Adding Family History of Colorectal Cancer to the FIT-based Screening Program in a Dutch Colorectal Cancer Population Sample

This is a prospective CRC population-based screening trial, using FIT and a family history questionnaire, inviting those with a positive FIT, a familial CRC syndrome, or both for colonoscopy.

All invitees receive an invitation to complete a FIT (FOB-Gold) and a validated, online family history questionnaire. Answers from the questionnaire are compared with the Dutch criteria for referral for genetic testing and/or surveillance colonoscopies for persons at a potential familial risk for CRC. Invitees are invited to perform both tests, but if they only perform one this will be assessed. Participants with a positive FIT (cut-off value 275 ng/ml) and/or a positive family history and a diagnosis of familial CRC by a clinical geneticist will be referred for colonoscopy.

The investigators will compare the diagnostic yield, participation rate, and positive predictive value for two strategies: referral for colonoscopy based on FIT and family history versus referral for colonoscopy based on FIT-only.

The sample size is justified by the anticipated gain in diagnostic yield from combining FIT-only with family history screening. With a sample size of 6.000 persons the investigators expect to detect 130 persons with advanced neoplasia after a positive FIT (21.7 per 1.000 invitees). By adding the family history questionnaire, the investigators anticipate detecting an additional number of 11 persons with advanced neoplasia, resulting in a yield of combined FIT-based and family history screening of 141 persons with advanced neoplasia (23.5 per 1.000 invitees), which is an increase of 1.8 per 1.000 invitees. Using the McNemar test with the significance level set at 5%, there is a power of 91% to show that the gain in diagnostic yield is statistically significant. Previous CRC test screening studies using FIT showed that participation rates are approximately 60% in a first screening round. The participation rate in the Dutch national CRC screening in 2014 (the first year of introduction) was even higher, 68%. The investigators expect that 80% of these persons will fill out the familial risk questionnaire, resulting in 3.264 persons who perform both tests. As the investigators recommend invitees to perform the questionnaire in addition to FIT, the investigators assume that the percentage of persons only filling out the questionnaire and not performing FIT are negligible. Based on previous screening studies using a FIT positivity cut-off value of 50 ng Hb/mL for the OC-Sensor, it is assumed that 8% of these participants have a positive FIT result. In the current Dutch screening program, a different FIT is used (FOB-Gold) with a cut-off value of 275 ng Hb/ml. The FIT positivity rate with this test and with this cut-off value is unknown. However, a governmental announcement recently reported that the positivity rate is expected to remain 8% with this cut-off value. Therefore, the investigators expect 326 FIT positives in this study (8% of 4.080). Based on previous studies, approximately 40% of all persons with a positive FIT result will have advanced neoplasia. In our sample this concerns 130 persons. The percentage of persons with a familial CRC risk requiring surveillance colonoscopies in a screening population varies between 2.3% and 4%. Differences in the definition of an increased familial CRC risk could explain the variability in reported numbers. As the investigators are using a relatively large range in criteria for fulfilling the definition of a familial CRC risk, the highest reported percentage (4%) is assumed, which concerns 131 persons. The previously developed family history questionnaire has a sensitivity of 90% and 100% in identifying those persons qualifying for referral, in two subsequent validation phases respectively. For the purpose of this study, this validated questionnaire is adjusted to the recently renewed referral criteria. The investigators make the assumption that the questionnaire can identify all persons with a familial CRC syndrome based on referral criteria. Persons with a familial CRC syndrome who do not fulfill these criteria will be missed. This concerns an unknown number and will not be taken into account. A previous screening study showed that 6% of those with a positive FIT have an increased familial CRC risk. One can calculate that 3.8% of persons with a negative FIT have a familial CRC risk: 4% minus 6% of 8% (FIT positives), divided by 92% (FIT negatives). If one assumes that there is no difference in familial CRC risk between those who do and those who do not fill out the questionnaire, this involves 114 of 3.264 persons. Based on unpublished data by this research group it is estimated that 10% of those with a negative FIT have a familial CRC risk: 11 persons (10% of 114). A similar percentage was found in another screening study. Additionally, all participants with a familial CRC syndrome will be offered surveillance recommendations. In this sample, 131 persons are estimated to have a familial CRC syndrome. Of these, 16 persons (12%) will undergo a colonoscopy due to a positive FIT (6% of 8% FIT positives) and their familial CRC risk could be addressed at their visit to a colonoscopy center. The other 115 persons (88%) are expected to have a negative FIT and these persons (and their family members) would otherwise be unlikely to be identified as having a familial CRC syndrome.