Hui Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Hui Chen has a total of 287 co-authors. Here are the top co-authors by number of shared publications:
- Honglei Liu
Capital Medical University
13 shared publications Xiaojie Huang
Center for Infectious Diseases, Beijing You'an Hospital, Capital Medical University, Beijing, China
9 shared publications- Hao Wu
Center for Infectious Diseases, Beijing You'an Hospital, Capital Medical University, Beijing, China
8 shared publications - Muyu Wang
School of Biomedical Engineering, Capital Medical University, No.10, Xitoutiao, You An Men, Fengtai District, Beijing 100069, China; Beijing Key Laboratory of Fundamental Research on Biomechanics in Clinical Application, Capital Medical University, No.10, Xitoutiao, You An Men, Fengtai District, Beijing 100069, China
8 shared publications Tong Zhang
Center for Infectious Diseases, Beijing You'an Hospital, Capital Medical University, Beijing, China
8 shared publications- Xiaolu Fei
Information Center, Xuanwu Hospital, Capital Medical University, Beijing, China
8 shared publications
Latest Publications
- Open Access
- 20 July 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Knowledge-Augmented Spatiotemporal Modeling of Electronic Medical Records for Heart Failure Prognosis Prediction: Development and Validation of KA-STNet (Preprint)
BACKGROUND Heart failure (HF) is a severe and progressive clinical syndrome associated with high mortality, frequent hospitalizations, and substantial healthcare burden worldwide. Identification of critically ill HF patients and accurate assessment of cardiac function grade at discharge are essential for timely intervention, risk stratification, and optimization of post-discharge management. OBJECTIVE This study aims to develop and validate KA-STNet, an interpretable knowledge-augmented spatiotemporal framework for in-hospital mortality prediction and discharge cardiac function grade prediction in patients with HF. METHODS To address the heterogeneity of EMR data, KA-STNet adopts distinct graph representation strategies for static and time-series data. For static data, it integrates external medical knowledge with data-driven statistical associations to learn knowledge-augmented patient representations. For time-series data, it employs a spatiotemporal graph neural network to capture spatial dependencies among laboratory indicators and temporal dynamics in longitudinal laboratory sequences, while incorporating patient-level static representations to guide inter-laboratory association modeling. Interpretability analyses based on Shapley additive explanations (SHAP), GNNExplainer, and Leiden community detection are conducted to identify key predictive factors and clinically meaningful laboratory communities. RESULTS KA-STNet outperformed all baseline models across all evaluated tasks. In the in-hospital mortality risk prediction task, it attained a mean AUROC of 0.968 on the local dataset and 0.897 on the public dataset. In the discharge cardiac function grade prediction task, KA-STNet achieved a mean AUROC of 0.925. Anemia-related variables, indicators of renal function and infection-related markers were identified as important contributors to adverse outcome prediction in patients with HF. Leiden community detection further revealed meaningful multi-organ interaction patterns. Adverse outcomes were associated with network reorganization involving acid–base/electrolyte, inflammatory, and hepatic indicators. CONCLUSIONS KA-STNet provides an effective and interpretable framework for predicting HF outcomes using heterogeneous EMRs. These findings support the value of combining knowledge-augmented static representation learning with spatiotemporal dependency modeling for interpretable prediction of clinically relevant HF outcomes. CLINICALTRIAL Not applicable. This was a retrospective study based on electronic medical records and was not a clinical trial. INTERNATIONAL REGISTERED REPORT RR2-
- 20 July 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessKnowledge-Augmented Spatiotemporal Modeling of Electronic Medical Records for Heart Failure Prognosis Prediction: Development and Validation of KA-STNet (Preprint)
