James Allison, M.D., is Professor Emeritus of medicine in the gastroenterology department at the University of California, San Francisco, and a leading expert on screening tests for colorectal cancer (CRC). We recently spoke with him about the fecal immunochemical test (FIT), which screens for human blood in the stool.
Why is FIT a good CRC screening test?
There are many reasons. Here are a few:
- FIT is a simple stool test that can be easily done at home.
- With FIT, preparing and sending one stool sample is sufficient. Older stool (guaiac) tests required 3 samples, and the sampling process was not appealing to most patients.
- It is cheap and covered by all insurance.
- It does not require an invasive procedure unless it is positive, in which case a colonoscopy is recommended.
- Compared with older stool tests, FIT is much more accurate in detecting cancer and polyps that are most likely to become cancer. It is more accurate because unlike the older tests, it identifies only human blood that may be in the stool, and it won’t give false-positive results due to food or medication. It detects only blood originating in the colon or rectum. The older tests can produce false-positive and false-negative results, depending on what people eat before testing. Red meat, such as beef or lamb, and foods such as cabbage, radishes, and fruits (melon) can result in a false-positive result, while vitamin C can result in a false-negative.
- The processing and reading of the test can be automated for quality assurance. Studies have shown that the human factor involved in older stool tests can lead to false-positive and false-negative results.
Is the FIT test always analyzed by automation?
No. Automated processing is done in large health-care organizations like Kaiser Permanente and the VA and in many other countries with CRC screening programs. In the U.S., stool tests are frequently developed and interpreted by a physician or his/her staff or sent to a commercial lab. A lot of published research shows that there's a wide variation in the interpretation of these results when the test is analyzed by minimally trained personnel. The most accurate results are those evaluated in laboratories either by automation or by well-trained certified personnel.
Why is FIT underutilized as a screening test?
A very important factor has been the message that colonoscopy is the “best” or “gold standard” CRC screening test. In 2000, two studies on population screening with colonoscopy and an editorial were published in the New England Journal of Medicine. Shortly thereafter, gastroenterology societies, CRC advocacy groups, and the media started discouraging eligible people from selecting screening tests other than colonoscopy and were told these were inferior. In 2008, the American Cancer Society, the U.S. Multi-Society Task Force, and the American College of Radiology stated that noninvasive tests, including FIT, were less likely to prevent cancer compared with invasive tests, like colonoscopy. This advice was not based on published evidence and has yet to be proven by “gold standard,” randomized, controlled clinical studies.
Since 2012, it has become clear and more universally accepted that there isn’t one “best” CRC screening test. The message being promoted is no longer “get a colonoscopy,” but “get screened for colorectal cancer via any of the guideline-recommended tests.” According to the American Cancer Society, American College of Physicians, Centers for Disease Control and Prevention (CDC), National Colorectal Cancer Roundtable, American Medical Association, and many recently published reviews of FIT and CRC screening guidelines, annual FIT tests are believed to decrease the incidence and mortality from CRC because randomized controlled trials using less sensitive guaiac fecal occult blood tests have shown these results. In modeling studies, annual FIT testing has been shown to be equally effective as colonoscopy every 10 years.
If I went to my doctor wanting information about FIT, what should I ask?
Questions such as these: Is the FIT you recommend among those recommended by latest U.S. Preventive Services Task Force’s CRC Screening Guidelines? These guidelines recommend the OC-Light and the OC-Auto FIT tests. The CDC uses these guidelines for their FIT recommendations to state Departments of Public Health. Other recommended FIT tests may not have proof of their performance characteristics or quality of their development and interpretation in large average-risk populations.
Also ask where can you obtain the test. Ask your doctor for a prescription for the test and to suggest where you can get it, whether at a lab or a drugstore. Mail in the test for interpretation or development to your health plan’s lab or the lab recommended by your physician. Be certain that the test results are sent to your primary-care provider and to you. If the results are positive, you must have a colonoscopy.
What about over-the-counter (OTC) FIT stool tests?
I don’t approve of nor recommend OTC FIT tests. Though they have been “FDA-cleared,” FIT tests sold over the counter in drugstores or online have not been adequately evaluated for quality and efficacy in large average-risk populations. The recommended FIT tests that have the best test performance characteristics are not sold directly to consumers.
One problem with FIT seems to be that it should be done every year. How will we make sure that people keep doing it? Any practical suggestions?
If you go to an organized health care center, like Kaiser Permanente or the VA, they have methods in place for reminding patients and their primary care providers of the need for an annual FIT. For example, every time you go in for an appointment at Kaiser, you get a sheet with the screenings you are due for, including FIT. Many physician offices and federally qualified health centers have hardware and software in place to do the same thing. Another way to ensure annual testing is to ask your doctor for a FIT test when you come in for your annual flu shot.
