Abstract: PO0964
New Diagnostic Model for the Differentiation of Diabetic Nephropathy from Non-Diabetic Nephropathy in Chinese Patients
Session Information
- Diabetic Kidney Disease: Clinical - 1
October 22, 2020 | Location: On-Demand
Abstract Time: 10:00 AM - 12:00 PM
Category: Diabetic Kidney Disease
- 602 Diabetic Kidney Disease: Clinical
Authors
- Liu, Xiaomin, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Zhang, Weiguang, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Dong, Zheyi, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Wang, Qian, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Chen, Yizhi, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Wang, Yong, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Chen, Liangmei, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Cai, Guangyan, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
- Chen, Xiangmei, Department of Nephrology Chinese PLA General Hospital, Chinese PLA Institute of Nephrology, State Key Laboratory of Kidney Diseases, National Clinical Research Center for Kidney Diseases, Beijing, Beijing, Beijing, China
Background
The differential diagnostic criteria of non-diabetic nephropathy (NDRD) and diabetic nephropathy (DN) usually depend on the 2007 KDOQI guideline, which is not accurate enough. Renal pathological biopsy is the gold standard for diagnosis, which is an invasive method and may cause several complications. This study aimed to construct a new noninvasive evaluation method for the differentiation of DN and NDRD.
Methods
We retrospectively screened 1030 patients (January 2005-March 2017). Variables were ranked in terms of importance, and random forest (RF) and support vector machine(SVM) were then used to construct the models. The final model was validated using an external group (338 patients, April 2017-April 2019), and compared with previous models.
Results
A total of 929 patients were assigned for model development. Ten variables were selected for model development. The area under the receiver operating characteristic curve (AUCROC) for the RF and SVM methods were 0.953 and 0.947. A total of 329 patients were analyzed for external validation. The AUCROC for the external validation of the RF and SVM method were 0.920 and 0.911.
Conclusion
We successfully constructed predictive model for DN and NDRD by machine learning methods, which were better than traditional ways.
Performance for SVM and other models in external validation
| Models | Sensitivity | Specificity | PPV | NPV | AUCROC | |
| Isolated DN vs. isolated NDRD | SVM | 0.8671 | 0.8889 | 0.9257 | 0.8073 | 0.9108 |
| RF | 0.9048 | 0.8636 | 0.8986 | 0.8716 | 0.9203 | |
| Model-2008 | 0.8926 | 0.7059 | 0.7297 | 0.8807 | 0.8855 | |
| Model-2014 | 0.8581 | 0.8529 | 0.8986 | 0.7982 | 0.9167 | |
| Isolated DN vs. non-DN | SVM | 0.7174 | 0.8897 | 0.8919 | 0.7127 | 0.8462 |
| RF | 0.7348 | 0.8986 | 0.8986 | 0.7348 | 0.8548 | |
| Model-2008 | 0.7324 | 0.7647 | 0.7027 | 0.7901 | 0.8206 | |
| Model-2014 | 0.6875 | 0.8832 | 0.8919 | 0.6685 | 0.8412 |
AUCROC, area under the ROC curve; NPV, negative predictive value; PPV, positive predictive value; SVM, support vector machine; RF, random forest
Analysis flow for the development and evaluation of the model
Funding
- Government Support - Non-U.S.