Early mortality in metastatic lung cancer: A SEER population data analysis.

Abstract

9103 Background: Millions of dollars invested in improving outcomes for metastatic lung cancer patients are essentially aimed at extending long term survival, with significant benefits being achieved over the last decade. However, little is known about lung cancer patients who die rapidly after diagnosis, potentially being deprived of these advances. We analyzed population-based data to describe real-world outcomes in metastatic lung cancer patients focusing on patients with early mortality. Methods: Using the Survival, Epidemiology and End Results (SEER) Database we analyzed adult metastatic lung cancers diagnosed between 1994-2014. This period was divided into 3 equal time periods: 1994-2000 (TP1), 2001-2007 (TP2) and 2008-2014 (TP3). Early mortality was defined as death within 2 months of diagnosis. Correlations between categorical variables were analyzed with chi squared tests and survival was analyzed using the Kaplan-Meier method. Results: Of 276,527 patients diagnosed with metastatic lung cancer, median age was 67 (range 20-105) and 154,465 (56%) were males. Thirty eight percent (103,830) of all patients died within 2 months of diagnosis. Of these early deaths, 96,344 (92.8%) were due to lung cancer. While the 2-year survival almost doubled from TP1 compared to TP3 (6% vs 11%, p < 0.001), the percentage of patients who died within 2 months only marginally improved (39.7% vs. 36.2% in TP1 vs TP3, respectively). For patients surviving at least 2 months, 2-year survival increased from 10% to 18% in TP1 vs TP3 (p < 0.001). Factors associated with early mortality include age > 65 (45% vs 31%), unmarried status (42% vs 34%), male sex (39% vs 36%), liver metastases (47% vs 32%) and large cell carcinoma vs adenocarcinoma (44% vs 36%) (all p < 0.001). Conclusions: While there has been a steady improvement in the long-term overall survival of patients with metastatic lung cancer, over one third of patients still die within 2 months of diagnosis. This has only marginally improved in the last 20 years. Research is urgently needed to identify causative and treatable factors.