Fengchan Xi

Dr. Fengchan Xi


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
31 March 2023 in JMIR Formative Research
Verified
Journal Article
Open access
JLJiang Li
Jiang Li
FXFengchan Xi
Fengchan Xi
WYWenkui Yu
Wenkui Yu
CSChuanrui Sun
Chuanrui Sun
XWXiling Wang
Xiling Wang
Background: Sepsis is a leading cause of death in patients with trauma, and the risk of mortality increases significantly for each hour of delay in treatment. A hypermetabolic baseline and explosive inflammatory immune response mask clinical signs and symptoms of sepsis in trauma patients, making early diagnosis of sepsis more challenging. Machine learning–based predictive modeling has shown great promise in evaluating and predicting sepsis risk in the general intensive care unit (ICU) setting, but there has been no sepsis prediction model specifically developed for trauma patients so far. Objective: To develop a machine learning model to predict the risk of sepsis at an hourly scale among ICU-admitted trauma patients. Methods: We extracted data from adult trauma patients admitted to the ICU at Beth Israel Deaconess Medical Center between 2008 and 2019. A total of 42 raw variables were collected, including demographics, vital signs, arterial blood gas, and laboratory tests. We further derived a total of 485 features, including measurement pattern features, scoring features, and time-series variables, from the raw variables by feature engineering. The data set was randomly split into 70% for model development with stratified 5-fold cross-validation, 15% for calibration, and 15% for testing. An Extreme Gradient Boosting (XGBoost) model was developed to predict the hourly risk of sepsis at prediction windows of 4, 6, 8, 12, and 24 hours. We evaluated model performance for discrimination and calibration both at time-step and outcome levels. Clinical applicability of the model was evaluated with varying levels of precision, and the potential clinical net benefit was assessed with decision curve analysis (DCA). A Shapley additive explanation algorithm was applied to show the effect of features on the prediction model. In addition, we trained an L2-regularized logistic regression model to compare its performance with XGBoost. Results: We included 4603 trauma patients in the study, 1196 (26%) of whom developed sepsis. The XGBoost model achieved an area under the receiver operating characteristics curve (AUROC) ranging from 0.83 to 0.88 at the 4-to-24-hour prediction window in the test set. With a ratio of 9 false alerts for every true alert, it predicted 73% (386/529) of sepsis-positive timesteps and 91% (163/179) of sepsis events in the subsequent 6 hours. The DCA showed our model had a positive net benefit in the threshold probability range of 0 to 0.6. In comparison, the logistic regression model achieved lower performance, with AUROC ranging from 0.76 to 0.84 at the 4-to-24-hour prediction window. Conclusions: The machine learning–based model had good discrimination and calibration performance for sepsis prediction in critical trauma patients. Using the model in clinical practice might help to identify patients at risk of sepsis in a time window that enables personalized intervention and early treatment.

14 February 2023 in Journal of Clinical Medicine
Verified
Journal Article
Open access
PDF file
JLJiang Li
Jiang Li
FXFengchan Xi
Fengchan Xi
YHYuanchen He
Yuanchen He
CSChuanrui Sun
Chuanrui Sun
WYWenkui Yu
Wenkui Yu
XWXiling Wang
Xiling Wang
Overweight and moderate obesity confer a survival benefit in chronic diseases such as coronary artery disease and chronic kidney disease, which has been termed the “obesity paradox”. However, whether this phenomenon exists in trauma patients remains controversial. We performed a retrospective cohort study in abdominal trauma patients admitted to a Level I trauma center in Nanjing, China between 2010 and 2020. In addition to the traditional body mass index (BMI) based measures, we further examined the association between body composition-based indices with clinical severity in trauma populations. Body composition indices including skeletal muscle index (SMI), fat tissue index (FTI), and total fat-to-muscle ratio (FTI/SMI) were measured using computed tomography. Our study found that overweight was associated with a four-fold risk of mortality (OR, 4.47 [95% CI, 1.40–14.97], p = 0.012) and obesity was associated with a seven-fold risk of mortality (OR, 6.56 [95% CI, 1.07–36.57], p = 0.032) compared to normal weight. Patients with high FTI/SMI had a three-fold risk of mortality (OR, 3.06 [95% CI, 1.08–10.16], p = 0.046) and double the risk of an intensive care unit length of stay ≥ 5 d (OR, 1.75 [95% CI, 1.06–2.91], p = 0.031) compared to patients with low FTI/SMI. The obesity paradox was not observed in abdominal trauma patients, and high FTI/SMI ratio was independently associated with increased clinical severity.
1

