Risk Prediction for Rapidly Progressive Interstitial Lung Disease in Anti-MDA5-Positive Dermatomyositis: The CRAFT Model

Abstract

Background Anti-melanoma differentiation-associated protein 5-positive dermatomyositis (MDA5 + DM) is characterized by a life-threatening complication of rapidly progressive interstitial lung disease (RP-ILD). Early prediction of RP-ILD can enhance diagnostic accuracy and therapeutic efficacy. This study was conducted to develop a nomogram model for predicting RP-ILD in patients with MDA5 + DM. Material/Methods We retrospectively analyzed 53 patients with MDA5 + DM, of whom 21 patients were diagnosed with RP-ILD between January 2018 and January 2021. Univariate analysis ( t test, Mann-Whitney U test, chi-squared test, or Fisher’s exact test) and receiver operating characteristic (ROC) analysis were used to select candidate variables. Multivariate logistic regression analysis was conducted to construct a prediction model, which was subsequently transformed into a nomogram. ROC analysis, calibration curve and decision curve analysis were performed to evaluate the model’s performance. The bootstrapping method (resampling=500) was used for internal validation. Results We successfully established a nomogram, called the CRAFT model, to predict RP-ILD in MDA5 + DM patients. The model included 4 variables, namely C-reactive protein-to-albumin ratio, red blood cell distribution width-coefficient of variation, fever status, and CD3 + T cells. The model presented high predictive power and a good performance in calibration curve and decision curve analysis. In addition, the model had a good predictive ability in internal validation. Conclusions The CRAFT model could help to predict RP-ILD in patients with MDA5 + DM.