Natural history of early colorectal cancer - Evolution of a growth curve

PURPOSE: There are very few studies on the development of early colorectal cancers, although we have previously reported growth speeds of early colore ctal cancer in a radiographic retrospective study. The aim of this study wa s to estimate a statistical curve for cancer growth from mucosal cancer. ME THODS: Subjects of the study were 31 patients with cancer in which initial lesions were diagnosed as mucosal cancer. These lesions were overlooked in the first or second investigations, but were detected later. Initial radiog raphic features were as follows; 4 pedunculated lesions, 1 semipedunculated lesion, 6 sessile lesions, 9 superficially elevated Lesions, and 11 superf icially depressed lesions. The diameters of the initial Lesions were 12.1 /- 6.1 mm. The final depths of invasion were 6 mucosal cancers, 12 submucos al cancers, 6 muscularis propria cancers, and 7 serosal cancers. The observ ation period between the initial and final examination was 41.5 +/- 25.8 mo nths. The growth curve was estimated by an exponential curve with the natur al logarithm of d = e((a + b x t)), where a is the intercept (initial tumor size) and b is the regression coefficient (growth speed). RESULTS: A growt h curve was obtained as follows: diameter = 12.5 x 2((t/77)) (r = 0.448, P < 0.0001) and 95 percent confidence interval of rime = 53 to 173 months. Su bsequently, volume = 1 X 10(3) X 2((t/26)) and the 95 percent confidence in terval of time = 18 to 58 months. CONCLUSION: Growth speed of early colorec tal cancer was estimated through a statistically significant growth curve. Estimated doubling time of the volume of early colorectal cancer was 26 (95 percent confidence interval, 18-58) months. From these results we could ob tain a rational cancer surveillance program using appropriate procedures wi th different sensitivities.