Robert Greenlee

Marshfield Clinic Research Foundation

PLCO (Learn more about this study)

2006-0113

Jul 1, 2006

The Association of Family History and SES with Advanced/Aggressive Prostate Tumors in PLCO

There is considerable geographic and demographic variation in prostate cancer (PCa) incidence and mortality in the United States; the reasons for this are not well understood (Hsing and Devesa 2001). Accordingly, national prevention research priorities for PCa include identifying factors responsible for the observed variation in occurrence and death rates across people and places (National Cancer Institute 2002). It is clear that effective prevention research efforts must not only include better stratification of general PCa risk in the population, but also a better understanding of factors leading to development of advanced or aggressive disease, which is more strongly linked to fatal outcomes (National Cancer Institute 2000). Two factors of particular interest for risk stratification include family history and measures of socioeconomic status (SES). Greater understanding of the role family history plays in determining disease risk and disease severity may be helpful at the population level in improved targeting for prevention and screening (Yoon et al. 2002). Similarly, there is strong interest for better characterization of the relations between SES and PCa risk, severity, and mortality to improve cancer control (Centers for Disease Control and Prevention 2000; Krieger et al. 2003). PLCO provides a rare opportunity to evaluate the relations of family history of PCa and SES with risk of diagnosis and with presence of accepted indicators of tumor severity at diagnosis. This is due to the large number of well-defined PCa cases, the availability of cases other than just those undergoing prostatectomy, and the inclusion of individual level data for personal and sociodemographic characteristics. Also, unlike recently reported studies, PLCO maintains a reduced influence of ascertainment bias through regimented screening intervals for the intervention arm participants, which will allow identification of factors more strongly affiliated with the inherent risk of advanced/aggressive disease rather than just ascertainment delay.

1. To characterize the extent to which men with a family history of PCa, and men of low SES (education and occupation) are more or less likely to be diagnosed with PCa.
• Men with a family history of PCa will be more likely to be diagnosed with prostate cancer in the overall analysis.
• Men of lower SES will be less likely to be diagnosed with prostate cancer in the overall analysis.

2. To characterize the extent to which men with a family history of PCa, and men of low SES are more or less likely to be diagnosed with advanced stage or aggressive grade tumors, given that they have prostate cancer.
• Cases with a family history of PCa will appear to be unassociated with, or even independently protected against, diagnosis with advanced/aggressive disease in the overall analysis.
• Cases of lower SES will show elevated odds of advanced and aggressive PCa independent of other exposures in the overall analysis.

3. To evaluate how these relations may vary across key strata defined to represent differing levels of influence of ascertainment delay: screening year (t0 vs. t1-t3), screening compliance (fully, partial, not), and timely follow-up (biopsy) to abnormal suspicious screens (yes vs. no).
• Within strata less influenced by ascertainment bias, men with a family history of PCa will still be at greater risk of diagnosis with prostate cancer, but not to the extent observed in the overall analysis.
• Within strata less influenced by ascertainment bias, men of low SES will not be less likely to be diagnosed with prostate cancer, contrary to the overall analysis.
• Cases with a family history of PCa will not be protected against, and may even have elevated odds of advanced or aggressive PCa, within strata less influenced by ascertainment bias.
• The relation between lower SES and advanced/aggressive PCA will be reduced or eliminated within strata less influenced by diagnostic delay.
• Changes in the strength of associations across study strata will be more pronounced for the stage outcome than for Gleason score.