Gut microbiome features and metabolites in non-alcoholic fatty liver disease among community-dwelling middle-aged and older adults.

BACKGROUND: The specific microbiota and associated metabolites linked to non-alcoholic fatty liver disease (NAFLD) are still controversial. Thus, we aimed to understand how the core gut microbiota and metabolites impact NAFLD. METHODS: The data for the discovery cohort were collected from the Guangzhou Nutrition and Health Study (GNHS) follow-up conducted between 2014 and 2018. We collected 272 metadata points from 1546 individuals. The metadata were input into four interpretable machine learning models to identify important gut microbiota associated with NAFLD. These models were subsequently applied to two validation cohorts [the internal validation cohort (n = 377), and the prospective validation cohort (n = 749)] to assess generalizability. We constructed an individual microbiome risk score (MRS) based on the identified gut microbiota and conducted animal faecal microbiome transplantation experiment using faecal samples from individuals with different levels of MRS to determine the relationship between MRS and NAFLD. Additionally, we conducted targeted metabolomic sequencing of faecal samples to analyse potential metabolites. RESULTS: Among the four machine learning models used, the lightGBM algorithm achieved the best performance. A total of 12 taxa-related features of the microbiota were selected by the lightGBM algorithm and further used to calculate the MRS. Increased MRS was positively associated with the presence of NAFLD, with odds ratio (OR) of 1.86 (1.72, 2.02) per 1-unit increase in MRS. An elevated abundance of the faecal microbiota (f__veillonellaceae) was associated with increased NAFLD risk, whereas f__rikenellaceae, f__barnesiellaceae, and s__adolescentis were associated with a decreased presence of NAFLD. Higher levels of specific gut microbiota-derived metabolites of bile acids (taurocholic acid) might be positively associated with both a higher MRS and NAFLD risk. FMT in mice further confirmed a causal association between a higher MRS and the development of NAFLD. CONCLUSIONS: We confirmed that an alteration in the composition of the core gut microbiota might be biologically relevant to NAFLD development. Our work demonstrated the role of the microbiota in the development of NAFLD.

Transparent and Planar Cs3Cu2Cl5 Crystals for Micrometer-Resolution X-ray Imaging Screen.

Copper-based (I) halide perovskites have emerged as a promising candidate for scintillation screens in X-ray inspection and imaging areas due to their solution processability and high light yield. Here, a centimeter-sized Cs3Cu2Cl5 single crystal was grown by a slow-cooling method. The planar orientation was controlled in a space-confined chamber, generating a planar crystal which is readily used for a scintillation screen without any further shaping. The crystal exhibited a unity photoluminescence quantum yield and superior scintillation performance. The Cs3Cu2Cl5 single crystal exhibited a high light yield up to 95,000 photons/MeV, which enabled an X-ray detector of a detection limit down to 2.7 µGyair/s. The homemade imager demonstrated a spatial resolution of 105 lp/mm, representing an unprecedented micrometer resolution in laboratory. Importantly, the stability of Cs3Cu2Cl5 was significantly improved by a new surface passivation procedure, whereby the passivated crystal reserved its phase after 6 months' storage in a vial. This work introduced a new solution-based synthetic method for two-dimensional scintillating crystals, opening many avenues to high-performance X-ray imaging applications.

Afterglow Phosphor Goes Transparent.

Recently, transparent afterglow phosphors have attracted increasing interest due to the mitigated self-absorption and the ensuing improved light output, which have inspired many advanced applications, including volumetric display and three-dimensional optical encryption. To date, the most successful afterglow phosphors remain those traditional oxide, nitride, or sulfide powders which are not transparent due to a severe scattering effect. By reduction of the number of interfaces and engineering the refractive index, the scattering effect could be circumvented effectively. To this end, four material systems, including transparent afterglow single crystals, transparent phosphorescent organics, transparent afterglow glass, and luminescent nanocomposites, were reviewed in this Perspective. We started with the discussion of the nontransparency origin. Through a careful inspection of Rayleigh scattering theory, a general solution involving both refractive index and particle size was proposed to reduce the scattering effect. Many representative works on transparent afterglow phosphors were systematically reviewed, where the typical synthesis methods and the advantages and disadvantages of each system were critically presented. In the last part, bottlenecks, prospects, and future development directions based on transparent afterglow phosphors are proposed.

