Search

Backgrounds: The in-hospital mortality in lung cancer patients admitted to intensive care unit (ICU) is extremely high. This study intended to adopt machine learning algorithm models to predict in-hospital mortality of critically ill lung cancer for providing relative information in clinical decision-making. Methods: Data were extracted from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) for a training cohort and data extracted from the Medical Information Mart for eICU Collaborative Research Database (eICU-CRD) database for a validation cohort. Logistic regression, random forest, decision tree, light gradient boosting machine (LightGBM), eXtreme gradient boosting (XGBoost), and an ensemble (random forest+LightGBM+XGBoost) model were used for prediction of in-hospital mortality and important feature extraction. The AUC (area under receiver operating curve), accuracy, F1 score and recall were used to evaluate the predictive performance of each model. Shapley Additive exPlanations (SHAP) values were calculated to evaluate feature importance of each feature. Results: Overall, there were 653 (24.8%) in-hospital mortality in the training cohort, and 523 (21.7%) in-hospital mortality in the validation cohort. Among the six machine learning models, the ensemble model achieved the best performance. The top 5 most influential features were the sequential organ failure assessment (SOFA) score, albumin, the oxford acute severity of illness score (OASIS) score, anion gap and bilirubin in random forest and XGBoost model. The SHAP summary plot was used to illustrate the positive or negative effects of the top 15 features attributed to the XGBoost model. Conclusion: The ensemble model performed best and might be applied to forecast in-hospital mortality of critically ill lung cancer patients, and the SOFA score was the most important feature in all models. These results might offer valuable and significant reference for ICU clinicians' decision-making in advance. [ABSTRACT FROM AUTHOR]

Background: We aimed to evaluate the prognostic ability of blood urea nitrogen (BUN) to serum albumin ratio (BAR) to predict in-hospital mortality in patients with lung cancer in the intensive care unit (ICU). Methods: Medical Information Mart for Intensive Care IV (MIMIC-IV v1.0) database was used to identify patients who were diagnosed with lung cancer. The primary outcome was in-hospital mortality. Multivariate COX regression was used to investigate the association between BAR and in-hospital mortality and propensity score matching (PSM) and inverse probability of treatment weighting (IPTW) were also used to ensure the robustness of our findings. eICU-CRD database (validation cohort) was also applied to validate our findings. Results: The optimal cut-off value for BAR was 6.8mg/g. Among 1202 patients who were diagnosed with lung cancer, 287 high-BAR group (≥ 6.8mg/g) patients and 287 low-BAR group (< 6.8mg/g) patients, who had similar propensity scores were included in this study. After matching, the high-BAR group had significantly higher in-hospital mortality (hazard ratio, HR, 2.24, 95% confidence index, 95% CI, 1.57– 3.19, P< 0.001) even after adjustment for confounding factors. Moreover, the performance of BAR was superior to that of BUN and serum albumin alone and could add net benefit in predicting in-hospital mortality. Those results were further confirmed in the validation cohort. Conclusion: As an easily accessible and cost-effective parameter, BAR could serve as a good prognostic predictor for lung cancer patients in ICU. [ABSTRACT FROM AUTHOR]