Purpose: Fecal immunochemical tests (FITs) have been developed to address analytical problems inherent in the older guaiac-based fecal occult blood tests (g-FOBTs). Our aim was to compare the performance characteristics of one g-FOBT (Hemoccult II) and two FITs (the Hemoccult ICT and MagStream HemSp) relative to colonoscopy for the detection of colorectal cancer and significant precursor lesions. We also examined whether a 1-day collection strategy would negatively impact test diagnostic performance.
Methods: We used a prospective observational cohort design in a Canadian population eligible for screening. All participants received colonoscopy after performing the occult blood tests.
Results: One thousand seventy-five individuals were enrolled (mean age 56.3 years, 53.8 % females). Using colonoscopy as the gold standard, the sensitivity for screen-relevant neoplasm was determined for Hemoccult II (7.2, 95 % CI: 1.1-13.4), Hemoccult ICT (23.2 %: 13.2-33.1), and MagStream HemSp using 67 μg/gram stool as the cut-off (23.2 %: 13.2-33.1). The Magstream HemSp, using a cut-off threshold of 30 μg/gram stool, had the lowest specificity at 87.6 % (85.4-89.6), while the Hemoccult II had the highest at 98.8 % (98.1-99.5). Single-day stool testing reduced the false-positive rates of all tests without significantly reducing the sensitivity.
Conclusion: We found that FITs have a significantly increased sensitivity but reduced specificity for screen-relevant neoplasm compared to g-FOBT using colonoscopy as the gold standard. Optimal threshold levels for hemoglobin detection depend on the desired trade off between sensitivity and false-positive rate. Single-day testing with an FIT may be an option to enhance population compliance with screening.