Long Non-Coding RNA ZSCAN16-AS1 Promotes the Malignant Progression of Melanoma Through Regulating the miR-503-5p/ARL2 Axis.

Background: LncRNA zinc finger and SCAN domain containing 16 antisense RNA 1 (ZSCAN16-AS1), a newly identified lncRNA, has been proven to accelerate hepatocellular carcinoma progression. However, the function and molecular mechanism of ZSCAN16-AS1 in melanoma are still unknown. Methods: The level of ZSCAN16-AS1 in melanoma tissues was detected and reported in The Cancer Genome Atlas (TCGA) and GEO#GSE15605. CCK-8, Transwell and flow cytometry assays were used to explore the role of ZSCAN16-AS1 in melanoma cells. Luciferase reporter assays and RNA pull-down assays were used to verify the molecular mechanism of ZSCAN16-AS1. Results: Here, we found that ZSCAN16-AS1 expression was increased in melanoma. We confirmed that ZSCAN16-AS1 promotes the growth and metastasis of melanoma. ZSCAN16-AS1 exerts its pro-tumour role through sponging of miR-503-5p to liberate ADP-ribosylation factor-like protein 2 (ARL2) mRNA transcripts. Conclusion: These results demonstrated the role and molecular mechanism of ZSCAN16-AS1 in the occurrence and development of melanoma. Therefore, ZSCAN16-AS1 may be used as a specific biomarker in the diagnosis and treatment of melanoma patients.

Prediction differences and implications of acute kidney injury with and without urine output criteria in adult critically ill patients.

BACKGROUND: Due to the convenience of serum creatinine (SCr) monitoring and the relative complexity of urine output (UO) monitoring, most studies have predicted acute kidney injury (AKI) only based on SCr criteria. This study aimed to compare the differences between SCr alone and combined UO criteria in predicting AKI. METHODS: We applied machine learning methods to evaluate the performance of 13 prediction models composed of different feature categories on 16 risk assessment tasks (half used only SCr criteria, half used both SCr and UO criteria). The area under receiver operator characteristic curve (AUROC), the area under precision recall curve (AUPRC) and calibration were used to assess the prediction performance. RESULTS: In the first week after ICU admission, the prevalence of any AKI was 29% under SCr criteria alone and increased to 60% when the UO criteria was combined. Adding UO to SCr criteria can significantly identify more AKI patients. The predictive importance of feature types with and without UO was different. Using only laboratory data maintained similar predictive performance to the full feature model under only SCr criteria [e.g. for AKI within the 48-h time window after 1 day of ICU admission, AUROC (95% confidence interval) 0.83 (0.82, 0.84) vs 0.84 (0.83, 0.85)], but it was not sufficient when the UO was added [corresponding AUROC (95% confidence interval) 0.75 (0.74, 0.76) vs 0.84 (0.83, 0.85)]. CONCLUSIONS: This study found that SCr and UO measures should not be regarded as equivalent criteria for AKI staging, and emphasizes the importance and necessity of UO criteria in AKI risk assessment.

Mixed traffic flow of human-driven vehicles and connected autonomous vehicles: String stability and fundamental diagram.

The introduction of connected autonomous vehicles (CAVs) gives rise to mixed traffic flow on the roadway, and the coexistence of human-driven vehicles (HVs) and CAVs may last for several decades. CAVs are expected to improve the efficiency of mixed traffic flow. In this paper, the car-following behavior of HVs is modeled by the intelligent driver model (IDM) based on actual trajectory data. The cooperative adaptive cruise control (CACC) model from the PATH laboratory is adopted for the car-following model of CAVs. The string stability of mixed traffic flow is analyzed for different market penetration rates of CAVs, showing that CAVs can effectively prevent stop-and-go waves from forming and propagating. In addition, the fundamental diagram is obtained from the equilibrium state, and the flow-density chart indicates that CAVs can improve the capacity of mixed traffic flow. Furthermore, the periodic boundary condition is designed for numerical simulation according to the infinite length platoon assumption in the analytical approach. The simulation results are consistent with the analytical solutions, suggesting the validity of the string stability and fundamental diagram analysis of mixed traffic flow.