BACKGROUND Heart failure (HF) is a severe and progressive clinical syndrome associated with high mortality, frequent hospitalizations, and substantial healthcare burden worldwide. Identification of critically ill HF patients and accurate assessment of cardiac function grade at discharge are essential for timely intervention, risk stratification, and optimization of post-discharge management. OBJECTIVE This study aims to develop and validate KA-STNet, an interpretable knowledge-augmented spatiotemporal framework for in-hospital mortality prediction and discharge cardiac function grade prediction in patients with HF. METHODS To address the heterogeneity of EMR data, KA-STNet adopts distinct graph representation strategies for static and time-series data. For static data, it integrates external medical knowledge with data-driven statistical associations to learn knowledge-augmented patient representations. For time-series data, it employs a spatiotemporal graph neural network to capture spatial dependencies among laboratory indicators and temporal dynamics in longitudinal laboratory sequences, while incorporating patient-level static representations to guide inter-laboratory association modeling. Interpretability analyses based on Shapley additive explanations (SHAP), GNNExplainer, and Leiden community detection are conducted to identify key predictive factors and clinically meaningful laboratory communities. RESULTS KA-STNet outperformed all baseline models across all evaluated tasks. In the in-hospital mortality risk prediction task, it attained a mean AUROC of 0.968 on the local dataset and 0.897 on the public dataset. In the discharge cardiac function grade prediction task, KA-STNet achieved a mean AUROC of 0.925. Anemia-related variables, indicators of renal function and infection-related markers were identified as important contributors to adverse outcome prediction in patients with HF. Leiden community detection further revealed meaningful multi-organ interaction patterns. Adverse outcomes were associated with network reorganization involving acid–base/electrolyte, inflammatory, and hepatic indicators. CONCLUSIONS KA-STNet provides an effective and interpretable framework for predicting HF outcomes using heterogeneous EMRs. These findings support the value of combining knowledge-augmented static representation learning with spatiotemporal dependency modeling for interpretable prediction of clinically relevant HF outcomes. CLINICALTRIAL Not applicable. This was a retrospective study based on electronic medical records and was not a clinical trial. INTERNATIONAL REGISTERED REPORT RR2-
- Open Access
- 29 April 2026
- Published by Elsevier in SSRN Electronic Journal
Time-Updated Prediction of Crohn’s Disease Progression Using Multicenter Longitudinal Electronic Health Records
Background: Crohn’s disease (CD) has an evolving course and may progress to stricturing or penetrating complications, intestinal surgery, and treatment escalation. Existing models often rely on baseline or time-agnostic data, limiting risk assessment during routine follow-up. We developed and externally validated a time-updated model CD-ProFormer for prediction of CD progression using longitudinal electronic health records (EHRs).Methods: We assembled multicenter EHR cohorts: a development cohort from Peking Union Medical College Hospital (PUMCH) (761 patients; 11,984 visits) and external validation cohorts from Guizhou (74 patients; 510 visits) and Nanjing (271 patients; 3934 visits). Using each visit as index, CD-ProFormer, a time-aware Transformer that used EHR trajectories to predict 1-, 3-, and 5-year risks of behavior progression, CD-related intestinal surgery, and major medication initiation. Global SHAP, temporal SHAP, decision curve analysis, and representative cases assessed interpretability and clinical utility.Findings: CD-ProFormer showed strong internal discrimination for behavior progression, with AUROCs of 0.979, 0.951, and 0.910 at 1, 3, and 5 years; macro-area under the curve for first-progression time-window classification was 0.811. External validation remained acceptable, with AUROCs of 0.725–0.879 in Guizhou and 0.856–0.913 in Nanjing, consistently exceeding recurrent baselines. Performance for surgery and medication initiation was comparable to recurrent models and better than conventional models.Interpretation: In this prognostic study, CD-ProFormer enabled time-updated, horizon-specific prediction of CD behavior progression, with surgery and medication initiation assessed in parallel. These findings suggest that prognosis in CD can be updated at follow-up visits using routinely collected longitudinal data and may support risk stratification, monitoring intensity, and timely treatment planning.