30 January 2023 in Journal of Clinical Medicine
Verified
Journal Article
Open access
PDF file
CSChuanrui Sun
Chuanrui Sun
FXFengchan Xi
Fengchan Xi
JLJiang Li
Jiang Li
WYWenkui Yu
Wenkui Yu
XWXiling Wang
Xiling Wang
This study aimed to identify the long-term D-dimer trajectory patterns and their associations with in-hospital all-cause mortality in abdominal trauma patients. This is a retrospective cohort study of general adult abdominal trauma patients admitted to Jinling Hospital (Nanjing, China) between January 2010 and April 2020. Group-based trajectory modeling was applied to model D-dimer trajectories over the first 50 days post-trauma. A multivariable logistic regression was performed to estimate the associations between D-dimer trajectories and in-hospital all-cause mortality. A total of 309 patients were included. We identified four distinct D-dimer trajectories: group 1 (57.61%; “stable low”), group 2 (28.16%; “moderate-decline”), group 3 (8.41%; “high-rapid decline”), and group 4 (5.83%; “high-gradual decline”). The SOFA score (p = 0.005) and ISS (p = 0.001) were statistically higher in groups 3 and 4 than in groups 1 and 2. The LMWH and UFH did not differ between groups 3 and 4. Compared with the patients in group 1, only the patients in group 4 were at a higher risk of in-hospital all-cause mortality (OR = 6.94, 95% CI: 1.20–40.25). The long-term D-dimer trajectories post-trauma were heterogeneous and associated with mortality. An initially high and slowly-resolved D-dimer might function as the marker of disease deterioration, and specific interventions are needed.

13 April 2026 in International Journal for Vitamin and Nutrition Research
Verified
Journal Article
Open access
PDF file
FXFengchan Xi
Fengchan Xi
NZNan Zheng
Nan Zheng
DWDi Wang
Di Wang
RTRan Teng
Ran Teng
XWXiling Wang
Xiling Wang
STShanjun Tan
Shanjun Tan
Background: Currently, vitamin D deficiency is highly prevalent among women of childbearing age. However, the relationship between vitamin D levels and muscle mass reduction in women of childbearing age remains unclear. This study aimed to investigate the association between vitamin D levels and muscle mass reduction in women of childbearing age. Methods: This study included 3934 participants from the National Health and Nutrition Examination Survey (NHANES) 2011–2018. Weighted linear regression and weighted chi-square tests were used to compare baseline characteristics between groups for continuous and categorical variables, respectively. Weighted logistic regression analysis was employed to assess the association between vitamin D levels and muscle mass reduction. Results: The study population was divided into two groups, 2803 non-muscle mass reduction and 1131 muscle mass reduction participants. The quartile of appendicular skeletal muscle mass (ASM)/body mass index (BMI) for diagnosing muscle mass reduction was determined to be 0.579. Compared with the non-muscle mass reduction group, muscle mass reduction women of childbearing age had a higher prevalence of vitamin D deficiency (p < 0.001) and lower use of vitamin D supplements (p = 0.018). Serum vitamin D levels in the muscle mass reduction group were significantly lower than those in the non-muscle mass reduction group (p < 0.001). After adjusting for confounding variables, weighted logistic regression analysis revealed that insufficient (odds ratio (OR) = 1.34, 95% confidence interval (CI): 1.01–1.78, p = 0.042) and deficient (OR = 1.54, 95% CI: 1.16–2.05, p = 0.003) vitamin D levels were still associated with muscle mass reduction. Conclusions: Insufficient and deficient vitamin D levels are associated with muscle mass reduction in women of childbearing age, highlighting the potential role of vitamin D in muscle mass reduction management. Vitamin D supplementation may improve muscle mass reduction in vitamin D-deficient women of childbearing age, and its effectiveness needs to be further verified by randomized controlled trials.

