A Randomized Controlled Trial Comparing Participation and Diagnostic Yield in Colonoscopy and CT-Colonography for Population Based Colorectal Cancer Screening

Abstract

Background: Semi-quantitative immunochemical fecal occult blood tests (I-FOBT) are increasingly used for colorectal cancer (CRC) screening. Their true performance characteristics have been determined clinically or sometimes by colonoscopy as “gold standard”. However, occasionally colonoscopy has false-negative results for CRC & Cancer Registry follow-up can provide information relevant to both tests. Aims: Evaluate true performance characteristics of both I-FOBT & colonoscopy examinations for detecting CRC & advanced adenomatous polyps (AAP). Methods: We reanalyzed a published file of ambulatory colonoscopy patients who also prepared 3 I-FOBTs processed by OC-MICRO automated instrument (Eiken, Tokyo, Japan) with ≥50ngHb/mL buffer in any test determining positivity (Rozen P, et al Cancer 2010). These patients were followed through the Israel National Cancer Registry (INCR) to identify newly diagnosed CRC. Analyses of sensitivities for CRC & AAP refer to results of I-FOBTs & initial colonoscopy. Results: 1630 persons performed both IFOBT & total colonoscopy, mean age 62.7 years, SD 11.9 & 50.1% males. They were followed through the INCR for mean 44.3 months SD13.4 & during this period 25 CRC patients were identified, 41% in the proximal colon & 86% were Stages 1 or 2. Follow-up Data Within 12 months both colonoscopy & I-FOBTs identified 21/22 CRCs, 2 initially classified as AAPs; neither test identified a Stage 1 rectal CRC. Between 12-24 months 2 more CRCs were diagnosed, both had elevated I-FOBTs & initially been identified as AAPs. Between 25-48 months a CRC was diagnosed after 47 months, it did not have elevated IFOBT & initially been identified as AAP. Analyses: At 1 year, sensitivity for CRC by both I-FOBT & initial colonoscopy was 95.5% (95%CI 86.8, 104) & at 24 months & 36 months it was 95.8% (95%CI 87.8, 104). Within 48 months sensitivity for CRC by I-FOBT was 92% (95%CI 81.4, 103) & 96% by colonoscopy. At 1 year 121 patients were identified as having AAPs, I-FOBT was positive in 70, & so had a sensitivity of 63.6% (95%CI 55.8, 71.5) for all significant neoplasms, CRC or AAP. At 24 months I-FOBT had a sensitivity of 63.2% (95%CI 55.8, 71.5) for all significant neoplasms. Conclusions: In this population, neither initial colonoscopy nor I-FOBTs identified all CRC patients. I-FOBTs examined at 50ngHb/mL buffer threshold identified almost all CRC diagnosed within 36 months. Colonoscopy identified most significant neoplasms, but only by clinical follow-up & repeated colonoscopy were AAPs reclassified correctly as CRCs. From this small but carefully documented study, it would be appropriate to repeat I-FOBTs after 1 year so as to identify any CRC missed initially by either test.