Predicting Lymph Node Metastasis in Endometrial Cancer Using Serum CA125 Combined with Immunohistochemical Markers PR and Ki67, and a Comparison with Other Prediction Models.

Abstract

We aimed to evaluate the value of immunohistochemical markers and serum CA125 in predicting the risk of lymph node metastasis (LNM) in women with endometrial cancer and to identify a low-risk group of LNM. The medical records of 370 patients with endometrial endometrioid adenocarcinoma who underwent surgical staging in the Obstetrics & Gynecology Hospital of Fudan University were collected and retrospectively reviewed. Immunohistochemical markers were screened. A model using serum cancer antigen 125 (CA125) level, the immunohistochemical markers progesterone receptor (PR) and Ki67 was created for prediction of LNM. A predicted probability of 4% among these patients was defined as low risk. The developed model was externally validated in 200 patients from Shanghai Cancer Center. The efficiency of the model was compared with three other reported prediction models. Patients with serum CA125 50% and Ki67 < 40% in cancer lesion were defined as low risk for LNM. The model showed good discrimination with an area under the receiver operating characteristic curve of 0.82. The model classified 61.9% (229/370) of patients as being at low risk for LNM. Among these 229 patients, 6 patients (2.6%) had LNM and the negative predictive value was 97.4% (223/229). The sensitivity and specificity of the model were 84.6% and 67.4% respectively. In the validation cohort, the model classified 59.5% (119/200) of patients as low-risk, 3 out of these 119 patients (2.5%) has LNM. Our model showed a predictive power similar to those of two previously reported prediction models. The prediction model using serum CA125 and the immunohistochemical markers PR and Ki67 is useful to predict patients with a low risk of LNM and has the potential to provide valuable guidance to clinicians in the treatment of patients with endometrioid endometrial cancer.