Establishment and Validation of a Model for Disease-Free Survival Rate Prediction Using the Combination of microRNA-381 and Clinical Indicators in Patients with Breast Cancer

Abstract

Objective: We have found that miR-381 can regulate the proliferation of breast cancer cells by regulating TWIST protein, it can serve as a potential marker for the tumor progression. Thus, we herein establishment and validation of a model for predicting disease progression in patients with breast cancer using a combination of microRNA-381 (miR-381) and clinical indicators. Methods: Data from 160 breast cancer patients in the First People’s Hospital of Lianyungang were collected, The relationship between miR-381 expression and tumor subtype was analyzed. The Kaplan–Meier (K–M) curve method was used to investigate the disease-free survival rate, while multivariate Cox regression analysis was used to investigate the risk factors affecting the prognosis of the patients. A model for predicting disease progression was subsequently established and validated. Results: The miR-381 was significantly higher in the stage I patients than stage II/III patients. The miR-381 level of triple-negative breast cancer (TNBC) type was significantly decreased. The miR-381 could be used to effectively predict the prognosis, using cut-off value of 0.2515, with a sensitivity of 65.38% (51.8–76.85%), specificity of 75.00% (46.77–91.11%). The K–M survival curve indicated that the patients with higher miR-381 expression had a better prognosis. The miR-381+Ki-67+TN model and TN (T and N in TNM staging) model were established and subsequently compared. The TN model had an area under the curve (AUC) of 0.479 (95% CI 0.329, 0.629); in comparison, the our model had an AUC of 0.719 (95% CI 0.580, 0.857), showing better performance. Conclusion: The miR-381 expression was correlated with different (TNM) stages and tumor subtypes. The higher the TNM stage, the lower the miR-381 expression in the tumor tissue, while it was significantly decreased in TNBC. A prediction model consisting of combination of miR-381 and Ki-67 and TN indicators could predict disease progression more effectively.