Evaluation of a selective screening for colorectal carcinoma - The Taiwan Multicenter Cancer Screening (TAMCAS) project

BACKGROUND. Although the efficacy of mass screening for colorectal carcinom a (CRC) with a fecal occult blood test has been demonstrated in several ran domized trials, a mass screening approach used in countries with intermedia te or low incidence of CRC might be costly. Screening high risk people may be an alternative approach, to aid in the prevention of death from CRC. How ever, the efficacy of CRC screening for high risk people in such countries is uncertain. METHODS. For this study, a multicenter design was devised to identify high risk groups without clinical symptoms related to CRC; these subjects were i dentified through the study of index cases of CRC in Taiwan. Colonoscopy, i n combination with a fecal occult blood test or double-contrast barium enem a, was used to screen high risk groups. A total of 8909 subjects were invit ed to attend screening. Of 8909, 81 with asymptomatic CRC were detected in one-shot screening. Markov models, in conjunction with a simulated approach , were proposed to estimate relevant parameters in relation to disease prog ression and to assess the effect of the interval between screenings on the efficacy of CRC screening for these high risk groups. RESULTS. The estimated preclinical incidence rate was 0.00396 (95% confiden ce interval [CI], 0.002944-0.004985), which was 21 times that reported from a cancer registry in 1994. The simultaneous estimations of mean sojourn ti me (the average duration between the preclinical screen-detectable phase an d the clinical phase) and sensitivity were 2.8 years (95% CI, 2.15-4.30) an d 95.0% (95% CI, 24.4-99.9%), respectively. Predictions of mortality reduct ion for people who received annual, biennial, and triennial screening regim es compared with controls were 26% (95% CI, 0-50%), 23% (95% CI, 0-48%), an d 21% (95% CI, 0-47%), respectively. CONCLUSIONS. The efficacy of selective colorectal carcinoma screening has b een demonstrated in this study. A high preclinical CRC incidence rate also suggests that such a screening strategy might be cost-effective for countri es with intermediate or low incidence of CRC. Methods proposed in this stud y can be used to evaluate the efficacy of CRC screening in similar screenin g trials. Cancer 1999;86:1116-28. (C) 1999 American Cancer Society.