2.1. Patient Selection from SEER database

The Surveillance, Epidemiology, and End Results (SEER) Program of the United States National Cancer Institute is an authoritative source which collects patient demographic information, cancer diagnostic information, and outcomes from 18 cancer registries in the United States, thus including approximately 28% of the US population. In this study, the first cohort we used was from the SEER database. The SEER database does not contain any identifiers and is publicly available for the studies of cancer-based epidemiology. In the present study, the National Cancer Institute's SEER-Stat software (version 8.3.5) was used to get access to SEER database.

As shown in Figure ​Figure11, 40968 patients, diagnosed with stage IIA (T3N0M0) colon cancer between January 1, 2004 and December 31, 2010, were identified. Patients with unknown tumor grade were excluded from the study. Other exclusion criteria were: unknown race; unknown tumor location or appendix; non-adenocarcinoma histology and not active follow-up.

Schematic representation of patient population selected from SEER database.

In this study, we stratified the “patient had chemotherapy” as ACT group and “no evidence of chemotherapy was found in the medical records examined” as a non-ACT group in the variable “chemotherapy recode” in SEER cohort. The primary endpoint in SEER cohort was cause-specific survival (CSS). The cause of death was categorized as colon cancer specific or non-colon cancer related. The CSS was calculated from the date of diagnosis to the date of colon cancer death. Patients who died of other causes were censored at the date of death. Well, moderately, poorly, and undifferentiated tumor grade accounted for 7.9% (n=3035), 74.9% (n=28756), 15.8% (n=6080) and 1.3% (n=513) of the whole cohort, respectively. Then we combined well and moderately tumor grade which were traditionally associated with better prognosis as the same group, poorly and undifferentiated tumor grade as the other group.