This is the authors' abstract. We don't add key points for this paper.
Background: There is still not a mortality prediction model built for breast cancer admitted to intensive care unit (ICU). Objectives: We aimed to build a prognostic model with comprehensive data achieved from eICU database. Methods: Outcome was defined as all-cause in-hospital mortality. Least absolute shrinkage and selection operator (LASSO) was conducted to select important variables which were then taken into logistic regression to build the model. Bootstrap method was then conducted for internal validation. Results: 448 patients were included in this study and 79 (17.6%) died in hospital. Only 5 items were included in the model and the area under the curve (AUC) was 0.844 (95% confidence interval [CI]: 0.804-0.884). Calibration curve and Brier score (0.111, 95% CI: 0.090-0.127) showed good calibration of the model. After internal validation, corrected AUC and Brier score were 0.834 and 0.116. Decision curve analysis (DCA) also showed effective clinical use of the model. The model can be easily assessed on website of https://breastcancer123.shinyapps.io/BreastCancerICU/. Conclusions: The model derived in this study can provide an accurate prognosis for breast cancer admitted to ICU easily, which can help better clinical management.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2022) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: