COLORECTAL CANCER SCREENING WITH REPEATED FECAL IMMUNOCHEMICAL TEST VERSUS SIGMOIDOSCOPY. INCIDENCE AND MORTALITY FOLLOW-UP OF THE ITALIAN MULTICENTER TRIALS

Background: Post-polypectomy colonoscopy surveillance can reduce the burden of colorectal cancer (CRC). However, the effectiveness of using the faecal immunochemical test (FIT) in the interval between surveillance colonoscopies for the detection of missed or rapidly progressing lesions in individuals at above-average risk of CRC, has not yet been adequately investigated. This study assessed the effectiveness of interval FIT for early detection of advanced neoplasia and the diagnostic accuracy of FIT in above-average risk people who undergo surveillance colonoscopy.

Methods: 7311 individuals from South Australia deemed to be at above-average risk of CRC due to a family history of CRC and/or a personal history of precursor lesions, and who had undergone at least 2 colonoscopy procedures were included. Individuals with a genetic syndrome, inflammatory bowel disease or previous CRC were excluded. A 2-sample FIT (OC Sensor, Eiken Chemical Company) was offered at 1-2 yearly intervals between colonoscopies, where a positive FIT result (≥20 µg Hb/g faeces) triggered an early colonoscopy from the scheduled surveillance time. FITs returned within 3 months before a complete colonoscopy were used to determine the diagnostic accuracy of FIT. Competing-risk regression was used to assess the association between interval FIT result and the risk of having advanced precursor lesions or CRC (advanced neoplasia).

Results: 9737 pairs of colonoscopies were included, with 38,424 years of total analysis time at risk. The mean (SD) age of participants was 62.7 y (±9.7) and 51% were female. Cumulative
FIT positivity was 18.1% (1388/7654). A positive FIT brought forward the time to diagnosis of cancer by 30 months (IQR: 22-38) and diagnosis of advanced precursor lesions by 20 months (IQR: 12-39). Individuals who had returned 4 consecutive negative FIT results had a significantly lower incidence of advanced neoplasia compared to individuals who did not complete a FIT (6.4% vs 10.2%, p<0.001) or individuals returning a positive FIT (6.4% vs 14.1%, p<0.001). The incidence of advanced neoplasia was lowest among those with 4 consecutive negative FIT results (subdistribution hazard ratio (SHR), 0.54; p = 0.001) and highest among those who returned 1 positive FIT (SHR, 2.62; p <0.001). Of the 689 FITs returned within 3 months prior to a complete colonoscopy, the test was positive for 4/6 cancers (sensitivity:66.7%) and 42/157 (sensitivity:26.8%) advanced precursor lesions.

Conclusion: Interval FIT enabled earlier detection of advanced neoplasia in individuals at elevated risk of CRC. Multiple rounds of negative FIT are likely to indicate a lower risk of advanced neoplasia compared to those not participating in interval FIT surveillance. The interval FIT is a significant predictor of advanced neoplasia risk and could be used as a tool to personalise surveillance colonoscopy intervals.

Background: Data about comparative effectiveness of screening with faecal immunochemical test (FIT) and with sigmoidoscopy (FS) are limited
Aims: To compare colorectal (CRC) incidence and mortality at 15-year follow-up among subjects enrolled in the FS and FIT arms in the SCORE2 (1) and SCORE 3 (2) trials.
Methods: Between November 1999 and June 2004, average risk men and women, aged 55 to 64, included in a general population sample, were enrolled in the SCORE2 and SCORE3 trials, aimed to comparative the effectiveness of different CRC screening strategies (fig.1). Invitation and examination procedures were the same in the two trials.
Only subjects randomly invited (by a personal invitation letter) to undergo FIT or FS screening are included in this analysis. Subjects with a negative FS were no longer invited, while those in the FIT arm were invited every 2 years, until age 69. The Immudia-HemSp method (Fujirebio Inc.) was used in the first two rounds of the SCORE2 and in the first round of the SCORE3 trial. All samples were processed in a centralized laboratory. The local programs subsequently adopted the OC Sensor method (Eiken co.), using 20 µg/gr. positivity cut-off. FS was performed by gastroenterologists in hospital endoscopy units. Bowel preparation was limited to a single enema self-administered at home 2 hours before the test. Colonoscopy was recommended if ≥1 polyp ≥10 mm, or ≥1 advanced adenoma, or ≥3 adenomas, or CRC was found at FS.
Follow-up was performed though automated record linkage of the trials database with the local population cancer and mortality registries. Time-to-event data were censored at time of CRC diagnosis (incidence), emigration, death, or 15 years, whichever came first. Outcome measures were CRC incidence and mortality. Adjusted hazard ratios (HR) were estimated using Cox proportional hazard models lumping the groups from the two trials. Individual patient data metanalysis is also planned.
Results: Overall, 23896 subjects were randomized to FS and 11236 to FIT screening. Participation in the initial round was 29% and 32% for FIT and 28% and 32% for FS in the SCORE2 and SCORE3 trials. During the 15-year follow-up at least 1 FIT was performed by 48% of subjects in the FIT arm performed at least 1 FIT (average FIT number=3) and by 46% of responders and 19% of non-responders in the FS arm. CRC incidence was reduced by 15% (HR: 0.85; 95%CI: 0.73-0.99) and mortality was reduced by 17% (HR:0.83; 95%CI:0.59-1.17) in the FS as compared to the FIT arm (fig.2).
Conclusions: CRC incidence was significantly reduced among subjects undergoing FS as compared to those undergoing FIT screening; the CRC mortality reduction did not reach the level of statistical significance. Participation in the FIT arms was lower than in many currently ongoing programs.
1) Segnan et al JNCI 2005 97(5)
2) Segnan et al. Gastroenterology 2007 131(7)