Using random forest algorithm for glomerular and tubular injury diagnosis

Abstract

Objectives: Chronic kidney disease (CKD) is a common chronic condition with high incidence and insidious onset. Glomerular injury (GI) and tubular injury (TI) represent early manifestations of CKD and could indicate the risk of its development. In this study, we aimed to classify GI and TI using three machine learning algorithms to promote their early diagnosis and slow the progression of CKD.

Methods: Demographic information, physical examination, blood, and morning urine samples were first collected from 13,550 subjects in 10 counties in Shanxi province for classification of GI and TI. Besides, LASSO regression was employed for feature selection of explanatory variables, and the SMOTE (synthetic minority over-sampling technique) algorithm was used to balance target datasets, i.e., GI and TI. Afterward, Random Forest (RF), Naive Bayes (NB), and logistic regression (LR) were constructed to achieve classification of GI and TI, respectively.

Results: A total of 12,330 participants enrolled in this study, with 20 explanatory variables. The number of patients with GI, and TI were 1,587 (12.8%) and 1,456 (11.8%), respectively. After feature selection by LASSO, 14 and 15 explanatory variables remained in these two datasets. Besides, after SMOTE, the number of patients and normal ones were 6,165, 6,165 for GI, and 6,165, 6,164 for TI, respectively. RF outperformed NB and LR in terms of accuracy (78.14, 80.49%), sensitivity (82.00, 84.60%), specificity (74.29, 76.09%), and AUC (0.868, 0.885) for both GI and TI; the four variables contributing most to the classification of GI and TI represented SBP, DBP, sex, age and age, SBP, FPG, and GHb, respectively.

Conclusion: RF boasts good performance in classifying GI and TI, which allows for early auxiliary diagnosis of GI and TI, thus facilitating to help alleviate the progression of CKD, and enjoying great prospects in clinical practice.

Keywords: auxiliary diagnosis; glomerular injury; machine learning; random forest; tubular injury.

FIGURE 1

Before and after SMOTE of response variables for GI and TI. SMOTE, Synthetic Minority Over-Sampling Technique. It’s a good and powerful way to handle imbalanced data, and it was conducted under the parameters of k = 5, C.perc = “balance,” dist = “Overlap.” (A) GI before SMOTE; (B) TI before SMOTE; (C) GI after SMOTE; (D) TI after SMOTE.

FIGURE 2

Workflow of the model construction.

FIGURE 3

Results of feature selection using LASSO. When Lamda is minimum, corresponding features were taken into model construction (14 features for GI, and 15 feature for TI). (A) Feature selection for GI; (B) feature selection for TI.

FIGURE 4

Comparison of the ROC curve areas of the three model classifiers. In model construction, 70% of samples were randomly divided as training set, and the rest 30% were as testing set. AUC (area under curve) was used to evaluate the performance of these three classifiers. (A) AUC of GI in the training set; (B) AUC of GI in the testing set; (C) AUC of TI in the training set; (D) AUC of TI in the testing set.

FIGURE 5

Contributions of explanatory variables to the random forest model. The “%IncMSE” is the increase in mean squared error, where the error of the model prediction is increased by randomly replacing the value of each predictor variable if it is more important. Therefore, a larger value indicates that the variable is more important.