- 29 April 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessTime-Updated Prediction of Crohn’s Disease Progression Using Multicenter Longitudinal Electronic Health Records
Background: Crohn’s disease (CD) has an evolving course and may progress to stricturing or penetrating complications, intestinal surgery, and treatment escalation. Existing models often rely on baseline or time-agnostic data, limiting risk assessment during routine follow-up. We developed and externally validated a time-updated model CD-ProFormer for prediction of CD progression using longitudinal electronic health records (EHRs).Methods: We assembled multicenter EHR cohorts: a development cohort from Peking Union Medical College Hospital (PUMCH) (761 patients; 11,984 visits) and external validation cohorts from Guizhou (74 patients; 510 visits) and Nanjing (271 patients; 3934 visits). Using each visit as index, CD-ProFormer, a time-aware Transformer that used EHR trajectories to predict 1-, 3-, and 5-year risks of behavior progression, CD-related intestinal surgery, and major medication initiation. Global SHAP, temporal SHAP, decision curve analysis, and representative cases assessed interpretability and clinical utility.Findings: CD-ProFormer showed strong internal discrimination for behavior progression, with AUROCs of 0.979, 0.951, and 0.910 at 1, 3, and 5 years; macro-area under the curve for first-progression time-window classification was 0.811. External validation remained acceptable, with AUROCs of 0.725–0.879 in Guizhou and 0.856–0.913 in Nanjing, consistently exceeding recurrent baselines. Performance for surgery and medication initiation was comparable to recurrent models and better than conventional models.Interpretation: In this prognostic study, CD-ProFormer enabled time-updated, horizon-specific prediction of CD behavior progression, with surgery and medication initiation assessed in parallel. These findings suggest that prognosis in CD can be updated at follow-up visits using routinely collected longitudinal data and may support risk stratification, monitoring intensity, and timely treatment planning.
- 6 April 2025
- Published by Elsevier in Journal of Biomedical Informatics
- Vol. 165, 104823
Multi-modal fusion model for Time-Varying medical Data: Addressing Long-Term dependencies and memory challenges in sequence fusion
- 6 April 2025
- Published by Elsevier in Journal of Biomedical Informatics
- Vol. 165, 104823
Multi-modal fusion model for Time-Varying medical Data: Addressing Long-Term dependencies and memory challenges in sequence fusion
- 22 February 2025
- Published by Elsevier in Journal of Biomedical Informatics
- Vol. 164, 104796
Missing-modality enabled multi-modal fusion architecture for medical data
- 22 February 2025
- Published by Elsevier in Journal of Biomedical Informatics
- Vol. 164, 104796
Missing-modality enabled multi-modal fusion architecture for medical data
- Open Access
- 18 December 2024
- Published by Elsevier in Computer Methods and Programs in Biomedicine
- Vol. 260, 108568
Robust multi-modal fusion architecture for medical data with knowledge distillation
- 18 December 2024
- Published by Elsevier in Computer Methods and Programs in Biomedicine
- Vol. 260, 108568
Open AccessRobust multi-modal fusion architecture for medical data with knowledge distillation
- Open Access
- 11 March 2024
- Published by Wolters Kluwer Health in International Journal of Surgery
- Vol. 110 (6) , 3237-3248
Multimodal learning system integrating electronic medical records and hysteroscopic images for reproductive outcome prediction and risk stratification of endometrial injury: a multicenter diagnostic study
Objective: To develop a multimodal learning application system that integrates electronic medical records (EMR) and hysteroscopic images for reproductive outcome prediction and risk stratification of patients with intrauterine adhesions (IUAs) resulting from endometrial injuries. Materials and methods: EMR and 5014 revisited hysteroscopic images of 753 post hysteroscopic adhesiolysis patients from the multicenter IUA database we established were randomly allocated to training, validation, and test datasets. The respective datasets were used for model development, tuning, and testing of the multimodal learning application. MobilenetV3 was employed for image feature extraction, and XGBoost for EMR and image feature ensemble learning. The performance of the application was compared against the single-modal approaches (EMR or hysteroscopic images), DeepSurv and ElasticNet models, along with the clinical scoring systems. The primary outcome was the 1-year conception prediction accuracy, and the secondary outcome was the assisted reproductive technology (ART) benefit ratio after risk stratification. Results: The multimodal learning system exhibited superior performance in predicting conception within 1-year, achieving areas under the curves of 0.967 (95% CI: 0.950–0.985), 0.936 (95% CI: 0.883–0.989), and 0.965 (95% CI: 0.935–0.994) in the training, validation, and test datasets, respectively, surpassing single-modal approaches, other models and clinical scoring systems (all PP=0.02), while low-risk patients exhibited good natural conception potential, with no significant increase in conception rates from ART treatment (P=1). Conclusions: The multimodal learning system using hysteroscopic images and EMR demonstrates promise in accurately predicting the natural conception of patients with IUAs and providing effective postoperative stratification, potentially contributing to ART triage after IUA procedures.
