Development and Validation of Novel Nomograms to Predict the Overall Survival and Cancer-Specific Survival of Cervical Cancer Patients With Lymph Node Metastasis

Abstract

Objective: The objective of this study was to establish and validate novel individualized nomograms for predicting the overall survival (OS) and cancer-specific survival (CSS) in cervical cancer patients with lymph node metastasis.

Methods: A total of 2,956 cervical cancer patients diagnosed with lymph node metastasis (American Joint Committee on Cancer, AJCC N stage=N1) between 2000 and 2018 were included in this study. Univariate and multivariate Cox regression models were applied to identify independent prognostic predictors, and the nomograms were established to predict the OS and CSS. The concordance index (C-index), calibration curves, and receiver operating characteristic (ROC) curves were applied to estimate the precision and discriminability of the nomograms. Decision-curve analysis (DCA) was used to assess the clinical utility of the nomograms.

Results: Tumor size, log odds of positive lymph nodes (LODDS), radiotherapy, surgery, T stage, histology, and grade resulted as significant independent predictors both for OS and CSS. The C-index value of the prognostic nomogram for predicting OS was 0.788 (95% CI, 0.762-0.814) and 0.777 (95% CI, 0.758-0.796) in the training and validation cohorts, respectively. Meanwhile, the C-index value of the prognostic nomogram for predicting CSS was 0.792 (95% CI, 0.767-0.817) and 0.781 (95% CI, 0.764-0.798) in the training and validation cohorts, respectively. The calibration curves for the nomograms revealed gratifying consistency between predictions and actual observations for both 3- and 5-year OS and CSS. The 3- and 5-year area under the curves (AUCs) for the nomogram of OS and CSS ranged from 0.781 to 0.828. Finally, the DCA curves emerged as robust positive net benefits across a wide scale of threshold probabilities.

Conclusion: We have successfully constructed nomograms that could predict 3- and 5-year OS and CSS of cervical cancer patients with lymph node metastasis and may assist clinicians in decision-making and personalized treatment planning.

Keywords: cancer-specific survival; cervical cancer; lymph node metastasis; nomogram; overall survival.

Figure 1

The flow chart of the patient’s selection process.

Figure 2

The optimal cutoff values for tumor size (A, B) and LODDS (C, D) via X-tile software analysis. The optimal tumor size cutoff values calculated by overall survival were 38 and 64 mm. The optimal LODDS cutoff values calculated by overall survival were -0.9 and -0.2. Tumor size was divided into ≤ 38 mm (sky blue), 39–64 mm (gray), and ≥ 65 mm (pink purple) subgroups. The LODDS was divided into three subgroups: LODDS1 (LODDS ≤ -0.9, sky blue), LODDS2 (-0.9 < LODDS ≤ -0.2, gray), and LODDS3 (LODDS > -0.2, pink purple). LODDS, log odds of positive lymph nodes.

Figure 3

Nomograms for predicting 3- and 5-year OS (A) and CSS (B) in cervical cancer patients with lymph node metastasis. OS, overall survival; CSS, cancer-specific survival; LODDS, log odds of positive lymph nodes; SCC, squamous cell carcinoma; AC, adenocarcinoma; ASC, adenosquamous carcinoma.

Figure 4

Calibration curves for 3- and 5-year OS and CSS of the prognostic nomograms. Calibration curves for 3- and 5-year OS prediction in the training cohort (A, C) and the validation cohort (B, D). Calibration curves for 3- and 5-year CSS prediction in the training cohort (E, G) and the validation cohort (F, H). OS, overall survival; CSS, cancer-specific survival.

Figure 5

ROC curves for 3- and 5-year OS and CSS of the prognostic nomograms. ROC curves for 3- and 5-year OS in the training cohort (A, C) and the validation cohort (B, D). ROC curves for 3- and 5-year CSS in the training cohort (E, G) and the validation cohort (F, H). OS, overall survival; CSS, cancer-specific survival.

Figure 6

Decision curve analysis for 3- and 5-year OS and CSS of the prognostic nomograms. Decision curves for 3- and 5-year OS in the training cohort (A, C) and the validation cohort (B, D). Decision curves for 3- and 5-year CSS in the training cohort (E, G) and the validation cohort (F, H). OS, overall survival; CSS, cancer-specific survival.