Multi-objective learning and explanation for stroke risk assessment in Shanxi province.

Stroke is the leading cause of death in China (Zhou et al. in The Lancet, 2019). A dataset from Shanxi Province is analyzed to predict the risk of patients at four states (low/medium/high/attack) and to estimate transition probabilities between various states via a SHAP DeepExplainer. To handle the issues related to an imbalanced sample set, the quadratic interactive deep model (QIDeep) was first proposed by flexible selection and appending of quadratic interactive features. The experimental results showed that the QIDeep model with 3 interactive features achieved the state-of-the-art accuracy 83.33%(95% CI (83.14%; 83.52%)). Blood pressure, physical inactivity, smoking, weight, and total cholesterol are the top five most important features. For the sake of high recall in the attack state, stroke occurrence prediction is considered an auxiliary objective in multi-objective learning. The prediction accuracy was improved, while the recall of the attack state was increased by 17.79% (to 82.06%) compared to QIDeep (from 71.49%) with the same features. The prediction model and analysis tool in this paper provided not only a prediction method but also an attribution explanation of the risk states and transition direction of each patient, a valuable tool for doctors to analyze and diagnose the disease.

Persistently Increased Resting Energy Expenditure Predicts Short-Term Mortality in Patients with Acute-on-Chronic Liver Failure.

OBJECTIVE: Hypermetabolism based on measurements of resting energy expenditure (REE) is suggested to be a potential biomarker for predicting the clinical outcomes of some diseases. We aimed to evaluate the potential value of hypermetabolism for predicting the short-term (28-day) mortality of patients with hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF). METHODS: A total of 105 HBV-ACLF patients, 30 chronic hepatitis B (CHB) patients and 30 healthy controls (HCs) were included in this study. The REE was measured using indirect calorimetry in the morning after 8-10 h of fasting. The predicted REE (REEHB) was determined using Harris-Benedict equation. Persistent hypermetabolism was defined as the REE:REEHB ratio > 1.20 at day 1 and day 7 after admission. The severity of liver disease was estimated using the Model for End-Stage Liver Disease (MELD). Clinical and biochemical variables were determined using blood samples ordered upon admission. These variables were compared between nonsurviving and surviving patients who were classified according to the 28-day mortality. RESULTS: The frequency of hypermetabolism at baseline was significantly higher in ACLF patients than that in HCs and CHB patients. Forty-six (43.8%) ACLF patients died within follow-up of 28 days. Persistent hypermetabolism (OR 2.10; 95% CI 1.15-3.69; p = 0.002) and MELD score (OR 1.93; 95% CI 1.47-3.51; p = 0.012) were independent predictive indicators of 28-day mortality. Furthermore, the performance of the 2 variables (persistent hypermetabolism and MELD) together with the area under the receiver operating curve (AUROC: 0.819) was significantly better than that of MELD alone -(AUROC: 0.694) for prediction of short-term mortality (p = 0.014). CONCLUSION: These findings indicate that persistent hypermetabolism is predictive of short-term mortality in this small population.