Lupus nephritis or not? A simple and clinically friendly machine learning pipeline to help diagnosis of lupus nephritis

• Published in: Inflammation research 2023-06, Vol.72 (6), p.1315-1324
• Main Authors: Wang, Da-Cheng, Xu, Wang-Dong, Wang, Shen-Nan, Wang, Xiang, Leng, Wei, Fu, Lu, Liu, Xiao-Yan, Qin, Zhen, Huang, An-Fang
• Format: Article
• Language: English
• Subjects: Accuracy; Allergology; Artificial neural networks; Bayes Theorem; Biomedical and Life Sciences; Biomedicine; Biopsy; Chronic conditions; Decision trees; Dermatology; Diagnosis; Feature selection; Humans; Immunology; Learning algorithms; Lupus; Lupus Erythematosus, Systemic; Lupus nephritis; Lupus Nephritis - diagnosis; Machine Learning; Nephritis; Neural networks; Neurology; Original Research Paper; Pharmacology/Toxicology; Proteinuria; Regression analysis; Rheumatology; Support vector machines; Systemic lupus erythematosus; Vitamin A
• ISSN: 1023-3830; 1420-908X
• DOI: 10.1007/s00011-023-01755-7

Objective Diagnosis of lupus nephritis (LN) is a complex process, which usually requires renal biopsy. We aim to establish a machine learning pipeline to help diagnosis of LN. Methods A cohort of 681 systemic lupus erythematosus (SLE) patients without LN and 786 SLE patients with LN was established, and a total of 95 clinical, laboratory data and 17 meteorological indicators were collected. After tenfold cross-validation, the patients were divided into training set and test set. The features selected by collective feature selection method of mutual information (MI) and multisurf were used to construct the models of logistic regression, decision tree, random forest, naive Bayes, support vector machine (SVM), light gradient boosting (LGB), extreme gradient boosting (XGB), and artificial neural network (ANN), the models were compared and verified in post-analysis. Results Collective feature selection method screens out antistreptolysin (ASO), retinol binding protein (RBP), lupus anticoagulant 1 (LA1), LA2, proteinuria and other features, and the hyperparameter optimized XGB (ROC: AUC = 0.995; PRC: AUC = 1.000, APS = 1.000; balance accuracy: 0.990) has the best performance, followed by LGB (ROC: AUC = 0.992; PRC: AUC = 0.997, APS = 0.977; balance accuracy: 0.957). The worst performance is naive Bayes model (ROC: AUC = 0.799; PRC: AUC = 0.822, APS = 0.823; balance accuracy: 0.693). In the composite feature importance bar plots, ASO, RF, Up/Ucr, and other features play important roles in LN. Conclusion We developed and validated a new and simple machine learning pathway for diagnosis of LN, especially the XGB model based on ASO, LA1, LA2, proteinuria, and other features screened out by collective feature selection.