Factors influencing recurrence after an ischemic stroke vary by sex.

OBJECTIVES: To determine the factors that affect recurrent stroke after acute ischemic stroke, specifically between male and female groups. METHODS: We examined relative factors associated with recurrent stroke in Chinese patients with first-ever ischemic stroke. LASSO (least absolute shrinkage and selection operator) Cox regression were used to determine the predictors of recurrent stroke in the male and female groups. Next, We used Kaplan-Meier survival curves and interactions among these predictors to assess the association between relapse-related factors and recurrent stroke. RESULTS: During one year of follow-up, we documented 42 incidents of recurrent stroke in males and 15 in females. There was no significant difference in the overall recurrence rate between men and women. We finally identified three variables in males and one variable in females associated considerably with recurrent stroke by LASSO Cox regression. In females, good sleep appeared to be the most significant protective factor against recurrent stroke(hazard ratio [HR], 0.21; 95% CI, 0.08-0.57). In the male group, we found two risk factors: atherosclerotic burden (HR, 2.42; 95% CI, 1.30-4.51) and coronary heart disease (HR, 2.98; 95% CI, 1.16-7.66); and one protective factor: domestic/physical activities (HR, 0.45; 95% CI, 0.24-0.83). We also found an interaction between good sleep and domestic/physical activities in males (Pinteraction = 0.016). DISCUSSION: Our data indicate that the factors for recurrent stroke may differ by sex. Engaging in domestic/physical activities may substantially lower recurrent strokes in Chinese adult males. And good sleep in females appears to be more important in preventing stroke recurrence.

Development and validation of a practical machine learning model to predict sepsis after liver transplantation.

BACKGROUND: Postoperative sepsis is one of the main causes of mortality after liver transplantation (LT). Our study aimed to develop and validate a predictive model for postoperative sepsis within 7 d in LT recipients using machine learning (ML) technology. METHODS: Data of 786 patients received LT from January 2015 to January 2020 was retrospectively extracted from the big data platform of Third Affiliated Hospital of Sun Yat-sen University. Seven ML models were developed to predict postoperative sepsis. The area under the receiver-operating curve (AUC), sensitivity, specificity, accuracy, and f1-score were evaluated as the model performances. The model with the best performance was validated in an independent dataset involving 118 adult LT cases from February 2020 to April 2021. The postoperative sepsis-associated outcomes were also explored in the study. RESULTS: After excluding 109 patients according to the exclusion criteria, 677 patients underwent LT were finally included in the analysis. Among them, 216 (31.9%) were diagnosed with sepsis after LT, which were related to more perioperative complications, increased postoperative hospital stay and mortality after LT (all p < .05). Our results revealed that a larger volume of red blood cell infusion, ascitic removal, blood loss and gastric drainage, less volume of crystalloid infusion and urine, longer anesthesia time, higher level of preoperative TBIL were the top 8 important variables contributing to the prediction of post-LT sepsis. The Random Forest Classifier (RF) model showed the best overall performance to predict sepsis after LT among the seven ML models developed in the study, with an AUC of 0.731, an accuracy of 71.6%, the sensitivity of 62.1%, and specificity of 76.1% in the internal validation set, and a comparable AUC of 0.755 in the external validation set. CONCLUSIONS: Our study enrolled eight pre- and intra-operative variables to develop an RF-based predictive model of post-LT sepsis to assist clinical decision-making procedure.

Postoperative sepsis is one of the main causes of mortality after liver transplantation (LT).Our results revealed that a larger volume of red blood cell infusion, ascitic removal, blood loss and gastric drainage, less volume of crystalloid infusion and urine, longer anesthesia time, higher level of preoperative TBIL were the top 8 important variables contributing to the prediction of post-LT sepsis.The Random Forest Classifier (RF) model showed the best overall performance to predict sepsis after LT in our study, which could assist in the clinical decision-making procedure.

Derivation and Validation of a Prognostic Scoring Model Based on Clinical and Pathological Features for Risk Stratification in Oral Squamous Cell Carcinoma Patients: A Retrospective Multicenter Study.

OBJECTIVE: To develop and validate a simple-to-use prognostic scoring model based on clinical and pathological features which can predict overall survival (OS) of patients with oral squamous cell carcinoma (OSCC) and facilitate personalized treatment planning. MATERIALS AND METHODS: OSCC patients (n = 404) from a public hospital were divided into a training cohort (n = 282) and an internal validation cohort (n = 122). A total of 12 clinical and pathological features were included in Kaplan-Meier analysis to identify the factors associated with OS. Multivariable Cox proportional hazards regression analysis was performed to further identify important variables and establish prognostic models. Nomogram was generated to predict the individual's 1-, 3- and 5-year OS rates. The performance of the prognostic scoring model was compared with that of the pathological one and the AJCC TNM staging system by the receiver operating characteristic curve (ROC), concordance index (C-index), calibration curve, and decision curve analysis (DCA). Patients were classified into high- and low-risk groups according to the risk scores of the nomogram. The nomogram-illustrated model was independently tested in an external validation cohort of 95 patients. RESULTS: Four significant variables (physical examination-tumor size, imaging examination-tumor size, pathological nodal involvement stage, and histologic grade) were included into the nomogram-illustrated model (clinical-pathological model). The area under the ROC curve (AUC) of the clinical-pathological model was 0.687, 0.719, and 0.722 for 1-, 3- and 5-year survival, respectively, which was superior to that of the pathological model (AUC = 0.649, 0.707, 0.717, respectively) and AJCC TNM staging system (AUC = 0.628, 0.668, 0.677, respectively). The clinical-pathological model exhibited improved discriminative power compared with pathological model and AJCC TNM staging system (C-index = 0.755, 0.702, 0.642, respectively) in the external validation cohort. The calibration curves and DCA also displayed excellent predictive performances. CONCLUSION: This clinical and pathological feature based prognostic scoring model showed better predictive ability compared with the pathological one, which would be a useful tool of personalized accurate risk stratification and precision therapy planning for OSCC patients.