statistical analysis
Randomization was performed centrally, and the
minimization method was used to balance treatment
allocation according to the TNM stage (T2 or
T3 vs. T4 and N0, N1, or N2), the presence or absence
of bowel obstruction or tumor perforation,
and the medical center. The sample size of 2200
patients was calculated under the assumptions of a
three-year disease-free survival rate of 73 percent in
the control group and 79 percent in the group given
FL plus oxaliplatin, with a ratio of stage II disease
to stage III disease of 0.4:0.6, an enrollment
period and a follow-up period of three years, a decrease
in the risk of relapse after three years, a statistical
power of 90 percent, and an alpha value of
0.05 and two-sided P values derived with the use of
the log-rank test. The primary efficacy variable was
disease-free survival, defined as the time from randomization
to relapse or death, whichever occurred
first. Second colorectal cancers were considered
relapses, whereas noncolorectal tumors were disregarded
in the analyses.