Protein-ligand binding affinity prediction with edge awareness and supervised attention.

Accurate prediction of protein-ligand binding affinity is crucial in structure-based drug design but remains some challenges even with recent advances in deep learning: (1) Existing methods neglect the edge information in protein and ligand structure data; (2) current attention mechanisms struggle to capture true binding interactions in the small dataset. Herein, we proposed SEGSA_DTA, a SuperEdge Graph convolution-based and Supervised Attention-based Drug-Target Affinity prediction method, where the super edge graph convolution can comprehensively utilize node and edge information and the multi-supervised attention module can efficiently learn the attention distribution consistent with real protein-ligand interactions. Results on the multiple datasets show that SEGSA_DTA outperforms current state-of-the-art methods. We also applied SEGSA_DTA in repurposing FDA-approved drugs to identify potential coronavirus disease 2019 (COVID-19) treatments. Besides, by using SHapley Additive exPlanations (SHAP), we found that SEGSA_DTA is interpretable and further provides a new quantitative analytical solution for structure-based lead optimization.

Analysis of Energy and Density in Treating Hypertrophic Scar After Burn in Children with CO2 Dot Matrix Laser.

Objective: To analyze and compare the effect of the combination of energy and density parameters of CO2 dot matrix laser in the hyperplastic stage of pediatric burn. Materials and Methods: A total of 160 pediatric patients with hypertrophic scar after limb burn from 2017 to 2020 were randomly divided into four parameter groups (n = 40). The patients were treated with ablative fraction carbon dioxide laser, once every 10 weeks. During the interval of laser treatment, Compound Heparin Sodium and Allantoin Gel (Contractubex) was applied externally, tid, and elastic cover or elastic bandage is attached to the affected limb. Scoring based on the Vancouver Scar Scale is performed before each laser treatment, The score before the first treatment was the initial score, which was scored by two people separately, and the average score was calculated. Subsequently, the patients were treated four times and scored. The differences between each treatment and the first score of each parameter group were compared. Under the same energy and different treatment density, the scores after each treatment were compared. Under the same density and different energy, the scores after each treatment were compared. The bleeding and pigmentation of each parameter group were compared. Results: The increase of density can show the therapeutic effect earlier than the increase of energy, and 25mj energy and 10% density have better intervention effect. With the course of disease and the progress of treatment, the correlation between intervention effect and parameters tends to weaken. Comparing the number of cases with different scores between each treatment and the first time, the score in the 5% density group was lower than that in the 10% density group, but there was no significant difference between the 25mj and 17.5mj energy levels in the same density group. The intervention effect of the increase of density on scar was better than that of energy, and the increase of energy and density could aggravate the pain. Conclusion: In pediatric burn hypertrophic scars treated by CO2 dot matrix laser in exfoliation mode, the intervention effect of increasing density is better than that of energy. When setting laser treatment parameters, we should give priority to increasing density and adjust energy according to the effect of treatment and the condition of pain, bleeding and color precipitation. In this study, the best combination of parameters is 17.5mj/10%.

A Transfer Learning Approach to Correct the Temporal Performance Drift of Clinical Prediction Models: Retrospective Cohort Study.

BACKGROUND: Clinical prediction models suffer from performance drift as the patient population shifts over time. There is a great need for model updating approaches or modeling frameworks that can effectively use the old and new data. OBJECTIVE: Based on the paradigm of transfer learning, we aimed to develop a novel modeling framework that transfers old knowledge to the new environment for prediction tasks, and contributes to performance drift correction. METHODS: The proposed predictive modeling framework maintains a logistic regression-based stacking ensemble of 2 gradient boosting machine (GBM) models representing old and new knowledge learned from old and new data, respectively (referred to as transfer learning gradient boosting machine [TransferGBM]). The ensemble learning procedure can dynamically balance the old and new knowledge. Using 2010-2017 electronic health record data on a retrospective cohort of 141,696 patients, we validated TransferGBM for hospital-acquired acute kidney injury prediction. RESULTS: The baseline models (ie, transported models) that were trained on 2010 and 2011 data showed significant performance drift in the temporal validation with 2012-2017 data. Refitting these models using updated samples resulted in performance gains in nearly all cases. The proposed TransferGBM model succeeded in achieving uniformly better performance than the refitted models. CONCLUSIONS: Under the scenario of population shift, incorporating new knowledge while preserving old knowledge is essential for maintaining stable performance. Transfer learning combined with stacking ensemble learning can help achieve a balance of old and new knowledge in a flexible and adaptive way, even in the case of insufficient new data.

