Fecal Immunochemical Testing for Detection of Colorectal Cancers in Patients Undergoing Routine Screening Colonoscopy at a VA Hospital: Preliminary Data

Abstract

Introduction: Fecal Immunochemical Testing (FIT) is a current method approved for colorectal cancers (CRC) screening, according to U.S. Preventive Services Task Force (USPSTF). We wanted to assess the use of FIT for CRC screening to compare with previously published data in non-VA patients. Methods: Patients who were deemed average risk for CRC without any cancer history who presented to our VA's Colonoscopy Prep Clinic prior to CRC screening with colonoscopy were included in our study. 102 of them submitted a FIT prior to undergoing screening colonoscopy. FIT was deemed positive based on quantitative lab standards. Patients were deemed “true positives” if they had a positive FIT as well as positive findings on colonoscopy and/or histopathology (polyp noted ≥1cm and/or histopathology with villous, high grade dysplastic, and/or malignant features). Results: 19 of the 102 patients (18.6%) had positive FIT and positive colonoscopic and/or histopathologic findings (true positives). Sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio, and negative likelihood ratio were 84.2%, 44.6%, 25.7%, 92.5%, 1.52, and 0.35, respectively. Comparing patients with positive colonoscopy/histopathology to those with negative results there was a statistically significant association with PPI use (χ2 = 7.2, p = 7.4x10-3) and polysubstance use (χ2 = 15.4, p = 8.7x10-5) with negative colonoscopy/histopathology. A larger number of polyps per colonoscopy was associated positive result with mean of 4.1 ± 2.7 polyps (95% CI 2.82-5.29) compared to mean of 0.9 ± 1.1 polyps (95% 0.66-1.14) in negative colonoscopy/histopathology patients (t value of 4.92, p < 0.0001, 95% CI 1.81-4.50). Conclusion: Our observed sensitivity of 84.2% is similar to other published FIT sensitivities of 69-86%. However, our observed disease prevalence of positive colonoscopy (18.6%) was higher than disease prevalence in the non-VA population. Specificity was also significantly lower than suggested by previously published data, likely owing to the high number of false positive FIT results in our study population as well as the lack of power of our preliminary data. Further analysis with a higher powered study will be needed to assess the disease prevalence in the VA population as well as to assess quantitative thresholds of FIT in order to maximize positive likelihood ratios in our already complex VA patients.