Many readers have undoubtedly seen the TV commercials for Cologuard—another CRC screening test. What’s your opinion?
I do not recommend the Cologuard test for average-risk screening program participants. Cologuard is a stool test that combines a FIT with a test for DNA markers associated with cancers. A recommendation for its use came from only one comparative study, but proof that it is better than FIT is lacking. The FDA and Medicare have cleared the test to be repeated every three years, though there’s no published research proving that this is a safe screening interval.
The only published study on Cologuard screening for a large average-risk population detected more cancers and advanced polyps than one-time use of a FIT. The comparison was inappropriate, however, since it’s recommended that FIT be done annually, but the study didn’t present results comparing the results of three annual screenings with FIT to those of a Cologuard test done every three years. A longer-term study comparing Cologuard every three years to FIT screening repeated annually would have more real-world relevance. In addition, Cologuard produced more than twice as many false-positives as FIT in the published study, and 6 percent of people undergoing Cologuard testing submitted stool samples that could not be analyzed (compared to just 0.3 percent undergoing FIT testing). Most likely this was because the stool collection procedure for Cologuard is more complicated than for FIT and harder to do correctly.
Finally, Cologuard costs at least 20 to 30 times more than FIT—as much as $650. A recent study revealed that Cologuard was not cost-effective when compared to screening for CRC with FIT or colonoscopy.
Why isn't sigmoidoscopy used much in the U.S.?
You won’t be surprised, but it’s a five-letter word: money. Flexible sigmoidoscopy used to be the screening test of choice—for instance, at Kaiser Permanente from 1993 until 2005. Sigmoidoscopy examines only the lower part of the colon and is usually done every five years. It costs only a fraction as much as colonoscopy, is quicker, and can be done well by trained primary-care doctors and nurses. The prep is simpler, and sedation is usually not needed. In 2000, however, the editorial in the New England Journal of Medicine that I mentioned above made the erroneous comparison that having a sigmoidoscopy was like screening for breast cancer by performing mammography on one breast. In the early part of the new millennium, Congress cut reimbursements for sigmoidoscopy below the level where private physicians could afford performing them in their offices. That was the end of sigmoidoscopy. Then in 2001, after Congress was lobbied by gastroenterology societies and many lay organizations promoting CRC screening, it mandated that colonoscopy be a covered screening test by Medicare.
How would you yourself get screened for colon cancer?
My preference shouldn’t really influence anyone else’s choice. I know the best doctors doing colonoscopies in my area, those who will give me the best results and with minimal risk. I am educated about my risk and how it relates to my age, family history, and previous screenings. I can choose whatever test I want, and in fact I’ve had them all: sigmoidoscopy, FIT, and colonoscopy. I’m not having any more colonoscopies. I’m just doing FIT until such time as respected guidelines say my age makes it inadvisable for me to get further screening.
The CRC screening tests with the best record and the most data behind them are FIT and colonoscopy. What is important is that you should be screened. Still, the best may not be “the one that gets done”—as is often said—if it costs $650 (like Cologuard) and if you can get an equally good test (like FIT) that costs as little as $5. I realize that for most people, insurance pays for screening, but it still matters.
A FIT-based screening program is the most cost-effective. If you want to do a colonoscopy and your insurance covers it, then there's nothing wrong with colonoscopy (except for the very unlikely risk of perforation of the colon). But screening with colonoscopy has a problem: It has been shown to decrease mortality from cancers in the left side of the colon by about 80 percent but only by about 50 percent in the right side. This may be for several reasons, including that tumors on the left side are genetically and biologically different than those on the right side; they are more difficult to detect by colonoscopy; and a large percentage of them grow faster than those without these genetic and biologic differences. The idea that a colonoscopy done every 10 years is safe may not be true if you are unlucky enough to develop a fast-growing right-sided colon cancer.
Bottom line: The test you select for screening is an individual choice. As a general rule, I encourage my average-risk patients to go from least invasive to most invasive screening tests.
What do you think the future holds for CRC screening?
The Holy Grail would be a blood test for CRC. This has been the hope and goal of investigators for at least three decades. I’ve talked to people who know a lot more about molecular biology than I, and despite some advances, they feel that a sensitive and specific blood test for CRC and advanced polyps is unlikely any time soon. Currently researchers are investigating algorithms (formulae) to measure CRC risk and using the results to decide which average-risk screenees should have a colonoscopy. They input the results from FIT as well as age, gender, family history, and any possibly related symptoms. When all these are considered, the ability to predict whether anything important would be found on a colonoscopy is high.
Also see Virtual Colonoscopy: The Inside Scoop.