Development and Validation of an Interpretable Machine Learning Model for Predicting Hospital Mortality in ICU-Admitted Ovarian Cancer Patients: A Multicenter Study

Yihan Li, Miao Guo, Hefan Yang, Huajie Chi, Ping Chen,
Development and Validation of an Interpretable Machine Learning Model for Predicting Hospital Mortality in ICU-Admitted Ovarian Cancer Patients: A Multicenter Study,
International Journal of Gynecological Cancer,
2025,
102689,
ISSN 1048-891X,
https://doi.org/10.1016/j.ijgc.2025.102689.
(https://www.sciencedirect.com/science/article/pii/S1048891X25018092)
Abstract: Objective
To develop and validate an interpretable machine learning model for predicting hospital mortality in ovarian cancer patients admitted to intensive care units.
Methods
This retrospective multicenter study analyzed 433 ovarian cancer patients from MIMIC-IV (n=286) and eICU (n=147) databases who met inclusion criteria of first ICU admission, age >18 years, and ICU stay ≥24 hours. The Boruta algorithm identified important predictors from clinical and laboratory data collected within 24 hours of admission. Seven machine learning algorithms were evaluated using 10-fold cross-validation. MIMIC-IV served as the training dataset (70% training, 30% internal validation) with eICU providing external validation. Model interpretability was assessed using SHAP analysis to identify feature contributions and clinical thresholds.
Results
The final cohort had 9.5% hospital mortality (41/433). The Support Vector Machine model achieved superior performance with AUC 0.857 (95% CI: 0.737-0.976) in internal validation and 0.750 (95% CI: 0.642-0.858) in external validation, outperforming SOFA (0.578), SAPS II (0.674), and OASIS (0.671) scores. Red cell distribution width emerged as the most important predictor (mean importance 0.032), followed by bicarbonate (0.028) and chloride (0.020). SHAP analysis revealed critical thresholds: RDW >20% (particularly >25%), bicarbonate <20 mmol/L, and abnormal chloride levels significantly increased mortality risk. High-risk patients demonstrated RDW 30.8% (SHAP +0.675) and bicarbonate 12 mmol/L (SHAP +0.124).
Conclusion
The interpretable machine learning model accurately predicts hospital mortality in ICU-admitted ovarian cancer patients using readily available laboratory parameters, significantly outperforming traditional ICU scores and providing actionable clinical thresholds for risk stratification.
Keywords: Ovarian cancer; Machine learning; Intensive care unit; Hospital mortality; SHAP analysis