test results and complex chronic health conditions were excluded. IBI was defined as the growth of a pathogen in the blood or cerebrospinal fluid. We used a deep neural network to develop a predictive model for IBI and compared its performance to the IBI score and step-by-step approach. The SHapley Additive Explanations (SHAP) technique was used to explain the model's predictions at different levels. RESULTS: Our study included 1847 patients, 53 (2.7%) of whom had IBI. The deep learning model performed similarly to the IBI score and step-by-step approach in terms of sensitivity and negative predictive value, but provided better specificity (54%), positive predictive value (5%), and area under the receiver-operating characteristic curve (0.87). SHapley Additive exPlanations identified five influential predictive variables (absolute neutrophil count, body temperature, heart rate, age, and C-reactive protein). CONCLUSION: We have developed an explainable deep learning model that can predict IBI in febrile infants aged 0-60 days. The model not only performs better than previous scoring systems, but also provides insight into how it arrives at its predictions through individual features and cases.

Adsorption and purification of biogas inside graphitic nanopores: molecular dynamics simulation approach.

Biogas is one of the most common sources of biomass energy. Due to the associated environmental pollution and costs, desulfurization, and purification are the most important challenges of biogas power generation. Using all-atom molecular dynamics (MD), we systematically simulated the isothermal adsorption behavior of biogas (comprising CH4, CO2, H2O, H2S, and H2) in graphite (Gr) slit nanopores. The impact of slit width, system temperature, and moisture content on the adsorption energy, adsorption ratio, and diffusion coefficient of biogas molecules was investigated. Simulation results revealed that due to strong interactions between graphite and H2S, graphite slits of width d = 48 ~ 80 Å displayed significant selective adsorption of H2S molecules. At temperatures between 300 and 500 K, Gr slits can effectively separate H2S in biogas. Moreover, as the moisture content of biogas (vol%) increases from 0 to 20%, the formation and interactions of hydrogen bonds between water molecules create H2O films accumulating on the Gr surface and taking up the adsorption sites, which reduces the amount of hydrogen sulfide that can be adsorbed. Our findings provide important insights into the material design for biogas purification. A schematic representation of molecular interactions between adsorbates and the wall for biogas mixtures (comprising CH4, CO2, H2O, H2S, and H2) inside graphitic nanopores.

Cytosolic galectin-4 enchains bacteria, restricts their motility, and promotes inflammasome activation in intestinal epithelial cells.

Galectin-4, a member of the galectin family of animal glycan-binding proteins (GBPs), is specifically expressed in gastrointestinal epithelial cells and is known to be able to bind microbes. However, its function in host-gut microbe interactions remains unknown. Here, we show that intracellular galectin-4 in intestinal epithelial cells (IECs) coats cytosolic Salmonella enterica serovar Worthington and induces the formation of bacterial chains and aggregates. Galectin-4 enchains bacteria during their growth by binding to the O-antigen of lipopolysaccharides. Furthermore, the binding of galectin-4 to bacterial surfaces restricts intracellular bacterial motility. Galectin-4 enhances caspase-1 activation and mature IL-18 production in infected IECs especially when autophagy is inhibited. Finally, orally administered S. enterica serovar Worthington, which is recognized by human galectin-4 but not mouse galectin-4, translocated from the intestines to mesenteric lymph nodes less effectively in human galectin-4-transgenic mice than in littermate controls. Our results suggest that galectin-4 plays an important role in host-gut microbe interactions and prevents the dissemination of pathogens. The results of the study revealed a novel mechanism of host-microbe interactions that involves the direct binding of cytosolic lectins to glycans on intracellular microbes.

Tunable Strategy for the Asymmetric Synthesis of Sulfoglycolipids from Mycobacterium tuberculosis To Elucidate the Structure and Immunomodulatory Property Relationships.

We developed a versatile asymmetric strategy to synthesize different classes of sulfoglycolipids (SGLs) from Mycobacterium tuberculosis. The strategy features the use of asymmetrically protected trehaloses, which were acquired from the glycosylation of TMS α-glucosyl acceptors with benzylidene-protected thioglucosyl donors. The positions of the protecting groups at the donors and acceptors can be fine-tuned to obtain different protecting-group patterns, which is crucial for regioselective acylation and sulfation. In addition, a chemoenzymatic strategy was established to prepare the polymethylated fatty acid building blocks. The strategy employs inexpensive lipase as a desymmetrization agent in the preparation of the starting substrate and readily available chiral oxazolidinone as a chirality-controlling agent in the construction of the polymethylated fatty acids. A subsequent investigation on the immunomodulatory properties of each class of SGLs showed how the structures of SGLs impact the host innate immunity response.

