| Establishment and validation of diagnostic prediction model for IgA nephropathy based on noninvasive indicators |
| Received:March 05, 2022 Revised:July 11, 2022 Click here to download the full text |
| Citation of this paper:WANG Jiao,WANG Li-zhen,WANG Yu,HE Zheng-jia,BAO Xiao-rong.Establishment and validation of diagnostic prediction model for IgA nephropathy based on noninvasive indicators[J].Chinese Journal of Clinical Medicine,2022,29(4):603-609 |
| Hits: 933 |
| Download times: 296 |
| Author Name | Affiliation | E-mail | | WANG Jiao | Department of Nephrology, Jinshan Hospital, Fudan University, Shanghai 201508, China | | | WANG Li-zhen | Department of Nephrology, Jinshan Hospital, Fudan University, Shanghai 201508, China | | | WANG Yu | Department of Nephrology, Jinshan Hospital, Fudan University, Shanghai 201508, China | | | HE Zheng-jia | Department of Nephrology, Jinshan Hospital, Fudan University, Shanghai 201508, China | | | BAO Xiao-rong | Department of Nephrology, Jinshan Hospital, Fudan University, Shanghai 201508, China | growingfirefly@sina.cn |
|
| Abstract:Objective To establish a noninvasive predictive model for the diagnosis of IgA nephropathy. Methods Clinical data of 276 patients with primary glomerular disease diagnosed by renal biopsy in the Department of Nephrology, Jinshan Hospital, Fudan University from October 2015 to June 2021 were collected. All patients were randomly divided into model training set (n=181, 85 cases of IgA nephropathy and 96 cases of non-IgA nephropathy) and model validation set (n=95, 46 cases of IgA nephropathy and 49 cases of non-IgA nephropathy). Univariate and multivariate logistic regression analysis were performed in the training set to screen non-invasive clinical indicators of IgA nephropathy, and a non-invasive diagnostic model for IgA nephropathy was established, and the model was converted into a visual Nomogram by R language. In the validation set, the diagnostic prediction model established in the training set was used for external validation. The ROC curve was drawn and the area under the curve (AUC) was calculated to distinguish the evaluation model from the validation model. The calibration curve was drawn to evaluate the calibration model. Results Based on the multivariate logistic regression analysis, five predictors were included:blood IgA/C3, blood fibrinogen, hematuria (microscopic hematuria or gross hematuria), blood albumin, and blood high-density lipoprotein. The diagnostic model was established based on the above predictors and the model was transformed into a Nomogram.In the training set, the AUC of the model was 0.934 (P<0.001, 95%CI 0.899-0.970), which had a good degree of differentiation. The optimal diagnostic threshold was 0.437 according to the Youden index, with a sensitivity of 91.8% and a specificity of 85.4%. In the validation set, the AUC was 0.902 (P<0.001, 95%CI 0.837-0.968). When the diagnostic boundary value of the training set was used in the validation set, the sensitivity and specificity were 76.1% and 87.8%, respectively. The calibration curve showd a good agreement between the predicted probability and the actual probability. Conclusion Based on IgA/C3, plasma albumin and another three non-invasive indexes, this noninvasive diagnostic prediction model has high diagnostic efficiency and can be used for the diagnosis of IgA nephropathy. |
| keywords:IgA nephropathy diagnostic model logistic regression Nomogram figure |
| HTML View Full Text View/Add Comment Download reader |
|
|
|