Contemporary changes in localized and metastatic prostate cancer incidence by geographic area following decreased PSA screening.

Abstract

1567 Background: In the setting of decreased PSA screening, the incidence of metastatic prostate cancer has been increasing in the United States. This was chronologically proceeded by decreasing localized prostate cancer incidence. While decreased detection of localized disease is hypothesized to increase likelihood of metastatic disease at diagnosis, it is unclear whether the two are geographically connected. Methods: Prostate cancer incidence was obtained from the of Surveillance, Epidemiology, and End Results (SEER) database for men 70 years or older. SEER Summary Stage 2000 was used to classify localized (local) and metastatic (distant) prostate cancers. Changes in incidence rates were calculated by health services areas (HSA), which each represents a relatively self-contained region of hospital care. We chose a priori to examine most recent years 2012-2015 for changes in metastatic disease, and proceeding years 2008-2011 for changes in localized disease. Population-weighted linear regression that was robust to outliers was performed. Results: A total of over 66,600 cases of localized and 6,400 cases of metastatic prostate cancer from 200 HSAs were included for analysis. From 2008 to 2011, localized incidence decreased from 613.6 to 534.2 per 100,000 men overall, and for each HSA on average decreased by 30.3 per 100,000 men for each year. From 2012 to 2015, metastatic incidence increased from 54.7 to 62.1 per 100,000 men overall, and for each HSA on average increased by 2.1 per 100,000 men for each year. Linear regression between HSA-level changes in localized and metastatic disease revealed a correlation coefficient of -0.023 (SE = 0.017, p = 0.16, 95% CI -0.056 to 0.009), representing lack of a statistically significant relationship between decreases in localized disease and later increases in metastatic disease within each health services region. Conclusions: Despite concerns of increasing metastatic prostate cancer incidence coinciding with decreases in PSA screening and localized cancer incidence, we do not observe a statistically significant geographic and temporal relationship between metastatic and localized disease at the HSA level. Our study is limited by short lead time and thus this trend warrants continued surveillance.