CK19 protein expression: the best cutoff value on the prognosis and the prognosis model of hepatocellular carcinoma

Abstract

Background and objective: In clinical practice, CK19 can be an important predictor for the prognosis of HCC. Due to the high incidence and mortality rates of HCC, more effective and practical prognostic prediction models need to be developed urgently. Methods: A total of 1,168 HCC patients, who underwent radical surgery at the Guangxi Medical University Cancer Hospital, between January 2014 and July 2019, were recruited, and their clinicopathological data were collected. Among the clinicopathological data, the optimal cutoff value of CK19-positive HCC was determined by calculating the area under the curve (AUC) using survival analysis and time-dependent receiver operating characteristic (timeROC) curve analysis. The predictors were screened using univariate and multivariate COX regression and least absolute shrinkage and selection operator (LASSO) regression to construct nomogram prediction models, and their predictive potentials were assessed using calibration curves and AUC values. Results: The 0% positive rate of CK19 was considered the optimal cutoff value to predict the poor prognosis of CK19-positive HCC. The survival analysis of 335 CK19-positive HCC showed no significant statistical differences in the overall survival (OS) and disease-free survival (DFS) of CK19-positive HCC patients. A five-factor risk (CK19, CA125, Edmondson, BMI, and tumor number) scoring model and an OS nomograph model were constructed and established, and the OS nomograph model showed a good predictive performance and was subsequently verified. Conclusion: A 0% expression level of CK19 protein may be an optimal threshold for predicting the prognosis of CK19-positive HCC. Based on this, CK19 marker a good nomogram model was constructed to predict HCC prognosis.