- 11 March 2024
- Published by Wolters Kluwer Health in International Journal of Surgery
- Vol. 110 (6) , 3237-3248
Open AccessMultimodal learning system integrating electronic medical records and hysteroscopic images for reproductive outcome prediction and risk stratification of endometrial injury: a multicenter diagnostic study
Objective: To develop a multimodal learning application system that integrates electronic medical records (EMR) and hysteroscopic images for reproductive outcome prediction and risk stratification of patients with intrauterine adhesions (IUAs) resulting from endometrial injuries. Materials and methods: EMR and 5014 revisited hysteroscopic images of 753 post hysteroscopic adhesiolysis patients from the multicenter IUA database we established were randomly allocated to training, validation, and test datasets. The respective datasets were used for model development, tuning, and testing of the multimodal learning application. MobilenetV3 was employed for image feature extraction, and XGBoost for EMR and image feature ensemble learning. The performance of the application was compared against the single-modal approaches (EMR or hysteroscopic images), DeepSurv and ElasticNet models, along with the clinical scoring systems. The primary outcome was the 1-year conception prediction accuracy, and the secondary outcome was the assisted reproductive technology (ART) benefit ratio after risk stratification. Results: The multimodal learning system exhibited superior performance in predicting conception within 1-year, achieving areas under the curves of 0.967 (95% CI: 0.950–0.985), 0.936 (95% CI: 0.883–0.989), and 0.965 (95% CI: 0.935–0.994) in the training, validation, and test datasets, respectively, surpassing single-modal approaches, other models and clinical scoring systems (all PP=0.02), while low-risk patients exhibited good natural conception potential, with no significant increase in conception rates from ART treatment (P=1). Conclusions: The multimodal learning system using hysteroscopic images and EMR demonstrates promise in accurately predicting the natural conception of patients with IUAs and providing effective postoperative stratification, potentially contributing to ART triage after IUA procedures.
- 1 February 2024
- Published by SAGE Publications in AIDS Research and Human Retroviruses
- Vol. 40 (2) , 101-109
Plasma Cytokine Expression and Immune Reconstitution in Early and Delayed Anti-HIV 96-Weeks Treatment: A Retrospective Study
HIV is an immunodeficiency disease with emergence of inadequate corresponding reconstruction therapies. Pyroptosis of CD4+T cell is mainly caused by immune activation and inflammation that cannot be reduced by successful antiretroviral therapy (ART) alone. Co-infections because of CD4+T cell reconstitution failure can occur. Anti-inflammatory treatment determines the success of immune reconstitution. In our experiment, only a few cytokines could recover to normal level following with a 2-year antiretroviral treatment in early ART initiation, which is consistent with current findings about adjuvant HIV anti-inflammatory therapy. Early infection is often accompanied by a more severe inflammatory response. Innate immunity cytokines like GM-CSF, IP-10, and TNF-α exhibited the most elevated levels among all kinds of inflammatory cytokines. The correlation analysis showed at least eight cytokines contributing to the changes of CD4/CD8 ratio.
- 1 February 2024
- Published by SAGE Publications in AIDS Research and Human Retroviruses
- Vol. 40 (2) , 101-109
Plasma Cytokine Expression and Immune Reconstitution in Early and Delayed Anti-HIV 96-Weeks Treatment: A Retrospective Study
HIV is an immunodeficiency disease with emergence of inadequate corresponding reconstruction therapies. Pyroptosis of CD4+T cell is mainly caused by immune activation and inflammation that cannot be reduced by successful antiretroviral therapy (ART) alone. Co-infections because of CD4+T cell reconstitution failure can occur. Anti-inflammatory treatment determines the success of immune reconstitution. In our experiment, only a few cytokines could recover to normal level following with a 2-year antiretroviral treatment in early ART initiation, which is consistent with current findings about adjuvant HIV anti-inflammatory therapy. Early infection is often accompanied by a more severe inflammatory response. Innate immunity cytokines like GM-CSF, IP-10, and TNF-α exhibited the most elevated levels among all kinds of inflammatory cytokines. The correlation analysis showed at least eight cytokines contributing to the changes of CD4/CD8 ratio.