29 September 2025 in Journal of Medical Case Reports
Verified
Journal Article
Open access
PDF file
RTRan Teng
Ran Teng
NZNan Zheng
Nan Zheng
RYRuixuan Yu
Ruixuan Yu
FXFengchan Xi
Fengchan Xi
Hemolysis, elevated liver enzymes, and low platelet count syndrome and catastrophic antiphospholipid syndrome are severe complications of pregnancy with potentially life-threatening consequences. Currently, the timely recognition and management of pregnant women with hemolysis, elevated liver enzymes, and low platelet count syndrome and catastrophic antiphospholipid syndrome remains a challenge. In this case, we describe a 28-year-old East Asian pregnant woman with a history of antiphospholipid syndrome who developed hemolysis, elevated liver enzymes, and low platelet count syndrome and catastrophic antiphospholipid syndrome after delivery at 36 weeks of pregnancy. The patient presented with multiple red spots on her extremities and despite receiving conventional therapy, her condition deteriorated rapidly postpartum leading to multiple organ dysfunction syndrome. Notably, the onset of multiple organ dysfunction syndrome occurred within a narrow timeframe of 11 hours. This case underscores the importance of the early warning value of skin manifestations. At the same time, it is crucial to closely monitor pregnant women with antiphospholipid syndrome both during and after delivery. Furthermore, the rapid onset and progression of multiple organ dysfunction syndrome highlight the critical need for prompt initiation of organ support measures in such situations.

5 September 2022 in JMIR Preprints
Verified
Preprint
Open access
JLJiang Li
Jiang Li
FXFengchan Xi
Fengchan Xi
WYWenkui Yu
Wenkui Yu
CSChuanrui Sun
Chuanrui Sun
XWXiling Wang
Xiling Wang
BACKGROUND Sepsis is a leading cause of death in patients with trauma, and the risk of mortality increases significantly for each hour of delay in treatment. A hypermetabolic baseline and explosive inflammatory immune response mask clinical signs and symptoms of sepsis in trauma patients, making early diagnosis of sepsis more challenging. Machine learning–based predictive modeling has shown great promise in evaluating and predicting sepsis risk in the general intensive care unit (ICU) setting, but there has been no sepsis prediction model specifically developed for trauma patients so far. OBJECTIVE To develop a machine learning model to predict the risk of sepsis at an hourly scale among ICU-admitted trauma patients. METHODS We extracted data from adult trauma patients admitted to the ICU at Beth Israel Deaconess Medical Center between 2008 and 2019. A total of 42 raw variables were collected, including demographics, vital signs, arterial blood gas, and laboratory tests. We further derived a total of 485 features, including measurement pattern features, scoring features, and time-series variables, from the raw variables by feature engineering. The data set was randomly split into 70% for model development with stratified 5-fold cross-validation, 15% for calibration, and 15% for testing. An Extreme Gradient Boosting (XGBoost) model was developed to predict the hourly risk of sepsis at prediction windows of 4, 6, 8, 12, and 24 hours. We evaluated model performance for discrimination and calibration both at time-step and outcome levels. Clinical applicability of the model was evaluated with varying levels of precision, and the potential clinical net benefit was assessed with decision curve analysis (DCA). A Shapley additive explanation algorithm was applied to show the effect of features on the prediction model. In addition, we trained an L2-regularized logistic regression model to compare its performance with XGBoost. RESULTS We included 4603 trauma patients in the study, 1196 (26%) of whom developed sepsis. The XGBoost model achieved an area under the receiver operating characteristics curve (AUROC) ranging from 0.83 to 0.88 at the 4-to-24-hour prediction window in the test set. With a ratio of 9 false alerts for every true alert, it predicted 73% (386/529) of sepsis-positive timesteps and 91% (163/179) of sepsis events in the subsequent 6 hours. The DCA showed our model had a positive net benefit in the threshold probability range of 0 to 0.6. In comparison, the logistic regression model achieved lower performance, with AUROC ranging from 0.76 to 0.84 at the 4-to-24-hour prediction window. CONCLUSIONS The machine learning–based model had good discrimination and calibration performance for sepsis prediction in critical trauma patients. Using the model in clinical practice might help to identify patients at risk of sepsis in a time window that enables personalized intervention and early treatment.