A Web-based nomogram predicting para-aortic nodal metastasis in incompletely staged patients with endometrial cancer:
Sokbom Kang1, Jong-Min Lee, Jae-Kwan Lee
1*Center for Uterine Cancer, National Cancer Center, Goyang; †Department of Obstetrics and Gynecology, School of Medicine, Kyung Hee University; ‡Department of Obstetrics and Gynecology, Korea University College of Medicine; §Department of Obstetrics and Gynecology, Cancer Research Institute, College of Medicine, Seoul National University, Seoul; ∥Department of Obstetrics and Gynecology, Dongsan Medical Center, Keimyung University, Daegu; ¶Department of Obstetrics and Gynecology, Chonnam National University, Gwangju; #Department of Obstetrics and Gynecology, Gachon University Hospital, Incheon; and **Department of Obstetrics and Gynecology, Busan Paik Hospital, Inje University, Busan, South Korea.
Objective:
The purpose of this study is to develop a Web-based nomogram for predicting the individualized risk of para-aortic nodal metastasis in incompletely staged patients with endometrial cancer.
Methods:
From 8 institutions, the medical records of 397 patients who underwent pelvic and para-aortic lymphadenectomy as a surgical staging procedure were retrospectively reviewed. A multivariate logistic regression model was created and internally validated by rigorous bootstrap resampling methods. Finally, the model was transformed into a user-friendly Web-based nomogram (http://http://www.kgog.org/nomogram/empa001.html).
Results:
The rate of para-aortic nodal metastasis was 14.4% (57/397 patients). Using a stepwise variable selection, 4 variables including deep myometrial invasion, non-endometrioid subtype, lymphovascular space invasion, and log-transformed CA-125 levels were finally adopted. After 1000 repetitions of bootstrapping, all of these 4 variables retained a significant association with para-aortic nodal metastasis in the multivariate analysis-deep myometrial invasion (P = 0.001), non-endometrioid histologic subtype (P = 0.034), lymphovascular space invasion (P = 0.003), and log-transformed serum CA-125 levels (P = 0.004). The model showed good discrimination (C statistics = 0.87; 95% confidence interval, 0.82-0.92) and accurate calibration (Hosmer-Lemeshow P = 0.74).
Conclusions:
This nomogram showed good performance in predicting para-aortic metastasis in patients with endometrial cancer. The tool may be useful in determining the extent of lymphadenectomy after incomplete surgery.
