Tailored thresholds for FIT may increase colorectal cancer detection

Using age and sex thresholds to determine positive results on fecal immunochemical testing (FIT) may increase rates of colorectal cancer detection, a new study indicated.

Researchers performed a retrospective cohort study to determine optimal FIT quantitative thresholds for hemoglobin positivity, as well as FIT performance characteristics, since stool hemoglobin concentrations can vary according to demographics. Patients in Kaiser Permanente Northern and Southern California who were ages 50 to 75 years, who were eligible for screening, and who had baseline quantitative FIT results and two years of follow-up were included in the study.

The study's outcome measures were FIT programmatic sensitivity for colorectal cancer and the number of positive test results per case of cancer detected, overall and according to age and sex. Programmatic sensitivity was defined as the proportion of patients with a diagnosis of colorectal cancer within two years of FIT screening whose quantitative FIT result exceeded a given positivity threshold at baseline or during follow-up testing. Study results were published by Annals of Internal Medicine on Sept. 18.

Overall, 640,859 patients completed FIT at baseline and were followed for two years, and of these, 481,817 (75%) had had at least one additional FIT. A total of 1,245 patients (0.19%) received a diagnosis of colorectal cancer. When lower positivity thresholds for FIT were used, rates of cancer detection among this group increased (66.0% at 30 μg/g, 74.3% at 20 μg/g, and 79.3% at 10 μg/g). The authors determined that 3% more cancer cases would be detected and 23% more colonoscopies would be required with a decrease in positivity threshold from 20 μg/g to 15 μg/g. At 20 μg/g, the conventional threshold for a positive FIT result, programmatic sensitivity decreased with age (79.0% for ages 50 to 59 years, 73.4% for ages 60 to 69 years, and 68.9% for ages 70 to 75 years, respectively; P=0.009). Programmatic sensitivity with the conventional threshold was 77.0% in men and 70.6% in women (P=0.011).

The authors noted that quantitative FIT values were not available for all patients who were screened with FIT during the study period, that false-negative results for advanced adenomas could not be measured, and that most patients had previously been screened with FIT, among other limitations.

However, they concluded that in their patient cohort, rates of cancer detection and the number of positive test results per cancer case detected varied significantly by age and sex, “suggesting that modifications in positivity thresholds by subgroups might optimize screening program performance, albeit with effects on the number of false-positive test results.” They called for additional research on the cost-effectiveness of these potential changes, as well as their use in different settings with varying resources.

An accompanying editorial agreed that more research is needed and said a larger study could determine thresholds for individual ages by sex. Also, a longer study of sequential screening could help show whether risk assessment can be improved by trends in FIT concentration, the editorialist said. The editorialist also pointed out that screening models could incorporate medical and family history of colorectal cancer in addition to age, sex, and other variables.

Although these data and those from other studies support multivariate FIT-based screening for colorectal cancer, the editorialist wrote, such an approach would require a clear screening objective. “Maximizing the [colorectal cancer] detection rate within the context of a fixed colonoscopy resource is a practical objective, but where do [colorectal cancer] prevention by adenoma detection, quality-adjusted life-years, and cost-effectiveness fit within the screening objective?” the editorialist wrote.