Objectives: Stool testing is a well established method of screening for colorectal neoplasia. Emerging data suggest that novel biomarkers may offer performance advantages over fecal occult blood. In this large, prospective study, we assessed fecal calprotectin (a leukocyte-derived protein) as a screening biomarker for colorectal neoplasia. Fecal calprotectin was directly compared to fecal hemoglobin (Hb) and colonoscopy as the existing criterion standards for stool screening and structural evaluation, respectively.
Methods: Subjects included colonoscopy patients with a personal history of colorectal neoplasia, family history of colorectal cancer, or iron deficiency anemia. Stool specimens were collected before purgation, processed appropriately, and quantitatively analyzed for calprotectin (Nycomed Pharma, Oslo, Norway) and for Hb (Mayo Medical Laboratories, Rochester, MN) by masked technicians. Colonoscopies were performed by experienced endoscopists without prior knowledge of the fecal assay results.
Results: Among 412 subjects, 97 (24%) subjects had one or more colorectal neoplasms (including three with adenocarcinomas). Fecal calprotectin levels did not differ significantly between subjects with versus subjects without colorectal neoplasms (p = 0.33). Neither tumor number (p = 0.85) nor tumor size (p = 0.86) significantly influenced the observed fecal calprotectin concentrations. Estimates of the sensitivity, specificity, and positive and negative predictive values of fecal calprotectin for any colorectal neoplasms were 37%, 63%, 23%, and 76%, respectively. Comparable performance estimates for fecal Hb were 3%, 97%, 27%, and 77%, respectively.
Conclusions: In this cohort of colonoscopy patients at above average risk, fecal calprotectin was a poor screening biomarker for colorectal neoplasia. Further investigation of tumor-derived, rather than blood-based, biomarkers may be a more rewarding approach to stool screening for colorectal neoplasia.