- Open Access
- 18 January 2024
- Published by MDPI AG in Bioengineering
- Vol. 11 (1) , 94
Research on Multimodal Fusion of Temporal Electronic Medical Records
The surge in deep learning-driven EMR research has centered on harnessing diverse data forms. Yet, the amalgamation of diverse modalities within time series data remains an underexplored realm. This study probes a multimodal fusion approach, merging temporal and non-temporal clinical notes along with tabular data. We leveraged data from 1271 myocardial infarction and 6450 stroke inpatients at a Beijing tertiary hospital. Our dataset encompassed static, and time series note data, coupled with static and time series table data. The temporal data underwent a preprocessing phase, padding to a 30-day interval, and segmenting into 3-day sub-sequences. These were fed into a long short-term memory (LSTM) network for sub-sequence representation. Multimodal attention gates were implemented for both static and temporal subsequence representations, culminating in fused representations. An attention-backtracking module was introduced for the latter, adept at capturing enduring dependencies in temporal fused representations. The concatenated results were channeled into an LSTM to yield the ultimate fused representation. Initially, two note modalities were designated as primary modes, and subsequently, the proposed fusion model was compared with comparative models including recent models such as Crossformer. The proposed model consistently exhibited superior predictive prowess in both tasks. Removing the attention-backtracking module led to performance decline. The proposed model consistently shows excellent predictive capabilities in both tasks. The proposed method not only effectively integrates data from the four modalities, but also has a good understanding of how to handle irregular time series data and lengthy clinical texts. An effective method is provided, which is expected to be more widely used in multimodal medical data representation.
- 18 January 2024
- Published by MDPI AG in Bioengineering
- Vol. 11 (1) , 94
Open AccessResearch on Multimodal Fusion of Temporal Electronic Medical Records
The surge in deep learning-driven EMR research has centered on harnessing diverse data forms. Yet, the amalgamation of diverse modalities within time series data remains an underexplored realm. This study probes a multimodal fusion approach, merging temporal and non-temporal clinical notes along with tabular data. We leveraged data from 1271 myocardial infarction and 6450 stroke inpatients at a Beijing tertiary hospital. Our dataset encompassed static, and time series note data, coupled with static and time series table data. The temporal data underwent a preprocessing phase, padding to a 30-day interval, and segmenting into 3-day sub-sequences. These were fed into a long short-term memory (LSTM) network for sub-sequence representation. Multimodal attention gates were implemented for both static and temporal subsequence representations, culminating in fused representations. An attention-backtracking module was introduced for the latter, adept at capturing enduring dependencies in temporal fused representations. The concatenated results were channeled into an LSTM to yield the ultimate fused representation. Initially, two note modalities were designated as primary modes, and subsequently, the proposed fusion model was compared with comparative models including recent models such as Crossformer. The proposed model consistently exhibited superior predictive prowess in both tasks. Removing the attention-backtracking module led to performance decline. The proposed model consistently shows excellent predictive capabilities in both tasks. The proposed method not only effectively integrates data from the four modalities, but also has a good understanding of how to handle irregular time series data and lengthy clinical texts. An effective method is provided, which is expected to be more widely used in multimodal medical data representation.
- 18 June 2023
- Published by Elsevier in Journal of Biomedical Informatics
- Vol. 143, 104427
Representation of time-varying and time-invariant EMR data and its application in modeling outcome prediction for heart failure patients
- 18 June 2023
- Published by Elsevier in Journal of Biomedical Informatics
- Vol. 143, 104427
Representation of time-varying and time-invariant EMR data and its application in modeling outcome prediction for heart failure patients