Predictive value of S100A9 for lymph node metastasis in cervical cancer

Lymph node metastasis affects the initial treatment strategy for cervical cancer and is hard to be diagnosed in clinical practice.This paper aims to explore the relationship between calcium-binding A9 (S100A9) and lymph node metastasis (LNM) in cervical cancer, and to determine the predictive value of S100A9 for LNM in cervical cancer.

We performed a retrospective cohort study and collected the pathological data, follow-up data, and paraffin tissue samples of 99 patients with cervical cancer who underwent modified extensive or extensive hysterectomy plus pelvic lymphadenectomy at the Department of Gynecology, Xiangya Hospital, Central South University from January 2013 to December 2018. Immunohistochemistry was used to detect the expression of S100A9 in cervical cancer tissues, and the correlation between S100A9 expression and LNM of cervical cancer, or clinicopathological characteristics were analyzed. The receiver operating characteristic (ROC) curve was used to establish a predictive model for LNM of cervical cancer, and Chi-square test of four-grid table was used to evaluate the diagnostic value of S100A9 for LNM in cervical cancer.

The expression of S100A9 was significantly correlated with LNM. The S100A9 immunohistochemical semi-quantitative score of the LNM group was significantly higher than that in the non-lymph node metastasis group (P<0.001). Moreover, the expression of S100A9 was significantly correlated with histological type, stromal invasion, lymphatic vessel invasion, or LNM (P<0.05). The cut-off of the ROC curve for predicting LNM was 5, with the Youden index of 0.649 and the area under the ROC curve of 0.863. The disease-free survival and overall survival in the S100A9 positive group were significantly shorter than those in the negative group (P<0.05). S100A9 alone had a sensitivity of 71.4%, a specificity of 91.5%, and an accuracy of 85.1% for diagnosing LNM. Imaging had a sensitivity of 32.1%, a specificity of 74.6%, and an accuracy of 60.9%. Combination of S100A9 with image examination in parallel test had a sensitivity of 85.7%, a specificity of 71.2%, and an accuracy of 75.9%, while combination of S100A9 and image examination in serial test had a sensitivity of 17.9%, a specificity of 98.3%, and an accuracy of 72.4%.

S100A9 may be associated with LNM in cervical cancer. S100A9 shows a promising perspective in predicting LNM in cervical cancer. Combination of S100A9 and image examination in serial test has a high specificity for LNM.