Lung Metastases in Newly Diagnosed Esophageal Cancer: A Population-Based Study

Abstract

Background: Esophageal cancer is one of the most common cancer types, with its most common distant metastatic site being the lung. Currently, population-based data regarding the proportion and prognosis of patients with esophageal cancer with lung metastases (ECLM) at the time of diagnosis is insufficient. Therefore, we aimed to determine the proportion of patients with ECLM at diagnosis, as well as to investigate the prognostic parameters of ECLM.

Methods: This population-based observational study obtained data from the Surveillance, Epidemiology, and End Results (SEER) database registered between 2010 and 2016. Multivariable logistic regression was performed to identify predictors of the presence of ECLM at diagnosis. Multivariable Cox regression and competing risk analysis were used to assess prognostic factors in patients with ECLM. Median survival was estimated using Kaplan-Meier curves.

Results: Of 10,965 patients diagnosed with esophageal cancer between 2010 and 2016, 713 (6.50%) presented with initial lung metastasis at diagnosis. Lung metastasis represented 27.15% of all cases with metastatic disease to any distant site. Considering all patients with esophageal cancer, multivariable logistic regression indicated that pathology grade, pathology type, T staging, N staging, race, and number of extrapulmonary metastatic sites were predictive factors for the occurrence of lung metastases at diagnosis. The median survival time of patients with ECLM was 4.0 months. Patients receiving chemotherapy or chemoradiotherapy had the longest median overall survival, 7.0 months. Multivariable Cox regression indicated that age, histology type, T2 staging, number of extrapulmonary metastatic sites, and treatment (chemotherapy, radiotherapy, or chemoradiotherapy) were independent predictors for overall survival (OS). Multivariable competing risk analysis determined that age, number of extrapulmonary metastatic sites, and treatment were independent predictors for esophageal cancer-specific survival (CSS).

Conclusion: The findings of this study may provide important information for the early diagnosis of ECLM, as well as aid physicians in choosing appropriate treatment regimens for these patients.

Keywords: SEER program; esophageal cancer; lung metastases; survival; treatment.

Figure 1

Flowchart of the esophageal cancer patients selection from SEER database.

Figure 2

Identification of the optimal age cut-off points for esophageal cancer patients with lung metastasis. (A) X-tile plots based on age according to esophageal cancer-specific mortality. The plots show χ² log-rank values; the brightest pixel represents the maximum χ² log-rank value. (B) Distribution of patients according to age ranging from 25 to 95 years old. The optimal age cut-off of ECLM patients age is shown as 58 and 74 years old (χ² = 9.65, P < 0.001).

Figure 3

Kaplan–Meier analysis of overall survival among esophageal cancer patients with lung metastasis at diagnosis. (A) overall, (B) stratified by age, (C) stratified by the extent of extrapulmonary metastatic disease, and (D) stratified by type of treatment.