Flexible and Transparent Ceramic Nanocomposite for Laboratory X-ray Imaging of Micrometer Resolution.

Transparent nanocomposites have attracted considerable attention in many areas including X-ray imaging, wearable electronics, and volumetric display. However, both the transparency and the flexibility were largely jeopardized by the loading content of functional nanoparticles (NPs), posing a major challenge to material engineering. Herein, an ultra-high-loading-ceramic nanocomposite film was fabricated by a blade-coating technique. The film exhibited a high transparency over ∼89% in the whole visible region even with a fluoride-ceramic content up to ∼83 wt %. Based on a real-time investigation on the formation process of the film, the refractive-index difference between the nanoparticles and matrix was identified as the dominating factor to transparency. The transmittance spectra based on Rayleigh scattering theory were simulated to screen both nanoparticle radius and loading content, leading to the discovery of a transparency zone for film making. As a proof-of-concept experiment, the transparent film was used as an X-ray scintillation screen, which exhibited a comparable light yield to that of LYSO owing to the mitigated self-absorption effect. The homemade imager demonstrated a spatial resolution of 122 lp/mm, representing a record resolution of 4.1 µm for laboratory X-ray photography. Our work not only provided an experimental procedure to make high-loading functional films but also demonstrated a theoretical model to guide the search for gradients of transparent composites.

Trap-tuning in afterglow perovskite crystals through alkali metal ion doping.

Herein, we report a trap-tuning strategy involving alkali metal ions of Li+ and K+, which enabled a fine tuning of afterglow traps at 250 K. Through Li+/K+ doping, a red afterglow from Cs2AgInCl6:Mn was activated at room temperature, which was extended up to 3600 s at low temperature. Thermoluminescence measurement revealed a shallow trap at ∼0.50 eV below the conduction band, which could be shifted up- and down-ward by dopant species.

hUC-MSCs exosomal miR-451 alleviated acute lung injury by modulating macrophage M2 polarization via regulating MIF-PI3K-AKT signaling pathway.

In the previous study, we have proved that exosomal miR-451 from human umbilical cord mesenchymal stem cells (hUC-MSCs) attenuated burn-induced acute lung injury (ALI). However, the mechanism of exosomal miR-451 in ALI remains unclear. Therefore, this study aimed to study the molecular mechanism of hUC-MSCs-derived exosomal miR-451 on ALI by regulating macrophage polarization. Exosomes were isolated and identified by transmission electron microscope (TEM) and nanoparticle tracking analysis (NTA). The expression of miR-451, macrophage migration inhibitory factor (MIF) and PI3K/AKT signaling pathway proteins were detected by qRT-PCR and western blot. Flow cytometry was used to detect the CD80 and CD206 positive cells. Severe burn rat model was established and HE was used to detect the inflammatory cell infiltration and inflammatory injury. Dual luciferase reporter system was used to detect the regulation of miR-451 to MIF. The contents of cytokines were detected by ELISA. The results showed that hUC-MSCs exosomes promoted macrophage M1 to M2 polarization. Furthermore, hUC-MSCs-derived exosomal miR-451 alleviated ALI development and promoted macrophage M1 to M2 polarization. Moreover, MIF was a direct target of miR-451. Downregulation of MIF regulated by miR-451 alleviated ALI development promoted macrophage M1 to M2 polarization. In addition, we found that MIF and hUC-MSCs-derived exosomal miR-451 participated in ALI by regulating PI3K/AKT signaling pathway. In conclusion, we indicated that hUC-MSCs-derived exosomal miR-451 alleviated ALI by modulating macrophage M2 polarization via regulating MIF-PI3K-AKT signaling pathway, which provided great scientific significance and clinical application value for the treatment of burn-induced ALI.

