Diabetes Metab J.  2024 Jul;48(4):771-779. 10.4093/dmj.2023.0033.

Construction of Risk Prediction Model of Type 2 Diabetic Kidney Disease Based on Deep Learning

• Yun C
• Tang F
• Gao Z
• Wang W
• Bai F
• Miller JD
• Liu H
• Lee Y
• Lou Q

• KMID: 2558044
• DOI: http://doi.org/10.4093/dmj.2023.0033

Abstract

Background
This study aimed to develop a diabetic kidney disease (DKD) prediction model using long short term memory (LSTM) neural network and evaluate its performance using accuracy, precision, recall, and area under the curve (AUC) of the receiver operating characteristic (ROC) curve.
Methods
The study identified DKD risk factors through literature review and physician focus group, and collected 7 years of data from 6,040 type 2 diabetes mellitus patients based on the risk factors. Pytorch was used to build the LSTM neural network, with 70% of the data used for training and the other 30% for testing. Three models were established to examine the impact of glycosylated hemoglobin (HbA1c), systolic blood pressure (SBP), and pulse pressure (PP) variabilities on the model’s performance.
Results
The developed model achieved an accuracy of 83% and an AUC of 0.83. When the risk factor of HbA1c variability, SBP variability, or PP variability was removed one by one, the accuracy of each model was significantly lower than that of the optimal model, with an accuracy of 78% (P<0.001), 79% (P<0.001), and 81% (P<0.001), respectively. The AUC of ROC was also significantly lower for each model, with values of 0.72 (P<0.001), 0.75 (P<0.001), and 0.77 (P<0.05).
Conclusion
The developed DKD risk predictive model using LSTM neural networks demonstrated high accuracy and AUC value. When HbA1c, SBP, and PP variabilities were added to the model as featured characteristics, the model’s performance was greatly improved.

Keyword

Deep learning; Diabetes mellitus, type 2; Diabetic nephropathies; Risk assessment; Risk factors

Figure

• 

  Fig. 1. Comparison of precision, accuracy and recall of models constructed by long short term memory (LSTM) and support vector machine (SVM).

• 

  Fig. 2. Comparison of receiver operating characteristic curves for the long short term memory (LSTM) and support vector machine (SVM) models. AUC, area under the curve.

• 

  Fig. 3. Comparison of area under the curve areas among four diabetic kidney disease prediction models based on long short term memory (LSTM) neural network. PP, pulse pressure; SBP, systolic blood pressure; HbA1c, glycosylated hemoglobin.

Reference

1. Koye DN, Magliano DJ, Nelson RG, Pavkov ME. The global epidemiology of diabetes and kidney disease. Adv Chronic Kidney Dis. 2018; 25:121–32.

2. Sever MS, Jager KJ, Vanholder R, Stengel B, Harambat J, Finne P, et al. A roadmap for optimizing chronic kidney disease patient care and patient-oriented research in the Eastern European nephrology community. Clin Kidney J. 2020; 14:23–35.

3. Gaede P, Tarnow L, Vedel P, Parving HH, Pedersen O. Remission to normoalbuminuria during multifactorial treatment preserves kidney function in patients with type 2 diabetes and microalbuminuria. Nephrol Dial Transplant. 2004; 19:2784–8.

4. Jiang W, Wang J, Shen X, Lu W, Wang Y, Li W, et al. Establishment and validation of a risk prediction model for early diabetic kidney disease based on a systematic review and meta-analysis of 20 cohorts. Diabetes Care. 2020; 43:925–33.

5. Li L, Yang Y, Zhu X, Xiong X, Zeng L, Xiong S, et al. Design and validation of a scoring model for differential diagnosis of diabetic nephropathy and nondiabetic renal diseases in type 2 diabetic patients. J Diabetes. 2020; 12:237–46.

6. Sun Z, Wang K, Miller JD, Yuan X, Lee YJ, Lou Q. External validation of the risk prediction model for early diabetic kidney disease in Taiwan population: a retrospective cohort study. BMJ Open. 2022; 12:e059139.

7. Xin L, Jin L, Lei L, Liang C, Huiling R. Risk prediction models of type 2 diabetic nephropathy. Chin J Med Libr Inf Sci. 2019; 28:41–5.

8. Gaoxing Q. To construct a predictive diagnostic model of diabetic nephropathy. Available from: https: https://kns.cnki.net/KCMS/detail/detail.aspx?dbname=CMFD201902&filename=1019115309.nh (cited 2023 Jul 14).

9. Mpanya D, Celik T, Klug E, Ntsinjana H. Machine learning and statistical methods for predicting mortality in heart failure. Heart Fail Rev. 2021; 26:545–52.

10. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015; 521:436–44.

11. Assegie TA. An optimized K-Nearest Neighbor based breast cancer detection. J Robot Control. 2021; 2:115–8.

12. Subramanian R, Rubi D, Lakshmi RG, Jain P, Kanneganti SR. Breast cancer lesion detection and classification in radiology images using deep learning. Eur J Mol Clin Med. 2020; 7:677–84.

13. Zhou LQ, Wu XL, Huang SY, Wu GG, Ye HR, Wei Q, et al. Lymph node metastasis prediction from primary breast cancer us images using deep learning. Radiology. 2020; 294:19–28.

14. Guo A, Beheshti R, Khan YM, Langabeer JR 2nd, Foraker RE. Predicting cardiovascular health trajectories in time-series electronic health records with LSTM models. BMC Med Inform Decis Mak. 2021; 21:5.

15. Lewis M, Elad G, Beladev M, Maor G, Radinsky K, Hermann D, et al. Comparison of deep learning with traditional models to predict preventable acute care use and spending among heart failure patients. Sci Rep. 2021; 11:1164.

