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

Abstract

e18789 Background: Overall survival from cancer has continuously improved over the last thirty years. This is due, in part, to advances in systemic treatment of some cancers resulting in clinically significant increases in long-term survival. Despite these improvements in long-term survival, many patients die quickly after a diagnosis of metastatic cancer, possibly being denied the opportunity to benefit from these recent advances in cancer treatment. To identify the scope and predictors of early mortality in de novo metastatic cancer, we analyzed population-based data, focusing on patients with early mortality. Methods: Using the Survival, Epidemiology and End Results (SEER) Database we analyzed adult patients who presented with de-novo metastatic solid malignancies diagnosed between 2008-2016. We defined early mortality as 1- and 2-month mortality. Factors associated with early mortality were analyzed including primary site of cancer and demographics. Correlations between categorical variables were analyzed with chi squared tests and survival data was derived from life tables using STATA I/C 12.1. Results: Of 519,553 patients identified, 273,976 (53%) were males, 411,041 (79%) were white, 213,325 (41%) had lung cancer, 65,679 (13%) colorectal cancer and 49,139 (10%) pancreatic cancer. One - and two - month cancer specific mortality was 12.4% and 22.5% respectively for the entire cohort. Mortality from cancer in the first two months after diagnosis decreased only marginally between 2008 and 2016 (23.4% and 21.6 %, respectively). Cancers with the highest 2-month cancer specific mortality rates were liver (39.5%), pancreatic (36.3%), lung (26.8%), gastroesophageal (23.7%), and urothelial (21.2%). Cancers with the lowest 2-month mortality rate were breast (10.4%) and prostate cancer (5%). Patients who presented with liver metastases had a clinically significant higher 2- month cancer specific mortality rate (29% vs. 18% p < 0.001), brain metastases were associated with a milder increase in 2-month cancer specific mortality (25%vs 21% p < 0.001) and bone metastases were not associated with a clinically significant difference in 2-month cancer specific mortality (22% vs 21%, p = 0.047). Other factors associated with early mortality included increasing age, low income, and living in a rural area. Conclusions: Over 21% of newly diagnosed metastatic cancer patients die within 2 months of their diagnosis, with little improvement over the last decade, rendering this an urgent unmet need in cancer care. Further research is urgently required to better identify the causative and treatable factors of early mortality in cancer patients.