IgA Nephropathy Prediction in Children with Machine Learning Algorithms
Abstract
:1. Introduction
- A dataset about IgAN in children was created. And the chi-square test was used to extract the most useful features from the dataset.
- EXtreme Gradient Boosting (XGBoost) was adopted in order to predict whether IgAN disease in children patients would reach ESRD or not within five years using a new dataset instead of the traditional clinical pathology. A decision-making system that was based on the XGBoost algorithm was designed with the Django framework.
- Comparation of the performance of XGBoost with random forest (RF), CatBoost, support vector machines (SVM), k-nearest neighbor (KNN), and extreme learning machine (ELM) was conducted.
2. Materials and Methods
2.1. Dataset
2.2. Feature Selection
2.3. Model
2.4. Performance Evaluation
3. System Implementation
4. Results
5. Discussion
Author Contributions
Funding
Conflicts of Interest
References
- Damico, G. The Commonest Glomerulonephritis in the World: IgA Nephropathy. QJM Int. J. Med. 1987, 64, 709–727. [Google Scholar]
- Wyatt, R.J.; Kritchevsky, S.B.; Woodford, S.Y.; Miller, P.M.; Roy, S., III; Holland, N.H.; Jackson, E.; Bishof, N.A. IgA nephropathy: Long-term prognosis for pediatric patients. J. Pediatr 1995, 127, 913–919. [Google Scholar] [CrossRef]
- Coppo, R.; Robert, T. IgA nephropathy in children and in adults: Two separate entities or the same disease? J. Nephrol. 2020, 33, 1219–1229. [Google Scholar] [CrossRef] [PubMed]
- Coppo, R. Treatment of IgA nephropathy in children: A land without KDIGO guidance. Pediatr. Nephrol. 2020, 2020, 1–6. [Google Scholar] [CrossRef]
- Jonsson, A. Deep Reinforcement Learning in Medicine. Kidney Dis. 2019, 5, 18–22. [Google Scholar] [CrossRef]
- Toback, F.G. Regeneration after acute tubular necrosis. Kidney Int. 1992, 41, 226–246. [Google Scholar] [CrossRef] [Green Version]
- Agar, J.W.; Webb, G.I. Application of machine learning to a renal biopsy database. Nephrol. Dialys. Transplant. 1992, 7, 472–478. [Google Scholar] [CrossRef] [Green Version]
- Ali, F.; El-Sappagh, S.; Islam, S.M.R.; Ali, A.; Attique, M.; Imran, M.; Kwak, K.-S. An intelligent healthcare monitoring framework using wearable sensors and social networking data. Futur. Gener. Comput. Syst. 2020, 114, 23–43. [Google Scholar] [CrossRef]
- Ali, F.; El-Sappagh, S.H.A.; Islam, S.R.; Kwak, D.; Ali, A.; Imran, M.; Kwak, K.-S. A smart healthcare monitoring system for heart disease prediction based on ensemble deep learning and feature fusion. Inf. Fusion 2020, 63, 208–222. [Google Scholar] [CrossRef]
- Leung, R.K.; Wang, Y.; Ma, R.C.; Luk, A.O.; Lam, V.; Ng, M.; So, W.Y.; Tsui, S.K.; Chan, J.C. Using a multi-staged strategy based on machine learning and mathematical modeling to predict genotype-phenotype risk patterns in diabetic kidney disease: A prospective case–control cohort analysis. BMC Nephrol. 2013, 14, 162. [Google Scholar] [CrossRef]