16. Zeleznik R, Foldyna B, Eslami P, Weiss J, Alexander I, Taron J, et al. Deep convolutional neural networks to predict cardiovascular risk from computed tomography. Nat Commun. 2021; 12:715.

17. Gonzalez G, Ash SY, Vegas-Sanchez-Ferrero G, Onieva Onieva J, Rahaghi FN, Ross JC, et al. Disease staging and prognosis in smokers using deep learning in chest computed tomography. Am J Respir Crit Care Med. 2018; 197:193–203.

18. Bellemo V, Lim ZW, Lim G, Nguyen QD, Xie Y, Yip MYT, et al. Artificial intelligence using deep learning to screen for referable and vision-threatening diabetic retinopathy in Africa: a clinical validation study. Lancet Digit Health. 2019; 1:e35–44.

19. Nielsen KB, Lautrup ML, Andersen JKH, Savarimuthu TR, Grauslund J. Deep learning-based algorithms in screening of diabetic retinopathy: a systematic review of diagnostic performance. Ophthalmol Retina. 2019; 3:294–304.

20. Raman R, Srinivasan S, Virmani S, Sivaprasad S, Rao C, Rajalakshmi R. Fundus photograph-based deep learning algorithms in detecting diabetic retinopathy. Eye (Lond). 2019; 33:97–109.

21. Sayres R, Taly A, Rahimy E, Blumer K, Coz D, Hammel N, et al. Using a deep learning algorithm and integrated gradients explanation to assist grading for diabetic retinopathy. Ophthalmology. 2019; 126:552–64.

22. Jiang H, Ma H, Qian W, Gao M, Li Y, Hongyang Jiang, et al. An automatic detection system of lung nodule based on multigroup patch-based deep learning network. IEEE J Biomed Health Inform. 2018; 22:1227–37.

23. Miotto R, Wang F, Wang S, Jiang X, Dudley JT. Deep learning for healthcare: review, opportunities and challenges. Brief Bioinform. 2018; 19:1236–46.

24. Hochreiter S, Schmidhuber J. Long short-term memory. Neural Comput. 1997; 9:1735–80.

25. Hua Y, Zhao Z, Li R, Chen X, Liu Z, Zhang H. Deep learning with long short-term memory for time series prediction. IEEE Commun Mag. 2019; 57:114–9.

26. Sak H, Senior A, Beaufays F. Long short-term memory based recurrent neural network architectures for large vocabulary speech recognition. ArXiv. 2014; Feb. 5. [Preprint]. https://doi.org/10.48550/arXiv.1402.1128.

27. Makino M, Yoshimoto R, Ono M, Itoko T, Katsuki T, Koseki A, et al. Artificial intelligence predicts the progression of diabetic kidney disease using big data machine learning. Sci Rep. 2019; 9:11862.

28. Xue C, Qianqian Z, Huijun X, Xiaodan Y, Chao L, Taojun L, et al. Effect of glycated hemoglobin A1c variability on diabetic kidney disease in type 2 diabetes mellitus patients. Chin J Diabetes Mellit. 2020; 12:993–8.

29. Luk AO, Ma RC, Lau ES, Yang X, Lau WW, Yu LW, et al. Risk association of HbA1c variability with chronic kidney disease and cardiovascular disease in type 2 diabetes: prospective analysis of the Hong Kong Diabetes Registry. Diabetes Metab Res Rev. 2013; 29:384–90.

30. Viazzi F, Bonino B, Mirijello A, Fioretto P, Giorda C, Ceriello A, et al. Long-term blood pressure variability and development of chronic kidney disease in type 2 diabetes. J Hypertens. 2019; 37:805–13.

31. Microvascular Complications Group of Chinese Diabetes Society. Clinical guideline for the prevention and treatment of diabetic kidney disease in China (2021 edition). Chin J Diabetes Mellit. 2021; 13:762–84.

32. Lou Q, Chen X, Wang K, Liu H, Zhang Z, Lee Y. The impact of systolic blood pressure, pulse pressure, and their variability on diabetes retinopathy among patients with type 2 diabetes. J Diabetes Res. 2022; 2022:7876786.

33. Karim F, Majumdar S, Darabi H, Chen S. LSTM fully convolutional networks for time series classification. IEEE Access. 2018; 6:1662–9.

34. Zhu W, Xu L, Chen X, Lee YJ, Zhang Z, Lou Q. Effects of different blood pressures and their long-term variability on the development of diabetic kidney disease in patients with type 2 diabetes mellitus. Clin Exp Hypertens. 2022; 44:464–9.

35. Brownlee M. Biochemistry and molecular cell biology of diabetic complications. Nature. 2001; 414:813–20.

36. Ceriello A, Kilpatrick ES. Glycemic variability: both sides of the story. Diabetes Care. 2013; 36(Suppl 2):S272–5.

37. Chang CM, Hsieh CJ, Huang JC, Huang IC. Acute and chronic fluctuations in blood glucose levels can increase oxidative stress in type 2 diabetes mellitus. Acta Diabetol. 2012; 49 Suppl 1:S171–7.

38. Chiu WC, Lai YR, Cheng BC, Huang CC, Chen JF, Lu CH. HbA1C variability is strongly associated with development of macroalbuminuria in normal or microalbuminuria in patients with type 2 diabetes mellitus: a six-year follow-up study. Biomed Res Int. 2020; 2020:7462158.

39. Di Flaviani A, Picconi F, Di Stefano P, Giordani I, Malandrucco I, Maggio P, et al. Impact of glycemic and blood pressure variability on surrogate measures of cardiovascular outcomes in type 2 diabetic patients. Diabetes Care. 2011; 34:1605–9.