Abstract
OBJECTIVE: To investigate the construction and clinical application value of Nomogram predictive model for recurrent spontaneous abortion (RSA) based on the expressions of MALAT1 (Metastasis Associated Lung Adenocarcinoma Transcript 1), miR-515-5p, and MCL1 (Myeloid Cell Leukemia-1) mRNA. METHODS: According to the 7:3 ratio, the patients were divided into the training set (n = 135) and the validation set (n = 58) by the complete random method. In the training set, multivariate Logistic regression was used to analyze the risk factors for RSA and a Nomogram prediction model was constructed based on the expressions of MALAT1, miR-515-5p, and MCL1 mRNA. The receiver operating characteristic (ROC) curve and calibration curve were drawn to evaluate the prediction performance of the Nomogram model and verified in the validation set. At the same time, decision curve analysis (DCA) was used to evaluate the clinical application value of the Nomogram model. RESULTS: There were 38 cases (28.15%) of RSA in the training set and 17 cases (29.31%) in the validation set. During the training set, advanced maternal age, number of abortions >3, low progesterone level in early pregnancy, high expression of MALAT1, low expression of miR-515-5p and low expression of MCL1 mRNA were the independent risk factors for RSA (p < 0.05). We will further construct the Nomogram prediction model. In the training set and validation set, the C-index values of Nomogram model were 0.915 and 0.869, respectively. The calibration curve showed that the predicted values accorded well with the actual values, with the average absolute errors of 0.012 and 0.034, respectively. The p-values of Hosmer-Lemeshow test were 0.362 and 0.779, respectively, indicating that the model had good calibration and fitting performance. The ROC curve showed that the Nomogram models in the training and validation sets predicted AUC values for RSA to be 0.916 (95% CI, 0.858-0.974) and 0.867 (95% CI, 0.742-0.992), respectively. CONCLUSION: The Nomogram prediction model constructed based on the expressions of MALAT1, miR-515-5p, and MCL1 mRNA is helpful to predict the occurrence of RSA in the early stage and guide the making of appropriate clinical decisions.