Lung cancer survival in Kentucky: a multifactorial approach

The incidence and mortality rates oflung are cancer are higher in Kentucky than in any other state in the country. This disease is especially fatal, with very low rates of survival at only two years after diagnosis. Although survival is largely determined by the stage of disease at diagnosis, with Stage I and Stage II survival rates being much higher, there are other factors that can also influence survival. Risk factors and treatment modalities, for example, have been shown to affect survival among lung cancer patients. Because early stage lung cancers might become more common due to improved early detection, these other factors could become even more important in future attempts to improve lung cancer survival. In order to investigate factors related to early stage survival, we will analyze existing data on 5000 cancer cases from the Kentucky Cancer Registry (KCR). KCR data from 1998-2003 will supply data on early stage lung cancer cases for a retrospective study. In addition, in collaboration with the Kentucky Lung Cancer Research Program's Clinical Trials Network (CTN), we will collect risk factor data from 200 newly diagnosed early stage lung cancer patients. Questionnaires will be administered to patients at seven or more CTN sites representing multiple regions of Kentucky. The questions will address smoking history and habits, family history oflung cancer, occupational history, and potential exposures to lung carcinogens at baseline and over a six month period. The questionnaire data will then be linked to KCR data to create a more comprehensive database for this subgroup of recently diagnosed cases. This database will be used to conduct a holistic analysis of demographic factors, risk factors, treatment modalities, and other variables at both the individual and institutional levels, and their effects on early stage lung cancer survival. Analytical techniques will include Cox proportional hazards regression, Kaplan-Meier plots, and other statistical tools and models.