Nichole Tanner

M.D.

Medical University of South Carolina

Assistant Professor

NLST (Learn more about this study)

NLST-46

Jan 9, 2014

Development of a lung cancer risk score for Veterans

The purpose of this research is to compare the demographics of the elderly, Veterans, and minorities diagnosed with lung cancer to those diagnosed with lung cancer in the NLST. A cohort of patients with lung cancer diagnosed/treated at the VA from October 2002 until January 2012 will be identified from the VA Central Cancer Registry. The demographics, lung cancer specific and all-cause mortality outcomes, from this patient cohort will be compared with the same information from all participants in the NLST. Statistical analysis including t-test and Chi square test will be performed where appropriate to determine crude differences in the two populations. To determine whether there are differences in mortality (cause specific as well as all-cause), we will use Cox regression with time to death (measured in months) as an outcome in SAS PROC PHREG. The dependent variable (time to death) will be defined as time (months) between entry to the cohort and death. Those who did not die by the end of the study (January 2012) will be treated as censored observations.Utilizing published methodology from the PLCOm201213 we will determine if this previously validated lung cancer risk tools in a Veteran population diagnosed with stage 1A lung cancer would have predicted lung cancer in those diagnosed with lung cancer in the cohort. Lastly, Two cohorts of patients ages 65 to 69 and ages 70 to 74 diagnosed with stage 1A lung cancer from January 2000 to September 2010 will be created from the SEER-Medicare database. The demographics (including age, race, smoking history, and comorbidities), lung cancer specific and all-cause mortality outcomes from this patient database will be compared with the same information from all participants in the NLST. Statistical analysis including t-test and Chi square test will be performed where appropriate to determine crude differences in the two populations. To determine whether there are differences in mortality (cause specific as well as all-cause), we will use Cox regression with time to death (measured in months) as an outcome in SAS PROC PHREG. The dependent variable (time to death) will be defined as time (months) between entry to the cohort and death. Those who did not die by the end of the study (Sep 30, 2010) will be treated as censored observations.