Abstract
From April 2004 to December 2005, we performed a total of 111,510 colorectal cancer screening tests for residents in Miyagi prefecture. Personal information about the examinees was collated with the Miyagi Cancer Registry database and confirmed that 346 colorectal cancers had been diagnosed within one year of the two-day immunological fecal occult blood test (Magstream AS®; Cut-off point was usually set at 40 ng/ml). Any cancers diagnosed within one year of a negative screen or after a positive screen where further investigation was negative or was refused, were classified as interval cancers. Of 346 colorectal cancers, 304 were detected by screening and 42 were classified as interval cancers (36 had negative screening results, 4 refused further investigation, 2 were negative on further investigation). Program sensitivity was 87.9% for all cancers and 86.2% for all but those with intramucosal cancer. Sensitivity was lower for the older group (age 70 years or older) and those with proximal colon cancer, but there was no apparent difference between males and females. The interval group had a higher proportion of cases with metastatic lesions than the screen-detected group. The survival rate of the interval group was significantly lower than that of the screen-detected group, but equal to that of the cancers diagnosed outside of the screening program. Based on the balance between the decreased number of interval cancers and the increased number of diagnostic examinations, altering the cut-off point from 40 ng/ml to 30 ng/ml (a positivity rate of 5%) was considered to be advisable. Annual screening is effective for reducing interval cancers and interventions to promote continuous participation are required. For the valid comparison among various areas or facilities, it will be necessary to unify the operational definition and the analytical methods of interval cancer.
