Back
For best output, select "Paper Size" as "A4" and "Margin" as "0" or "None".
To save or print to PDF, please select Print Destination > Save as PDF, enable Background Graphics under "More Settings", then click "Save".
During the congress, E-Posters will be accessible to all participants on the congress website 24/7, as well as in the E-poster stations in the congress center.
Preparing your E-Poster
Please review the E-Poster format requirements carefully when preparing your E-Poster. Should your E-Poster not meet the mentioned requirements, it may not be displayed as described above.
E-Poster Submission Deadline
Please prepare and upload your E-Poster no later than March 14, 2026 11.59PM CET. After this date, you will no longer be able to prepare and upload your E-poster and it will not be displayed and accessible on the congress website.
Please follow the instructions below to input your abstract title.
Abstract titles should be brief and reflect the content of the abstract.
Chronic kidney disease (CKD) prevention is a critical public health challenge. Lifestyle-associated diseases are increasingly contributing to CKD onset. However, no existing tool specifically addresses lifestyle-associated risk factors to prevent CKD in young to middle-aged populations. Therefore, we developed an artificial intelligence (AI)-driven predictive system to identify CKD development in an apparently healthy population.
We analyzed annual health checkup data collected between 2017 and 2022 from Japanese adults aged 18–65 years. Individuals with an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2 or proteinuria ≥1+ at baseline (2017) were excluded. CKD development was defined as eGFR <60 mL/min/1.73 m2 or proteinuria in 2022. To construct a 5-year CKD risk predictive model, we applied supervised machine learning using blood test results, urine analyses, and self-administered questionnaires. Four algorithms—logistic regression, support vector machine (SVM), random forest, and XGBoost—were trained on data from 2017 to 2020. Model performance was assessed using receiver operating characteristic (ROC) curves to identify the most appropriate approach. To verify the versatility, the predictive models were further tested on additional datasets from annual health checkups of healthy populations.
Of 24,558 participants, 9,273 (93% male) met the eligibility criteria. The mean age was 36.3 years, and the mean eGFR at baseline was 84.2 mL/min/1.73 m2. During follow-up, 1,041 (11.2%) participants developed CKD. Seventy percent of the dataset was used for training, and 30% for testing. Training and testing accuracies were as follows: logistic regression, 89% and 88%; SVM, 89% and 89%; random forest, 99% and 100%; and XGBoost, 89% and 89%. The most important features of CKD development were baseline eGFR, dyslipidemia, and mean blood pressure in both the random forest and XGBoost models. ROC analysis showed that the area under the curve (AUC, 95% CI) for random forest was 1.00 (1.00–1.00), significantly higher than logistic regression [0.649 (0.619 – 0.683)], SVM [0.623 (0.605-0.641)], and XGBoost [0.635 (0.617 – 0.653)]. External validation using additional datasets (N = 7974) demonstrated accuracies of 92%–93% across all four predictive models.
All four predictive models demonstrated the ability to estimate CKD development over 5 years in an apparently healthy population, with random forest providing the highest accuracy. The eGFR at baseline, dyslipidemia, and mean blood pressure were identified as key predictors of CKD risk.
