Tailored FIT positivity thresholds could optimize colon cancer detection

Last Updated: 2018-09-17

By Will Boggs MD

NEW YORK (Reuters Health) - Different positivity thresholds for quantitative fecal immunochemical tests (FIT) yield different colorectal cancer (CRC) detection rates across age groups and sexes, suggesting the need for thresholds tailored to different individuals.

"Throughout the U.S., quantitative FIT are used as binary tests, most commonly using the threshold of 20 ug/g," Dr. Kevin Selby from Kaiser Permanente Northern California, Oakland, and University of Lausanne, Switzerland told Reuters Health. "I think some consideration should be given to providing patients with their quantitative results. It might allow for more nuanced decision-making. Important information is being lost."

Dr. Selby and colleagues assessed the balance between cancer detection and screening burden at varying FIT positivity thresholds in their retrospective study of more than 640,000 patients who had at least one quantitative FIT result.

Among these patients, 1,245 cancer cases were diagnosed at or within two years of the baseline FIT, according to the September 18th Annals of Internal Medicine report.

At the conventional threshold of 20 ug/g, 7.6% of patients had a positive result, and 76% of these underwent colonoscopy within two years (median, 33 days after receiving the result).

The proportion of cancer cases detected by FIT (i.e., programmatic sensitivity) decreased with higher FIT positivity thresholds, from 987 of 1245 (79.3%) at 10 ug/g to 925 (74.3%) at 20 ug/g and 822 (66.0%) at 30 ug/g.

The number of positive FIT results per cancer case increased substantially at thresholds below 20 ug/g. For example, reducing the threshold from 20 to 15 ug/g would have detected 25 additional cancer cases (a 3% increase) from 11,308 additional colonoscopies (a 23% increase), amounting to 452 colonoscopies per additional cancer case detected.

Programmatic sensitivity and specificity decreased significantly with increasing age. FIT had lower programmatic sensitivity and higher programmatic specificity among women than among men.

"The lower sensitivity in women is especially troubling, given that long-term follow-up of sigmoidoscopy screening suggests that it is less effective in women," Dr. Selby said. "Future modeling studies to inform guidelines should probably be sex specific."

"Despite the differences by sex and age, I still think that FIT is a good screening test across groups, especially if used annually," he said. "Currently, providers are just provided with positive/negative results. So in the immediate future, these results apply more to those organizing screening programs than individual providers, as they are the ones choosing thresholds. But long term, as mentioned, I think that providers should push to get quantitative results for their patients."

"Modifications in positivity thresholds by subgroups might optimize screening program performance, albeit with effects on the number of false-positive test results," the researchers conclude. "Further research is needed to assess the cost-effectiveness of such changes and their practicality across settings with differing resources."

Dr. Stephen P. Helloran from Surrey, UK, who wrote an editorial related to this report, told Reuters Health by email, "(My) editorial takes some of the new data from Kaiser Permanente (KP) and illustrates/highlights how use of data already available to KP could be combined with numeric FIT concentrations (or trends in FIT concentration), then include measures of adherence to screening, last date of screening, family and personal medical history, lifestyle information, if available, all of which can be utilized in a multivariable score of CRC risk as the indication for referral to colonoscopy."

"We already have evidence that the application of artificial neural networks to the 'big data' held by screening programs can refine the accuracy of screening," he said. "Better accuracy brings better targeted colonoscopy referrals, fewer false positives, and more cost-effective screening."

SOURCE: http://bit.ly/2xhr0CL

Ann Intern Med 2018.