Analysis of prognostic factors of metastatic endometrial cancer based on surveillance, epidemiology, and end results database

Abstract

Objective: To explore the risk factors for survival and prognosis of patients with metastatic endometrial cancer and to build and verify a reliable prediction model.

Methods: We retrospectively analyzed patients diagnosed with metastatic endometrial cancer in the US Surveillance, Epidemiology, and End Results (SEER) database between January 2010 and December 2015. Univariate and multivariate Cox regression analyses were used to assess clinical variables impact on survival and to construct nomograms. The results of the consistency index (C-index), subject operating characteristic (ROC) curve, and calibration curve were used to evaluate the predictive ability of the nomogram.

Results: This study included 3,878 patients with metastatic endometrial cancer. In the univariate analysis, variables associated with overall survival (OS) and cancer-specific survival (CSS) included age, race, marital status, pathological type, pathological grade, T-stage, N-stage, surgery, radiotherapy, chemotherapy, bone metastasis, brain metastasis, liver metastasis, and lung metastasis. In the multivariate analysis, age, race, pathological type, pathological grade, T-stage, N-stage, surgery, radiotherapy, chemotherapy, brain metastasis, liver metastasis, and lung metastasis were independent risk factors for OS and CSS (all P < 0.05). Combined with the results of the multiple factors, the 1-, 3-, 5-, and 8-year nomograms were constructed. For OS and CSS, T-stage had the greatest impact on the adverse prognosis of patients with metastatic endometrial cancer. The C-indexes of the OS and CSS nomograms in the training cohort were 0.749 (95% CI, 0.739-0.760) and 0.746 (95% CI, 0.736-0.756), respectively. The C-indices of OS and CSS in the validation cohort were 0.730 (95% CI, 0.714-0.746) and 0.728 (95% CI, 0.712-0.744), respectively. The ROC curve revealed our model's good prediction accuracy and clinical practicability. The calibration curve also confirmed the consistency between the model and actual existence. The Kaplan-Meier curves revealed statistically significant differences between the risk subgroups (P < 0.05).

Conclusion: Our SEER-based nomograms for predicting survival in patients with metastatic endometrial cancer were helpful for the clinical evaluation of patient prognosis.

Keywords: SEER database; endometrial cancer; metastasis; nomogram; predictive model fund program: natural science foundation of gansu province(17JR5RA242).