- Hamedan, F.; Orooji, A.; Sanadgol, H.; Sheikhtaheri, A. Clinical Decision Support System to Predict Chronic Kidney Disease: A Fuzzy Expert System Approach. Int. J. Med. Info. 2020, 138, 104134. [Google Scholar] [CrossRef] [PubMed]
- Gupta, R.; Sharma, A.; Singh, S.; Dinda, A.K. Rule-based decision support system in the biopsy diagnosis of glomerular diseases. J. Clin. Pathol. 2011, 64, 862. [Google Scholar] [CrossRef] [PubMed]
- Takayuki, K.; Masaki, O.; Akira, K.; Michiharu, K.; Kyoichi, H.; Jun, K.; Masaki, M.; Ryosuke, Y.; Atsushi, S. Risk Prediction of Diabetic Nephropathy via Interpretable Feature Extraction from EHR Using Convolutional Autoencoder. Stud. Health Technol. Info. 2018, 247, 106–110. [Google Scholar]
- Lemley, K.V.; Lafayette, R.A.; Derby, G.C.; Blouch, K.L.; Anderson, L.; Efron, B.; Myers, B.D. Prediction of early progression in recently diagnosed IgA nephropathy. Nephrol. Dialys. Transplant. 2007, 23, 213–222. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Noh, J.; Punithan, D.; Lee, H.; Lee, J.; Kim, Y.; Kim, D.; McKay, R.B. Machine Learning Models and Statistical Measures for Predicting the Progression of IgA Nephropathy. Int. J. Soft. Eng. Know. Eng. 2015, 25, 829–849. [Google Scholar] [CrossRef]
- Ducher, M.; Kalbacher, E.; François, C.; de Vilaine, J.F.; McGregor, B.; Fouque, D.; Fauvel, J.P. Comparison of a Bayesian Network with a Logistic Regression Model to Forecast IgA Nephropathy. BioMed Res. Int. 2013, 2013, 686150. [Google Scholar] [CrossRef] [Green Version]
- Barbour, S.J.; Reich, H.N. Risk Stratification of Patients with IgA Nephropathy. Am. J. Kidney Dis. 2012, 59, 865–873. [Google Scholar] [CrossRef] [Green Version]
- Barbour, S.J.; John, F. Predicting the future in immunoglobulin A nephropathy: A new international risk prediction tool. Nephrol. Dialys. Transplant. 2020, 35, 379–382. [Google Scholar] [CrossRef]
- Barbour, S.J.; Coppo, R.; Zhang, H.; Liu, Z.-H.; Suzuki, Y.; Matsuzaki, K.; Katafuchi, R.; Er, L.; Espino-Hernandez, G.; Kim, J.; et al. Evaluating a New International Risk-Prediction Tool in IgA Nephropathy. JAMA Intern. Med. 2019, 179, 942–952. [Google Scholar] [CrossRef]
- Chen, T.; Li, X.; Li, Y.; Xia, E.; Qin, Y.; Liang, S.; Xu, F.; Liang, D.; Zeng, C.; Liu, Z. Prediction and Risk Stratification of Kidney Outcomes in IgA Nephropathy. Am. J. Kidney Dis. 2019, 74, 300–309. [Google Scholar] [CrossRef]
- Liu, Y.; Zhang, Y.; Liu, D.; Tan, X.; Tang, X.; Zhang, F.; Xia, M.; Chen, G.; He, L.; Zhou, L.; et al. Prediction of ESRD in IgA Nephropathy Patients from an Asian Cohort: A Random Forest Model. Kidney Blood Press. Res. 2018, 43, 1852–1864. [Google Scholar] [CrossRef] [PubMed]
- Han, X.; Zheng, X.; Wang, Y.; Sun, X.; Xiao, Y.; Tang, Y.; Qin, W. Random forest can accurately predict the development of end-stage renal disease in immunoglobulin a nephropathy patients. Annal. Translat. Med. 2019, 7, 234. [Google Scholar] [CrossRef] [PubMed]
