Nomogram for Predicting Lymph Node Metastasis in Cervical Cancer: Comparison of 2009 and 2018 Staging Systems

Abstract OBJECTIVES Lymph node metastasis (LNM) is crucial for personalized treatment in cervical cancer (CC) patients; this study evaluates the enhancement of LNM prediction models by integrating the 2018 FIGO staging system into those based on the 2009 FIGO cohort and assesses the models' net benefit. METHORDS: We included 691 patients: 348 with 2009 FIGO stage IB1-IIA2 CC (Cohort 1) for model development and external validation, and 343 with 2018 FIGO stage IB1-IIICr CC (Cohort 2) for external validation. Variables were selected using regression analyses and LASSO methods. Nomogram models were evaluated with ROC curves, C-index, Integrated Discrimination Improvement (IDI), and Net Reclassification Index (NRI). RESULTS: Model 1 included tumor size, depth of stromal invasion (DSI), SCC-Ag, and neutrophil-to-lymphocyte ratio (NLR). Model 2 incorporated the FIGO staging system. Model 1 showed suitable calibration with discrimination accuracy of 0.807 in the training cohort and 0.772 in the testing cohort. External validation yielded an AUC of 0.743. Model 2, with the 2018 FIGO image staging, demonstrated superior performance for predicting LNM, with a discrimination accuracy of 0.877, significant ΔC-index (P = 0.0024). Adding the 2009 FIGO clinical staging did not significantly enhance predictive value (ΔC-index P = 0.896). Model 2's predictive performance improved by 23.18% in the low-to-intermediate risk group and 17.24% in the high-risk group. CONCLUDSION: Incorporating the 2018 FIGO image staging system into the 2009 FIGO clinical staging model significantly improved LNM prediction, highlighting the importance of imaging-based staging systems and facilitating their application across different time periods.