Related Experiment Video
Updated: May 7, 2025
13:25
Acute Kidney Injury Model Induced by Cisplatin in Adult Zebrafish
Published on: May 15, 2021
5.6K
Prediction of Cisplatin-Induced Acute Kidney Injury Using an Interpretable Machine Learning Model and Electronic
Kaori Ambe1, Yuka Aoki1, Miho Murashima2
1Department of Regulatory Science, Nagoya City University Graduate School of Pharmaceutical Sciences, Nagoya, Japan.
Clinical and Translational Science
|January 6, 2025
Summary
Predicting cisplatin-induced acute kidney injury (Cis-AKI) is crucial. A machine learning model using clinical data achieved 0.78 AUC, identifying magnesium as potentially protective and diuretics as risk factors for Cis-AKI.
Area of Science:
- Nephrology
- Oncology
- Medical Informatics
Background:
- Cisplatin chemotherapy is associated with a significant risk of acute kidney injury (AKI).
- Early prediction of cisplatin-induced AKI (Cis-AKI) is essential for patient management and potentially mitigating renal damage.
- Interpretable machine learning offers a promising avenue for developing predictive models using complex clinical data.
Purpose of the Study:
- To develop and validate an interpretable machine learning model for predicting Cis-AKI in hospitalized patients.
- To identify key clinical variables and patient factors associated with the development of Cis-AKI.
- To explore potential preventative strategies based on model insights.
Main Methods:
- A retrospective study of 1253 adult patients receiving cisplatin chemotherapy at Nagoya City University Hospital (2011-2020).
- Development of a CatBoost classification model using 29 explanatory variables, including demographics, lab values, medications, and treatment details.
- Cis-AKI diagnosis based on Kidney Disease Improving Global Outcomes serum creatinine criteria within 14 days of cisplatin administration.
Main Results:
- 119 patients (9.5%) developed Cis-AKI, with a median onset of 7 days.
- The predictive model achieved an Area Under the Receiver Operating Characteristic Curve (ROC-AUC) of 0.78.
- SHapley Additive exPlanations revealed that concomitant use of intravenous magnesium preparations was negatively correlated with Cis-AKI, while loop diuretics were positively correlated.
Conclusions:
- The developed interpretable machine learning model demonstrates good performance in predicting Cis-AKI.
- Intravenous magnesium preparations may play a protective role in preventing Cis-AKI.
- Further investigation into the role of magnesium and diuretics is warranted for optimizing Cis-AKI prevention strategies.
