Development and validation of a noninvasive diagnostic model for IgA nephropathy

Objective To explore the development and clinical application value of Nomogram model, a noninvasive early diagnosis model, in IgA nephropathy. Methods The clinical data of 712 patients with primary glomerular disease diagnosed by renal histopathological examination in Affiliated Hospital of Qingdao University during October 1, 2010 to August 31, 2019 were collected retrospectively, including 241 cases of IgA nephropathy and 471 cases of non-IgA nephropathy. According to the time of case inclusion, the patients were divided into the training set (n=426, 156 cases of IgA nephropathy and 270 cases of non-IgA nephropathy) and the validation set (n=286, 85 cases of IgA nephropathy and 201 cases of non-IgA nephropathy). Univariate and multivariate logistic regression equations were used to analyze the risk factors for diagnosing IgA nephropathy in patients of training set. Nomogram model for noninvasive diagnosis of IgA nephropathy was established according to the akichi information criteria (AIC) and applied to the validation set for validation. The discriminant degree, calibration degree and clinical practicability of the model were verified and evaluated by receiver operating characteristic curve (ROC), calibration curve and decision curve analysis (DCA), respectively. Results Multivariate logistic regression results showed that the age (OR=0.966, 95%CI 0.947-0.985, P=0.001), IgA/C3 ratio (OR=1.889, 95%CI 1.468-2.432, P＜0.001), serum albumin (OR=1.091, 95%CI 1.047-1.136, P＜0.001), total cholesterol (OR=0.810, 95%CI 0.694-0.946, P=0.008), and gross hematuria (OR=6.858, 95%CI 1.867-25.189, P=0.004) of patients with primary glomerular disease were independent factors for the diagnosis of IgA nephropathy. Nomogram diagnostic model was constructed based on the above indicators, and the areas under ROC curve were 0.880 and 0.887 respectively in the training set and the validation set. The calibration curve showed that the predicted probability of the model was in good agreement with the actual probability. DCA showed that the safety and clinical net benefit of the model were higher. Conclusions The Nomogram model has high accuracy and clinical practicality in diagnosing IgA nephropathy, and can be used for noninvasive and early diagnosis of IgA nephropathy to enable patients to receive early treatment.