Prediction of overall survival of non-small cell lung cancer with bone metastasis: an analysis of the Surveillance, Epidemiology and End Results (SEER) database

Abstract

Background: The prognosis of non-small cell lung cancer (NSCLC) patients with bone metastasis is extremely repulsive. The aim of this study was to potentially characterize the prevalence, associated factors and to establish a prognostic nomogram to predict the overall survival (OS) of NSCLC patients with bone metastasis.

Methods: The Surveillance, Epidemiology and End Results (SEER) database was used to collected NSCLC cases during 2010-2015. The cases with incomplete clinicopathological information were excluded. Finally, 484 NSCLC patients with bone metastasis were included in the present study and randomly divided into the training (n=340) and validation (n=144) cohorts in a ratio of 7:3 based on R software. NSCLC patients with bone metastasis were selected to investigate predictive factors for OS and cancer-specific survival (CSS) using the multivariable Cox proportional hazards regression. A nomogram incorporating these prognostic factors was developed and evaluated by a concordance index (C-index), calibration plots, and risk group stratifications.

Results: In the Cox proportional hazards model, sex, race, American Joint Committee on Cancer (AJCC) N, T stage, liver metastasis, and chemotherapy were regarded as prognostic factors of OS. The nomogram based on sex, race, AJCC N, T stage, liver metastasis and chemotherapy was developed for cancer-specific death to predict 1-, 3-, and 5-year survival rate with good performance. The C-index of established nomogram was 0.695 for cancer-specific death in the study population with an acceptable calibration.

Conclusions: The female gender, the patients with chemotherapy and not liver metastasis may indicate improved survival. However, the global prospective data with the latest tumor, node, metastasis (TNM) classification is needed to further improve this model.

Keywords: Nomogram; Surveillance, Epidemiology and End Results (SEER) bone metastasis; non-small cell lung cancer (NSCLC); overall survival (OS).

Figure 1

The subject selection algorithm. C34.1, upper lobe, lung; C34.2, middle lobe, lung; C34.3, lower lobe, lung; C34.8, overlapping lesion of lung. NSCLC, non-small cell lung cancer; SEER, Surveillance, Epidemiology and End Results database; AJCC, American Joint Committee on Cancer; TNM, tumor node metastasis.

Figure 2

A nomogram to predict the OS of NSCLC patients with bone metastasis in the training cohort. The total points were calculated by adding the points of each prognostic factor, and correspond to the possibilities of 1-, 3-, and 5-year OS of NSCLC patients with bone metastasis in the training cohort. OS, overall survival; NSCLC, non-small cell lung cancer.

Figure 3

ROC curves. ROC curves for predicting 1-, 3-, and 5-year OS in the training cohort (A-C). ROC curves for predicting 1-, 3-, and 5-year OS in the validation cohort (D-F). AUC, area under the curve; ROC, receiver-operating characteristic; OS, overall survival.

Figure 4

Calibration curves. The calibration curves of the nomogram for the 1-, 3-, and 5-year OS prediction of the validation cohort (A-C), training cohort (D-F). OS, overall survival.

Figure 5

Predicted probability of OS by chemotherapy (A), absence of liver metastasis (B), N stage (C), T stage (D), race (E), sex (F) shown using Kaplan-Meier curve. OS, overall survival.