Xiu Zhang

Xiu Zhang

Auto-generated profile, based on publicly available information

Xiu Zhang

Department of Cardiology, West China Hospital, Sichuan University, Chengdu, China

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
2 May 2024 in Journal of Medical Internet Research
Verified
Journal Article
Open access
ZGZhenyue Gao
Zhenyue Gao
XLXiaoli Liu
Xiaoli Liu
YKYu Kang
Yu Kang
PHPan Hu
Pan Hu
XZXiu Zhang
Xiu Zhang
ZZZhengbo Zhang
Zhengbo Zhang
Background: Clinical notes contain contextualized information beyond structured data related to patients’ past and current health status. Objective: This study aimed to design a multimodal deep learning approach to improve the evaluation precision of hospital outcomes for heart failure (HF) using admission clinical notes and easily collected tabular data. Methods: Data for the development and validation of the multimodal model were retrospectively derived from 3 open-access US databases, including the Medical Information Mart for Intensive Care III v1.4 (MIMIC-III) and MIMIC-IV v1.0, collected from a teaching hospital from 2001 to 2019, and the eICU Collaborative Research Database v1.2, collected from 208 hospitals from 2014 to 2015. The study cohorts consisted of all patients with critical HF. The clinical notes, including chief complaint, history of present illness, physical examination, medical history, and admission medication, as well as clinical variables recorded in electronic health records, were analyzed. We developed a deep learning mortality prediction model for in-hospital patients, which underwent complete internal, prospective, and external evaluation. The Integrated Gradients and SHapley Additive exPlanations (SHAP) methods were used to analyze the importance of risk factors. Results: The study included 9989 (16.4%) patients in the development set, 2497 (14.1%) patients in the internal validation set, 1896 (18.3%) in the prospective validation set, and 7432 (15%) patients in the external validation set. The area under the receiver operating characteristic curve of the models was 0.838 (95% CI 0.827-0.851), 0.849 (95% CI 0.841-0.856), and 0.767 (95% CI 0.762-0.772), for the internal, prospective, and external validation sets, respectively. The area under the receiver operating characteristic curve of the multimodal model outperformed that of the unimodal models in all test sets, and tabular data contributed to higher discrimination. The medical history and physical examination were more useful than other factors in early assessments. Conclusions: The multimodal deep learning model for combining admission notes and clinical tabular data showed promising efficacy as a potentially novel method in evaluating the risk of mortality in patients with HF, providing more accurate and timely decision support.

9 September 2022 in ESC Heart Failure
Verified
Journal Article
Open access
XZXiu Zhang
Xiu Zhang
ZLZeruxin Luo
Zeruxin Luo
MYMengxuan Yang
Mengxuan Yang
WHWei Huang
Wei Huang
PYPengming Yu
Pengming Yu
During the coronavirus disease 2019 (COVID-19) pandemic, it has become difficult to provide centre-based cardiac rehabilitation for heart failure patients. Digital therapeutics is a novel concept proposed in recent years that refers to the use of evidence-based therapeutic interventions driven by high-quality software programs to treat, manage, or prevent a medical condition. However, little is known about the use of this technology in heart failure patients. This study aims to explore the safety and efficacy of digital therapeutics-based cardiac rehabilitation in heart failure patients and to provide new insights into a new cardiac rehabilitation model during the COVID-19 era. To identify technologies related to digital therapeutics, such as the use of medical applications, wearable devices, and the Internet, all relevant studies published on PubMed, EMBASE, Cochrane database, and China National Knowledge Internet were searched from the time the database was established until October 2021. The PEDro was used to assess the quality of included studies. We ultimately identified five studies, which included 1119 patients. The mean age was 66.37, the mean BMI was 25.9, and the NYHA classification ranged from I to III (I = 232, II = 157, III = 209). The mean 6-min walk distance was 397.7 m. The PEDro scores included in the study ranged from 4 to 8, with a mean of 5.8. Exercise training was performed in four studies, and psychological interventions were conducted in three studies. No death or serious adverse events were observed. Adherence was reported in three studies, and all exceeded 85%. The results of most studies showed that digital therapeutics-based cardiac rehabilitation significantly increases exercise capacity and quality of life in heart failure patients. Overall, although this study suggests that digital therapeutics-based cardiac rehabilitation may be a viable intervention for heart failure patients during the COVID-19 era, the efficacy of this new model in routine clinical practice needs to be further validated in a large clinical trial.