Using Deep Transfer Learning to Detect Hyperkalemia From Ambulatory Electrocardiogram Monitors in Intensive Care Units: Personalized Medicine Approach.

BACKGROUND: Hyperkalemia is a critical condition, especially in intensive care units. So far, there have been no accurate and noninvasive methods for recognizing hyperkalemia events on ambulatory electrocardiogram monitors. OBJECTIVE: This study aimed to improve the accuracy of hyperkalemia predictions from ambulatory electrocardiogram (ECG) monitors using a personalized transfer learning method; this would be done by training a generic model and refining it with personal data. METHODS: This retrospective cohort study used open source data from the Waveform Database Matched Subset of the Medical Information Mart From Intensive Care III (MIMIC-III). We included patients with multiple serum potassium test results and matched ECG data from the MIMIC-III database. A 1D convolutional neural network-based deep learning model was first developed to predict hyperkalemia in a generic population. Once the model achieved a state-of-the-art performance, it was used in an active transfer learning process to perform patient-adaptive heartbeat classification tasks. RESULTS: The results show that by acquiring data from each new patient, the personalized model can improve the accuracy of hyperkalemia detection significantly, from an average of 0.604 (SD 0.211) to 0.980 (SD 0.078), when compared with the generic model. Moreover, the area under the receiver operating characteristic curve level improved from 0.729 (SD 0.240) to 0.945 (SD 0.094). CONCLUSIONS: By using the deep transfer learning method, we were able to build a clinical standard model for hyperkalemia detection using ambulatory ECG monitors. These findings could potentially be extended to applications that continuously monitor one's ECGs for early alerts of hyperkalemia and help avoid unnecessary blood tests.

Machine learning models for predicting in-hospital mortality in patient with sepsis: Analysis of vital sign dynamics.

Purpose: To build machine learning models for predicting the risk of in-hospital death in patients with sepsis within 48 h, using only dynamic changes in the patient's vital signs. Methods: This retrospective observational cohort study enrolled septic patients from five emergency departments (ED) in Taiwan. We adopted seven variables, i.e., age, sex, systolic blood pressure, diastolic blood pressure, heart rate, respiratory rate, and body temperature. Results: Among all 353,253 visits, after excluding 159,607 visits (45%), the study group consisted of 193,646 ED visits. With a leading time of 6 h, the convolutional neural networks (CNNs), long short-term memory (LSTM), and random forest (RF) had accuracy rates of 0.905, 0.817, and 0.835, respectively, and the area under the receiver operating characteristic curve (AUC) was 0.840, 0.761, and 0.770, respectively. With a leading time of 48 h, the CNN, LSTM, and RF achieved accuracy rates of 0.828, 0759, and 0.805, respectively, and an AUC of 0.811, 0.734, and 0.776, respectively. Conclusion: By analyzing dynamic vital sign data, machine learning models can predict mortality in septic patients within 6 to 48 h of admission. The performance of the testing models is more accurate if the lead time is closer to the event.

Thermal Conductivity of Helium and Argon at High Pressure and High Temperature.

Helium (He) and argon (Ar) are important rare gases and pressure media used in diamond-anvil cell (DAC) experiments. Their thermal conductivity at high pressure-temperature (P-T) conditions is a crucial parameter for modeling heat conduction and temperature distribution within a DAC. Here we report the thermal conductivity of He and Ar over a wide range of high P-T conditions using ultrafast time-domain thermoreflectance coupled with an externally heated DAC. We find that at room temperature the thermal conductivity of liquid and solid He shows a pressure dependence of P0.86 and P0.72, respectively; upon heating the liquid, He at 10.2 GPa follows a T0.45 dependence. By contrast, the thermal conductivity of solid Ar at room temperature has a pressure dependence of P1.25, while a T-1.37 dependence is observed for solid Ar at 19 GPa. Our results not only provide crucial bases for further investigation into the physical mechanisms of heat transport in He and Ar under extremes, but also substantially improve the accuracy of modeling the temperature profile within a DAC loaded with He or Ar. The P-T dependences of the thermal conductivity of He are important to better model and constrain the structural and thermal evolution of gas giant planets containing He.