- Bertram, J.F.; Goldstein, S.L.; Pape, L.; Schaefer, F.; Shroff, R.C.; Warady, B.A. Kidney disease in children: Latest advances and remaining challenges. Nat. Rev. Nephrol. 2016, 12, 182–191. [Google Scholar] [CrossRef] [PubMed]
- Coppo, R.; Troyanov, S.; Camilla, R.; Hogg, R.J.; Cattran, D.C.; Terence Cook, H.; Feehally, J.; Roberts, I.S.D.; Amore, A.; Alpers, C.E.; et al. The Oxford IgA nephropathy clinicopathological classification is valid for children as well as adults. Kidney Int. 2010, 77, 921–927. [Google Scholar] [CrossRef] [PubMed]
- Shima, Y.; Nakanishi, K.; Sato, M.; Hama, T.; Mukaiyama, H.; Togawa, H.; Tanaka, R.; Nozu, K.; Sako, M.; Iijima, K.; et al. IgA nephropathy with presentation of nephrotic syndrome at onset in children. Pediatr. Nephrol. 2017, 32, 457–465. [Google Scholar] [CrossRef] [PubMed]
- Chen, T.; Guestrin, C. XGBoost: A Scalable Tree Boosting System. arXiv 2016, arXiv:1603.02754. [Google Scholar] [CrossRef] [Green Version]
- Hanley, J.A.; Mcneil, B.J. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982, 143, 29–36. [Google Scholar] [CrossRef] [Green Version]
Variable | Range | Mean | Standard Deviation |
---|---|---|---|
Age (years) | 0–18 | 14 | 3.59 |
Scr (μmol/L) | 16.00–1154.50 | 71.12 | 67.86 |
CHOL (mmol/L) | 0.20–19.00 | 5.49 | 2.70 |
TG (mmol/L) | 0.17–22.00 | 1.62 | 1.25 |
UA (μmol/L) | 0.53–777.00 | 347.87 | 97.63 |
ALB (g/mL) | 7.60–65.20 | 41.88 | 120.48 |
Complement C3 (g/L) | 0.24–25.00 | 1.06 | 0.74 |
FSGS | 0.00–61.50 | 4.20 | 8.75 |
Spherically sclerotic | 0.00–84.20 | 3.75 | 9.21 |
Crescent ratio | 0.00–77.80 | 5.25 | 9.50 |
eGFR (mL·min−1·(1.72 m2)−1) | 4.90–141.75 | 111.77 | 21.11 |
Variable | Possible Values | Numeric Value | Number of Records | Percentage (%) |
---|---|---|---|---|
Gender | M/F | 1/0 | 776/369 | 67.78/32.22 |
Ur_C3 | normal/abnormal | 0/1 | 347/798 | 30.31/69.69 |
2-m | normal/abnormal | 0/1 | 214/931 | 18.69/81.31 |
Ur_NAG | normal/abnormal | 0/1 | 909/236 | 79.39/20.61 |
Ur_RBP | normal/abnormal | 0/1 | 164/981 | 14.32/85.68 |
M | 0/1/2/3 | 283/562/267/33 | 24.72/49.08/23.32/2.88 | |
E | 0/1/2 | 756/386/3 | 66.03/33.71/0.26 | |
S | 0/1 | 728/417 | 63.58/36.42 | |
T | 0/1/2/3 | 845/251/42/7 | 73.80/21.92/3.67/0.61 | |
C | 0/1/2 | 648/444/53 | 56.59/38.78/4.63 | |
IgA | 0/1/2/3 | 16/31/1082/16 | 1.40/2.70/94.50/1.40 | |
IgG | 0/0.5/1/2 | 882/1/177/85 | 77.03/0.09/15.46/7.42 | |
IgM | 0/0.5/1/2 | 669/13/418/45 | 58.43/1.14/36.50/3.93 | |
C3 | 0/0.5/1/2 | 207/9/173/756 | 18.08/0.79/15.11/66.02 | |
C4 | 0/1 | 1133/12 | 98.95/1.05 | |
C1q | 0/1/2 | 1094/49/2 | 95.55/4.28/0.17 | |
Hypertension | yes/no | 1/0 | 322/823 | 28.12/71.88 |
ACEI_ARB | 0/1 | 862/283 | 75.28/24.72 | |
Immunosuppressive therapy | 0/1 | 726/419 | 63.41/36.59 | |
Lipid lowering | 0/1 | 738/407 | 64.45/35.55 | |
Tonsillectomy | 0/1 | 1124/21 | 98.17/1.83 | |
Loop necrosis | 0/1/2/3 | 1039/91/7/8 | 90.74/7.95/0.61/0.70 | |
