Background: The objective of this study was to develop a nomogram that can predict lymph node metastasis (LNM) in patients with cervical adenocarcinoma (cervical AC).
Methods: A total of 219 patients with cervical AC who had undergone radical hysterectomy and lymphadenopathy between 2005 and 2021 were selected for this study. Both univariate and multivariate logistic regression analyses were performed to analyze the selected key clinicopathologic features and develop a nomogram and underwent internal validation to predict the probability of LNM.
Results: Lymphovascular invasion (LVI), tumor size ≥ 4 cm, and depth of cervical stromal infiltration were independent predictors of LNM in cervical AC. However, the Silva pattern was not found to be a significant predictor in the multivariate model. The Silva pattern was still included in the model based on the improved predictive performance of the model observed in the previous studies. The concordance index (C-index) of the model increased from 0.786 to 0.794 after the inclusion of the Silva pattern. The Silva pattern was found to be the strongest predictor of LNM among all the pathological factors investigated, with an OR of 4.37 in the nomogram model. The nomogram developed by incorporation of these four predictors performed well in terms of discrimination and calibration capabilities (C - index = 0.794; 95% confidence interval (CI), 0.727-0.862; Brier score = 0.127). Decision curve analysis demonstrated that the nomogram was clinically effective in the prediction of LNM.
Conclusion: In this study, a nomogram was developed based on the pathologic features, which helped to screen individuals with a higher risk of occult LNM. As a result, this tool may be specifically useful in the management of individuals with cervical AC and help gynecologists to guide clinical individualized treatment plan.
Copyright © 2022 Yongju Tian et al.