Mn2+-Activated Afterglow in a Transparent Perovskite Crystal.

Transparent crystals of long afterglow are keenly desired for advanced applications involving three-dimensional information storage and volumetric display. Here, we report the growth of a perovskite crystal, Cs2NaInCl6:Mn, via a facile hydrothermal reaction. The crystal featured a high transparency up to 90% in a broad range from 400 to 750 nm. Upon ultraviolet (UV) excitation, the crystal exhibited an overlapped emission from both self-trapped exciton (STE) and Mn ions. An unusual energy transfer from Mn ion to STE was revealed through steady-state and transient photoluminescence (PL) spectra. Through concentration tuning, the afterglow duration was extended to 2500 s with a signal-to-noise ratio over 20. The single dopant of Mn ions was found to play two roles in afterglow mechanism, including both electron traps and emitting activators. This work provided a facile method to synthesize transparent crystals of persistent luminescence, opening many avenues for three-dimensional information storage and volumetric display.

Nearly-Unity Quantum Yield and 12-Hour Afterglow from a Transparent Perovskite of Cs2 NaScCl6 :Tb.

Unlike afterglow phosphors, transparent crystals free of scattering issues hold great potential in applications such as volumetric displays and three-dimensional encryption. Here, a double-perovskite host of ca. 85 % transparency, a Cs2 NaScCl6 single crystal, was grown in aqueous solution. Through heavy doping of Tb ions, the crystal exhibited a high photoluminescence quantum yield (PL QY) up to 98.2 %. Both steady-state and transient PL spectroscopic investigation revealed a cross-relaxation efficiency (ηCR ≈99 %) between Tb ions, greatly mitigated the electron saturation on the 5 D3 level and encouraged the radiative route against quenching. The doped crystal displayed an intense afterglow of up to 12 h after ceasing the X-ray excitation, allowing it to be used in a radiation-storage battery which has a linear scaling capacity with loading weight. Our work opens possibilities in photo-stimulated volumetric display and three-dimensional information storage.

Development and Validation of a Personalized Model With Transfer Learning for Acute Kidney Injury Risk Estimation Using Electronic Health Records.

Importance: Acute kidney injury (AKI) is a heterogeneous syndrome prevalent among hospitalized patients. Personalized risk estimation and risk factor identification may allow effective intervention and improved outcomes. Objective: To develop and validate personalized AKI risk estimation models using electronic health records (EHRs), examine whether personalized models were beneficial in comparison with global and subgroup models, and assess the heterogeneity of risk factors and their outcomes in different subpopulations. Design, Setting, and Participants: This diagnostic study analyzed EHR data from 1 tertiary care hospital and used machine learning and logistic regression to develop and validate global, subgroup, and personalized risk estimation models. Transfer learning was implemented to enhance the personalized model. Predictor outcomes across subpopulations were analyzed, and metaregression was used to explore predictor interactions. Adults who were hospitalized for 2 or more days from November 1, 2007, to December 31, 2016, were included in the analysis. Patients with moderate or severe kidney dysfunction at admission were excluded. Data were analyzed between August 28, 2019, and May 8, 2022. Exposures: Clinical and laboratory variables in the EHR. Main Outcomes and Measures: The main outcome was AKI of any severity, and AKI was defined using the Kidney Disease: Improving Global Outcomes serum creatinine criteria. Performance of the models was measured with area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve, and calibration. Results: The study cohort comprised 76 957 inpatient encounters. Patients had a mean (SD) age of 55.5 (17.4) years and included 42 159 men (54.8%). The personalized model with transfer learning outperformed the global model for AKI estimation in terms of AUROC among general inpatients (0.78 [95% CI, 0.77-0.79] vs 0.76 [95% CI, 0.75-0.76]; P < .001) and across the high-risk subgroups (0.79 [95% CI, 0.78-0.80] vs 0.75 [95% CI, 0.74-0.77]; P < .001) and low-risk subgroups (0.74 [95% CI, 0.73-0.75] vs 0.71 [95% CI, 0.70-0.72]; P < .001). The AUROC improvement reached 0.13 for the high-risk subgroups, such as those undergoing liver transplant and cardiac surgery. Moreover, the personalized model with transfer learning performed better than or comparably with the best published models in well-studied AKI subgroups. Predictor outcomes varied significantly between patients, and interaction analysis uncovered modifiers of the predictor outcomes. Conclusions and Relevance: Results of this study demonstrated that a personalized modeling with transfer learning is an improved AKI risk estimation approach that can be used across diverse patient subgroups. Risk factor heterogeneity and interactions at the individual level highlighted the need for agile, personalized care.