Arterial hyaline degeneration | 0/1/5 | 1015/129/1 | 88.65/11.26/0.09 | |
Medullary interstitial fibrosis | 0/1/2/3 | 1090/39/11/5 | 95.20/3.41/0.96/0.43 | |
Thickening and stratification of elastic layer of interlobular artery | 0/1 | 1108/37 | 96.77/3.23 | |
Vacuolar degeneration of arteriole smooth muscle cells | 0/1 | 1085/60 | 94.76/5.24 |
Variable | Possible Values | Numeric Value | Number of Records | Percentage (%) |
---|---|---|---|---|
ESRD | yes | 1 | 567 | 49.52 |
no | 0 | 578 | 50.48 |
Variable | Score | p-Value |
---|---|---|
Scr | 827.885 | 0 |
FSGS | 716.552 | 0 |
Crescent ratio | 658.534 | 0 |
ALB | 621.352 | 0 |
UA | 461.223 | 0 |
Spherically sclerotic | 321.243 | 0 |
CHOL | 279.544 | 0 |
Ur_NAG | 119.224 | 0 |
eGFR | 163.170 | 0 |
TG | 56.108 | 0 |
E | 27.470 | 0 |
T | 23.479 | 0 |
C | 18.653 | 0 |
M | 18.451 | 0 |
IgM | 16.973 | 0 |
C3 | 13.189 | 0 |
Gender | 0.203 | 0.652 |
Ur_C3 | 0.024 | 0.878 |
2-m | 7.233 | 0.007 |
Ur_RBP | 5.823 | 0.016 |
Complement C3 | 0.003 | 0.953 |
S | 5.760 | 0.016 |
IgA | 0.025 | 0.875 |
IgG | 0.816 | 0.366 |
C4 | 0.001 | 0.973 |
C1q | 3.444 | 0.063 |
Hypertension | 11.592 | 0.001 |
ACEI_ARB | 0.211 | 0.646 |
Immunosuppressive therapy | 0.232 | 0.630 |
Lipid lowering | 9.725 | 0.002 |
Tonsillectomy | 0.488 | 0.485 |
Loop necrosis | 0.024 | 0.877 |
Arterial hyaline degeneration | 0.012 | 0.911 |
Medullary interstitial fibrosis | 5.655 | 0.017 |
Thickening and stratification of elastic layer of interlobular artery | 6.373 | 0.012 |
Vacuolar degeneration of arteriole smooth muscle cells | 0.490 | 0.484 |
Algorithm | Accuracy | Precision | Recall | F1_Score | AUC |
---|---|---|---|---|---|
XGBoost | 0.7860 | 0.7596 | 0.7670 | 0.7633 | 0.8511 |
RF | 0.7642 | 0.7426 | 0.7282 | 0.7353 | 0.8507 |
CatBoost | 0.7642 | 0.7379 | 0.7379 | 0.7379 | 0.8454 |
KNN | 0.7555 | 0.7327 | 0.7184 | 0.7255 | 0.8090 |
SVM | 0.7642 | 0.7333 | 0.7476 | 0.7404 | 0.8272 |
ELM | 0.7598 | 0.7264 | 0.7476 | 0.7368 | 0.8174 |
Variables | Importance Score |
---|---|
Ur_NAG | 0.191591 |
ALB | 0.175141 |
CHOL | 0.104368 |
Crescent ratio | 0.085010 |
FSGS | 0.067488 |
Scr | 0.062304 |
TG | 0.050906 |
Spherically sclerotic | 0.034839 |
UA | 0.034522 |
M | 0.033158 |
IgM | 0.032431 |
T | 0.031299 |
C3 | 0.025893 |
C | 0.025279 |
eGFR | 0.024145 |
E | 0.021627 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Zhang, P.; Wang, R.; Shi, N. IgA Nephropathy Prediction in Children with Machine Learning Algorithms. Future Internet 2020, 12, 230. https://doi.org/10.3390/fi12120230
Zhang P, Wang R, Shi N. IgA Nephropathy Prediction in Children with Machine Learning Algorithms. Future Internet. 2020; 12(12):230. https://doi.org/10.3390/fi12120230
Chicago/Turabian StyleZhang, Ping, Rongqin Wang, and Nianfeng Shi. 2020. "IgA Nephropathy Prediction in Children with Machine Learning Algorithms" Future Internet 12, no. 12: 230. https://doi.org/10.3390/fi12120230