Development and validation of a prognostic nomogram for predicting cancer-specific survival in patients with metastatic clear cell renal carcinoma: A study based on SEER database

Abstract

Objectives: Clear cell renal cell carcinoma (ccRCC) is highly prevalent, prone to metastasis, and has a poor prognosis after metastasis. Therefore, this study aimed to develop a prognostic model to predict the individualized prognosis of patients with metastatic clear cell renal cell carcinoma (mccRCC).

Patients and methods: Data of 1790 patients with mccRCC, registered from 2010 to 2015, were extracted from the Surveillance, Epidemiology and End Results (SEER) database. The included patients were randomly divided into a training set (n = 1253) and a validation set (n = 537) based on the ratio of 7:3. The univariate and multivariate Cox regression analyses were used to identify the important independent prognostic factors. A nomogram was then constructed to predict cancer specific survival (CSS). The performance of the nomogram was internally validated by using the concordance index (C-index), calibration plots, receiver operating characteristic curves, net reclassification improvement (NRI), integrated discrimination improvement (IDI), and decision curve analysis (DCA). We compared the nomogram with the TNM staging system. Kaplan-Meier survival analysis was applied to validate the application of the risk stratification system.

Results: Diagnostic age, T-stage, N-stage, bone metastases, brain metastases, liver metastases, lung metastases, chemotherapy, radiotherapy, surgery, and histological grade were identified as independent predictors of CSS. The C-index of training and validation sets are 0.707 and 0.650 respectively. In the training set, the AUC of CSS predicted by nomogram in patients with mccRCC at 1-, 3- and 5-years were 0.770, 0.758, and 0.757, respectively. And that in the validation set were 0.717, 0.700, and 0.700 respectively. Calibration plots also showed great prediction accuracy. Compared with the TNM staging system, NRI and IDI results showed that the predictive ability of the nomogram was greatly improved, and DCA showed that patients obtained clinical benefits. The risk stratification system can significantly distinguish the patients with different survival risks.

Conclusion: In this study, we developed and validated a nomogram to predict the CSS rate in patients with mccRCC. It showed consistent reliability and clinical applicability. Nomogram may assist clinicians in evaluating the risk factors of patients and formulating an optimal individualized treatment strategy.

Keywords: Surveillance, Epidemiology, and End Results (SEER); metastatic clear cell renal cell carcinoma; nomogram; prognosis; survival analysis.

Figure 1

Flowchart showing the selection of patients from the SEER database, based on the inclusion and exclusion criteria outlined above; 1,790 patients were included in this study.

Figure 2

Kaplan–Meier curves of cancer specific survival in patients with metastatic clear cell renal cell carcinoma stratified by age (A), sex (B), race (C), laterality (D), T-stage (E), N-stage (F), bone metastases (G), brain metastases (H), liver metastases (I), lung metastases (J), chemotherapy (K), radiotherapy (L), surgery (M) and histological grade (N).

Figure 3

Nomogram model was constructed using the independent prognostic factors predicting the 1-, 3- and 5-year CSS for patients with mccRCC.CSS, Cancer-Specific Survival; mccRCC, metastatic clear cell renal cell carcinoma.

Figure 4

Calibration plots for the nomogram. Calibration plots of 1-year (A), 3-year (B), and 5-year (C) CSS in the training set; Calibration plots of 1-year (D), 3-year (E), and 5-year (F) CSS in the validation set. CSS, Cancer-Specific Survival.

Figure 5

ROC curves of the nomogram for CSS compared with TNM staging. ROC curves comparation of the nomogram and TNM staging for 1-year (A), 3-year (B) and 5-year (C) CSS in the training set. ROC curves comparation of the nomogram and TNM staging for 1-year (D), 3-year (E) and 5-year (F) CSS in the validation set. AUC: area under the curve; ROC, receiver operating characteristic; CSS cancer specific survival.

Figure 6

DCA of the nomogram and AJCC TNM staging for 1-year (A), 3-year (B) and 5-year (C) CSS in training set, and for 1-year (D), 3-year (E) and5-year (F) CSS in the validation set. The red line represents the nomogram. The orange line represents AJCC TNM stage. CSS, cancer specific survival; DCA, decision curve analyses; AJCC, American Joint Committee on Cancer.

Figure 7

Kaplan–Meier survival analyses to test the risk stratification system within the total cohort. The blue line represents low-risk group, and the yellow line represents high-risk group. Low-risk group (score≤ 257); high-risk group (score >257).

Figure 8

The Kaplan-Meier curves of CSS in ccRCC patients according to metastatic status: 1 site versus 2 sites verses ≥3 sites (A), 1 site versus >1 sites (B), 2 sites versus >2 sites (C). The Kaplan-Meier curves of CSS in ccRCC patients according to metastatic status: with single site (D), with two sites (E) , with three sites (F).