Characterizing the temporal changes in association between modifiable risk factors and acute kidney injury with multi-view analysis.

BACKGROUND: Acute kidney injury (AKI) is a common life-threatening clinical syndrome in hospitalized patients. Advances in machine learning has demonstrated success in AKI risk prediction using electronic health records (EHRs). However, to prevent AKI, it is critical to identify clinically modifiable factors and understand their impact at different prevention windows. METHOD: We extracted 4129 clinical variables including demographics, social history, past diagnoses, procedures, labs, medications, vitals from EHRs for a cohort of 144,084 eligible inpatient encounters. We developed a multi-view learning framework for XGBoost (MV-XGB) to enhance algorithm attention on modifiable factors. To study effects of modifiable factors at different time points, we built AKI prediction models at 24-hours, 48-hours, 72-hours before AKI onset. To characterize the temporal changes in effect of modifiable factors on AKI, we derived two indicators, inter-class score-difference and exposed-score-difference, based on SHAP values to compare effects of modifiable factors in different windows. RESULT: MV-XGB effectively increased attention on modifiable factors (explained 92.4%-94.1% inter-class score-difference, i.e., predictive difference between AKI and non-AKI samples) while maintaining good predictive performance (AUROCs were 0.854, 0.798, 0.765 in models for 24-48-72 h AKI prediction respectively). We observed that 62% of predicted odds-ratio difference between AKI and non-AKI patients in 24 h can be explained by factors occurring between 24 and 72 h. Among the important modifiable factors, electrolyte balance explained 38.3% of the inter-class score difference increase between 24 h and 72 h, followed by high-risk medications (13.7%), care strategy (12.1%), blood pressure (10%), infection (7.8%), and anemia (5.4%). Effects of cardiac surgery or condition, respiratory ventilation, and anemia remained important longer than 72 h. CONCLUSION: Better understanding of the clinically modifiable factors is important to AKI prevention. The proposed multi-view learning approach improved the identification of modifiable factors of AKI and allowed characterization of the temporal dynamics of their potential benefit in intervention.

Analysis of the therapeutic effect of artificial leather embedding combined with fascial sleeve flap transplantation on chronic wounds of lower limbs with bone and plate exposure.

