Cutoff value determines the performance of a semi-quantitative immunochemical faecal occult blood test in a colorectal cancer screening programme

Compared with the guaiac-based FOBT (G-FOBT), one of the main advantages of some immunochemical faecal occult blood tests (iFOBTs) is that they allow haemoglobin quantification (Itoh et al, 1996; Castiglione et al, 2002; Guittet et al, 2007). The semi-quantitative nature of these tests permits adjustment of the cutoff value for the detection of colorectal cancer (CRC) in an effort to optimise screening programmes for specific populations and health-care practices. Changing the cutoff value can have considerable implications on the performance of the test in a screening population. In general, lowering the cutoff value will increase sensitivity, but consequently decrease specificity and vice versa. An increase in sensitivity means an increase in the detection of patients with colorectal cancer or advanced adenomas, but the consequential decrease in specificity results in more persons without relevant lesions undergoing a colonoscopy (false positives). Some studies in screening populations have been published on changing the cutoff value of iFOBTs. However, in these studies only a few selected cutoff values are presented, and the complete range of possible cutoff values is not addressed (Castiglione et al, 2002; Guittet et al, 2007; Grazzini et al, 2009). Furthermore, colonoscopy data, verifying the presence or absence of pathology, are usually presented for test results equal to or above the threshold that is recommended by the manufacturer. The two most frequently presented quantitative iFOBTs, the OC-Sensor (Eiken Chemical) and the Magstream 1000 (Fujirebio Diagnostics, Tokyo, Japan), were developed in Japan, where incidence rates for CRC are lower than those in Europe (Minami et al, 2006). Therefore, the cutoff value with optimal overall performance may be different in Europe compared with Japan. In a recent study including 1000 symptomatic and other high-risk patients in Israel, cutoff values below the recommended threshold of 100 ng ml−1 were evaluated (Levi et al, 2007). The authors concluded that the optimal cutoff value might be as low as 75 ng ml−1; they also noted that the test performance in average-risk patients in a screening population is unknown. Our aim was to evaluate the performance and efficiency of a semi-quantitative iFOBT in an average-risk screening population.