Purpose: The progression rates of colorectal cancer by Dukes' stage in a high-risk group were estimated and applied to evaluate the efficacy of different screening regimens.
Patients and methods: Of 6303 high-risk subjects invited to a colorectal cancer screening project with colonoscopy, 39 screen-detected cases and 16 postscreening cases were diagnosed with information available on Dukes' stage. A five-state Markov process was applied to estimate parameters pertaining to the disease natural history of colorectal cancer by Dukes' stage.
Results: The estimates of the mean sojourn time in years were 3.10 for preclinical Dukes' A and B and 1.92 for preclinical Dukes' stages C and D. The predicted reductions of Dukes' stages C and D achieved by annual, biennial, 3-yearly, and 6-yearly screening regimens against the control group were 60%, 49%, 40%, and 25%, respectively. These, in turn, yield the corresponding predicted mortality reductions of 39%, 33%, 28%, and 18%.
Conclusions: These findings suggest that to achieve a 30% mortality reduction, as observed in annual fecal occult blood testing, a prudent interscreening interval with colonoscopy for this high-risk group should not be longer than 3 years.