BACKGROUND: After severe trauma of lower limbs, bone, tendon or plate graft exposure is common. The traditional repair method is to use a variety of skin flap transplantation to cover the exposed part, but the wound often can not heal after operation, or the wound is cracked, ulcer, sinus, bone and steel plate are exposed again after wound healing. The reason for this result is that when the flap is covered, the space around the bone plate is not well closed, forming a dead cavity, blood and exudate accumulation, hematoma formation or infection, and finally the wound ruptures again. In addition, due to the swelling and contracture of the flap after operation, the suture tension between the flap and the receiving area becomes larger, the skin becomes thinner and broken, and then the wound is formed. In order to solve the above problems, we carried out the study of artificial true skin embedding combined with fascial sleeve flap transplantation in the treatment of chronic bone plate exposed wounds of lower limbs. METHODS: In this paper, 11 cases of chronic wounds with bone exposure and skin necrosis after steel plate implantation were selected. First stage is the wound bed preparation including primary wound expansion, removal of necrotic tissue and incision of sinus wall, removal of deep necrotic bone and fibrotic scarred skin on the outer wall of steel plate to normal tissue on the outer edge of the wound, removal of precipitated peptone and purulent fur in the hole, periphery and bone space of the steel plate, and removal of tendon tissue with basal necrosis and disintegration of the wound. After vacuum sealing drainage (VSD) 1-2 weeks, the peritraumatic basal granulation tissue grew well and there was no necrotic tissue in the wound. In the second stage, the exposed bone was covered with artificial dermis, the steel plate hole or the periphery and the basal space were filled, and the exposed steel plate was completely embedded, and then the fascia sleeve flap was transplanted to cover the wound. The sural neurovascular flap was performed in nine cases and the lateral superior malleolar artery perforator flap in two case. RESULTS: The flap survived well in all 11 cases. During the follow-up of 6 months to the removal of the plate, there was no case of rupture, exposure and sinus formation. CONCLUSIONS: Artificial dermal covering combined with fascial sleeve flap transplantation can effectively avoid wound dehiscence or sinus formation caused by foreign body retention, infection and flap contracture. It has good effect in repairing chronic wounds with bone plate exposure after severe trauma of lower limbs.

A hybrid adaptive approach for instance transfer learning with dynamic and imbalanced data.

Machine learning has demonstrated success in clinical risk prediction modeling with complex electronic health record data. However, the evolving nature of clinical practices can dynamically change the underlying data distribution over time, leading to model performance drift. Adopting an outdated model is potentially risky and may result in unintentional losses. In this paper, we propose a novel Hybrid Adaptive Boosting approach (HA-Boost) for transfer learning. HA-Boost is characterized by the domain similarity-based and class imbalance-based adaptation mechanisms, which simultaneously address two critical limitations of the classical TrAdaBoost algorithm. We validated HA-Boost in predicting hospital-acquired acute kidney injury using real-world longitudinal electronic health records data. The experiment results demonstrate that HA-Boost stably outperforms the competing baselines in terms of both AUROC and AUPRC across a 7-year time span. This study has confirmed the effectiveness of transfer learning as a superior model updating approach in dynamic environment.

Development of a knowledge mining approach to uncover heterogeneous risk predictors of acute kidney injury across age groups.

OBJECTIVES: Acute kidney injury (AKI) risk increases with age and the underlying clinical predictors may be heterogeneous across age strata. This study aims to uncover the AKI risk factor heterogeneity among general inpatients across age groups using electronic medical records (EMR). METHODS: Patient data (n = 179,370 encounters) were collected from an academic hospital between 2007 and 2016, and were stratified into four age groups: 18-35, 36-55, 56-65, and > 65. Potential risk factors extracted for the cohort included demographics, vital signs, laboratory values, past medical diagnoses, medications and admission diagnoses. We developed a data driven knowledge mining approach consisting of a machine learning algorithm to identify AKI predictors across age strata and a statistical method to quantify the impact of those factors on AKI risk. Identified predictors were evaluated for their predictability of AKI in terms of area-under-the-receiver-operating-characteristic-curve (AUC) and validated against expert knowledge. RESULTS: Among the final analysis cohort of 76,957 hospital admissions, AKI prediction across age groups 18-35 (16.73%), 36-55 (32.74%), 56-65 (23.52%), and > 65 years (27.01%) achieved AUC of 0.85 (95% CI, 0.80-0.88), 0.86 (95% CI, 0.83-0.89), 0.87 (95% CI, 0.86-0.90), and 0.87 (95% CI, 0.86-0.90), respectively. Compared to expert knowledge, absolute consistency rates of the top-150 identified risk factors for each group were 78.4%, 77.2%, 81.3%, and 79.5%, respectively. Impact of many predictors on AKI varied across age groups; for example, high body mass index (BMI) was found to be associated with higher AKI risk in elderly patients, but low BMI was found to be associated with higher AKI risk in younger patients. CONCLUSIONS: We verified the effectiveness of the knowledge mining method from the perspectives of accuracy, stability and credibility, and used this approach to clarify the heterogeneity of AKI risk factors between age groups. Future decision support systems need to consider such heterogeneity to enhance personalized patient care.