7 November 2023 in JMIR Preprints
Verified
Preprint
Open access
ZGZhenyue Gao
Zhenyue Gao
XLXiaoli Liu
Xiaoli Liu
YKYu Kang
Yu Kang
PHPan Hu
Pan Hu
XZXiu Zhang
Xiu Zhang
ZZZhengbo Zhang
Zhengbo Zhang
BACKGROUND Clinical notes contain contextualized information beyond structured data related to patients’ past and current health status. OBJECTIVE This study aimed to design a multimodal deep learning approach to improve the evaluation precision of hospital outcomes for heart failure (HF) using admission clinical notes and easily collected tabular data. METHODS Data for the development and validation of the multimodal model were retrospectively derived from 3 open-access US databases, including the Medical Information Mart for Intensive Care III v1.4 (MIMIC-III) and MIMIC-IV v1.0, collected from a teaching hospital from 2001 to 2019, and the eICU Collaborative Research Database v1.2, collected from 208 hospitals from 2014 to 2015. The study cohorts consisted of all patients with critical HF. The clinical notes, including chief complaint, history of present illness, physical examination, medical history, and admission medication, as well as clinical variables recorded in electronic health records, were analyzed. We developed a deep learning mortality prediction model for in-hospital patients, which underwent complete internal, prospective, and external evaluation. The Integrated Gradients and SHapley Additive exPlanations (SHAP) methods were used to analyze the importance of risk factors. RESULTS The study included 9989 (16.4%) patients in the development set, 2497 (14.1%) patients in the internal validation set, 1896 (18.3%) in the prospective validation set, and 7432 (15%) patients in the external validation set. The area under the receiver operating characteristic curve of the models was 0.838 (95% CI 0.827-0.851), 0.849 (95% CI 0.841-0.856), and 0.767 (95% CI 0.762-0.772), for the internal, prospective, and external validation sets, respectively. The area under the receiver operating characteristic curve of the multimodal model outperformed that of the unimodal models in all test sets, and tabular data contributed to higher discrimination. The medical history and physical examination were more useful than other factors in early assessments. CONCLUSIONS The multimodal deep learning model for combining admission notes and clinical tabular data showed promising efficacy as a potentially novel method in evaluating the risk of mortality in patients with HF, providing more accurate and timely decision support.

27 August 2025 in Journal of Medical Internet Research
Verified
Journal Article
Open access
MLMengwei Li
Mengwei Li
YKYu Kang
Yu Kang
SZShuanglin Zhao
Shuanglin Zhao
XZXiu Zhang
Xiu Zhang
ZLZheng Liu
Zheng Liu
ZZZhengbo Zhang
Zhengbo Zhang
Background: Excessive pulmonary congestion (PC) is a major contributor to heart failure (HF) deterioration and often necessitates emergency hospitalization. Early detection of PC-related respiratory abnormalities via wearable devices could enable prompt intervention and reduce admissions. However, the feasibility of using wearables to detect specific respiratory biomarkers of PC remains unclear. Objective: This study aimed to evaluate the feasibility of using wearables to monitor respiratory data in hospitalized patients with HF and to determine if these signals could distinguish patients with HF and PC. Methods: This single-center, observational, exploratory study enrolled hospitalized patients with HF without severe lung diseases or need for intensive care or ventilation. All participants wore a designated device for ≥24 hours starting within 24 hours after admission; only nighttime data were analyzed. Breathing patterns were quantified across the respiratory cycle, amplitude, and multiscale entropy (MSE). Patients underwent comprehensive evaluations, including vital signs, laboratories, echocardiography, and a 28-zone lung ultrasound to define PC (>5 B-lines) within the first 24 hours. Results: The study enrolled 62 patients with HF between May 2021 and November 2022, including 44 with PC. Compared with patients with non-PC, those with PC demonstrated significantly prolonged mean expiratory time (TE; mean 2.17, SD 0.43 s vs mean 1.94, SD 0.34 s, P=.03), elevated expiratory time ratio (TE_ratio; mean 59.12%, SD 2.94% vs mean 56.4%, SD 3.36%, P=.006), and higher area under the curves (AUC) of MSE values for scales 1 to 5 (area_1_5) and scales 6 to 20 (area_6_20) in respiratory amplitude (RA; area_1_5 mean 4.20, SD 1.52 vs area_1_5 mean 2.93, SD 1.03, P<.001; and area_6_20 mean 8.86, SD 3.14 vs area_6_20 mean 12.28, SD 4.84, P=.002). Logistic regression identified mean TE_ratio, RA area_1_5, and RA area_6_20 as significant predictors of PC (P<.05). After adjusting for clinical confounders, both RA area_1_5 and RA area_6_20 remained independently associated with PC. Receiver operating characteristic analysis revealed that RA area_1_5 had the largest AUC of 0.75 (95% CI 0.63‐0.88, P=.002), with 65.9% sensitivity and 73.3% specificity. For the multivariate logistic regression model constructed using combined parameters of diastolic blood pressure, logarithmically transformed N-terminal pro-B-type natriuretic peptide, New York Heart Association class IV, mean TE_ratio, RA area_1_5, and RA area_6_20, the AUC was 0.91 (95% CI 0.84-0.98, P<.001), sensitivity of 77.3%, and specificity of 88.9%. Conclusions: In this exploratory study, wearable-based MSE analysis distinguished hospitalized patients with HF with and those without PC, showing prolonged expiratory phases and increased respiratory amplitude complexity in the PC group.