AUTHOR=Xu Jie , Chen Can , Xiong Jing , Wang Hui , Linghu Hua 

TITLE=Predictive Value of Serum Cytokeratin 19 Level for the Feasibility of Conserving Ovaries in Endometrial Cancer

JOURNAL=Frontiers in Medicine

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2021.670109

DOI=10.3389/fmed.2021.670109

ISSN=2296-858X

ABSTRACT=Objective. To determine the predictive value of cytokeratin 19 (CK19) for evaluating the safety of ovarian preservation in patients with endometrial cancer(EC).
Methods: 517 EC patients hospitalized from November May 2010 to June 2016 were reviewed retrospectively. Preoperative tumor biomarkers including CA125, HE4, CK19 and CA19-9 were obtained. Predictive biomarkers associated with ovarian metastasis were selected using Univariate and multivariate Logistic regression. The cut-off values was determined by receiver operator curves (ROC). Kaplan-Meier method and Cox multivariate regression model was used to perform survival analysis.
Results：Among clinical parameters and biomarkers included, age>65, type II EC, CA125 ≥ 35u/ml, CK19 >3.3ng/ml and myometrial invasion≥50% depth appeared as significant predictors of the risk of ovarian involvement in univariable logistic analysis. In multivariable analysis, CK19 > 3.3ng/ml (OR=11.541, 95%CI: 1.968-67.668, P=0.007) and Type II EC (OR=8.336, 95%CI: 1.456-47.722, P=0.017) were independent risk predictors of ovarian metastasis in premenopausal women. In premenopausal women with Type I EC (n=142), CK19 level could satisfactorily predict the risk of ovarian metastasis (AUC=0.860, 95%CI: 0.792-0.912, P<0.001),and when the cut-off point was set as 2.45 ng/ml, the negative predictive value and negative likelihood ratio were 99% and 0.19, with the maximum Youden index of 0.598. 
Conclusions. The present study advocates the necessity of incorporating serum CK19 measurement into the preoperative evaluation of EC, especially as extension of current standard approach with ovarian preservation counseling.