Prediction model of pelvic lymph node metastasis in early stage cervical cancer and its clinical value.

ABSTRACT

AIM: This study was designed to investigate the risk factors of pelvic lymph node metastasis in early stage cervical cancer in order to establish a prediction model for this metastasis and to explore the feasibility of conservative surgery. METHODS: The records of 207 stage IB-IIA cervical cancer patients were retrospectivly analyzed. The risk factors of pelvic lymph node metastasis were analyzed using univariate and multivariate methods. The prediction model for pelvic lymph node metastasis was established by logistic regression. RESULTS: Without preoperative adjuvant therapy, the metastatic rate of pelvic lymph node in stage IB-IIA cervical cancer was 25.1%. The serum SCCAg, the tumor diameter, the depth of cervical stroma invasion, and the cervical canal involvement were revealed as the risk factors of pelvic lymph node metastasis by univariate analysis (P<0.05). Multivariate analysis showed that the serum SCCAg and the depth of cervical stroma invasion were the independent risk factors of pelvic lymph node metastasis (P<0.05, OR = 6.917, 2.227). The patients were divided into three groups according to different independent risk factors the low-risk group, the medium-risk group, and the high-risk group, which showed metastatic rates of pelvic lymph node of 5.7%, 16.9%, and 48.7%, respectively (P<0.001). A prediction model for pelvic lymph node metastasis was established as follows Logti(P) = -2.534 + serum SCCAg×1.934 + depth of cervical stroma invasion×0.801. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of this prediction model were 53.8%, 83.9 %, 52.8%, 84.4%, and 76.3%, respectively. CONCLUSION: The serum SCCAg and the depth of cervical stroma invasion were the independent risk factors of pelvic lymph node metastasis in early stage cervical cancer. The proposed prediction model may help to improve the conservative surgery for early stage cervical cancer.