An analysis of the COLOGIC test for colorectal cancer

I recently opened my office mail to find a direct-to-physician informational package from a local laboratory about COLOGIC (also known as CRC-446 or GTA-446). The brochure marketed this test as “A simple blood test to assess the risk of colorectal cancer for 50-75 year olds”. That’s certainly a big claim to make, and game-changing if true, so I decided to delve further into the evidence.

(Spoiler: it’s slick marketing and disappointing statistics, not a life safer. But read on to find out why.)

Colorectal cancer (CRC) screening is a hot-button topic to begin with. Colorectal cancer is the 3rd most common cancer in Canada (excluding non-melanoma skin cancer), the 2nd leading cause of death from cancer in men, and the 3rd leading cause of death from cancer in women. The “early detection” drum has been beaten loudly for colorectal screening, with government organizations employing organized screening programs and frequent advertising including a recent popular ad featuring Canadian comic Neil Crone. Colorectal screening has been shown to be cost effective at less than $30,000 per additional life year gained. A Cochrane review showed a 16% reduction in colorectal mortality with fecal occult blood testing (FOBT), despite the high number needed to screen of 1,176 to prevent one death over 10 years (1). A single FOBT test in isolation has a sensitivity of <20%. Despite erroneous claims from many cancer agencies that the 6-item FOBT kit will increase its sensitivity by 3-fold to nearly 90%, many studies show that the sensitivity remains 24-35% for advanced colonic lesions (specificity around 94%, positive predictive value (PPV) 36-39%). Most disappointing is its poor negative predictive value (NPV) of 87-89% (2,3,4,5). Fecal immunochemical testing (FIT) has a higher sensitivity than FOBT, but is more expensive than FOBT and we don’t yet have evidence of a decrease in mortality compared to FOBT. Fecal DNA testing has an even higher sensitivity, with a similar specificity, but has yet to be widely adopted. Colonoscopy screening for all low-risk patients is not currently recommended (and not likely feasible).

So we are left looking for other highly sensitive, highly specific, and cost-effective means of detecting clinically significant colorectal cancers. This is the niche that the COLOGIC test is desperately trying to fill.

The promotional material describes COLOGIC as a means of “categorizing a patient’s risk of developing colorectal cancer based on their level of GTA-446”, which is a “human-specific, hydroxylated, polyunsaturated, ultra-long chain fatty acid that is pro-apoptotic, anti-proliferative, anti-inflammatory”. It hits all of the buzzwords, but let’s dig a bit deeper.

All of the published studies on the link between GTA-446 and colorectal cancer have been authored by Dr. Shawn Ritchie, a biochemist from Saskatchewan who is the Director of Biomarker Discovery and Validation for Phenomenome, the company that discovered and markets the COLOGIC test. This connection does not itself invalidate any of the data produced to date, but with any studies that have ties to industry, we cannot take their final written conclusions for face value. We must dig into the data.

The first study describing this particular fatty acid pattern and a link to colorectal cancer was published in the open access BMC Medicine in 2012 (6). Dr. Ritchie and his colleagues took the blood samples of 222 patients with colorectal cancer and 220 controls (from 3 larger population data sets), and using mass spectroscopy identified the top 50 metabolites whose patterns were most significant, and from this group identified 13 metabolites that were common between the 3 data sets, which were all hydroxylated polyunsaturated ultra long-chain fatty acids containing between 28 and 36 carbons. The process of identifying a culprit metabolite was not determined a priori, so this entire study was merely hypothesis-generating. Their hypothesis is that low levels of these fatty acids with 28 carbons (which they call CRC-446 or GTA-446, 446 being the molecular weight) represents a compromised ability to protect against accumulating chronic inflammation and abnormal cell growth, which ultimately leads to a pro-cancer environment. There has been no basic science research to back up this specific claim to date. Interestingly, in this study they state that this group of metabolites was not present in the tumor itself, nor was it associated with tumor burden (high rate of association in stage I cancer, less in higher stage cancers).

Their next study in open-access BMC Gastroenterology in 2010 looked at blood samples of patients with CRC pre-treatment, post-treatment, and controls (7). They found that GTA-446 levels did not change post-treatment, again showing that this test is not a marker of tumour burden. They also showed that GTA-446 levels increase with age in healthy patients, but are stably low in CRC patients. They then take a huge leap, and begin to make projections on how many colonoscopies would be saved if we used the GTA-446 test instead of FOBT or FIT testing. Two observational studies, no prospective trials, yet the lead author is making grand pronouncements about the future utility of the test.

Finally the most recent study by Dr. Ritchie’s group from 2013 in the International Journal of Cancer (8). They looked at serum samples and pathology data from 4,923 representative subjects undergoing colonoscopy and from 964 subjects from the general population. (The general population data was useless, since it showed that non-CRC patients have higher GTA-446 levels, which we already knew. They didn’t do a colonoscopy on this general population, so it doesn’t give us any relevant screening data). Of the 4,923 subjects, 98 were found to have colon cancer, of which 84 had a GTA-446 level below their threshold, which is a positive test (not clear whether this threshold was determined a priori). 2,352 of the 4,923 patients had a GTA-446 level below the threshold yet did not have CRC (see Table 3 below, add up the numerators in the lower half of the All column). This gives a sensitivity of 85%, a NPV of 99%, but a horrific specificity of 53% and PPV of 3.4%. Shockingly, they did not include a specificity calculation in their paper. Once again their conclusion declares the utility of GTA-446 in CRC screening, despite the obvious flaws.

cologic1

Now to the fun part. The marketing.

The national laboratory promoting this test included “Post-Market Results” in their marketing material. Of the 14,995 COLOGIC tests ordered on a private pay basis between October 2012 and October 2014, 2233 were positive, of which 419 underwent colonoscopy (it is not clear what happened to the rest). Of the 419, 12 were found to have carcinoma, for a positive predictive value of 2.86%. For comparison, even the lacklustre PSA test has a PPV of 25%.

COLOGIC1

They break down the 12 carcinoma patients further:  8 had no previous FOBT, 1 had a previous positive FOBT, and 1 had a negative FOBT. They declare this as “92% of CRC cases detected by colonoscopy were missed by FOBT because the person did not perform the test (8 cases) or the test was a false negative (3 cases).” Intentionally misleading, and they are conflating issues of patient compliance with the quality of the COLOGIC test. Of course the COLOGIC is going to “pick up” cases missed by FOBT, since nearly half of all patients have a positive COLOGIC and will go on to colonoscopy!

COLOGIC2

The data of the patients who were COLOGIC negative was incomplete, so no legitimate analysis can be done from this data.

COLOGIC3

Here are some other amusing screenshots from their promotional material:

COLOGIC1

Yes, patients would like to take a blood test rather than collect their stool. I’m surprised it’s not higher than 90%. That doesn’t change the effectiveness of the test.

cologic2

I’m not sure where they are getting these numbers from. They compare sensitivity and specificity to FOBT, which is misleading to start (should be compared to gold standard colonoscopy), but from my calculations their sensitivity and specificity calculations are incorrect. If anyone can see an error I’ve made, I’d be happy to look at things again.

cologic3

This is essentially arguing that having millions of patients do a COLOGIC test (and thus hundreds of thousands of colonoscopies) will save lives. It is saving lives just by virtue of increasing the number of colonoscopies, not by any positive attributes of the test.

cologic4

This is the laboratory version of the free drug sample. This card was provided in the package. Get the physician comfortable with ordering the test, and they are likely to order and recommend it more frequently. This test is not cheap, and for that money, we need to be able to justify the test’s value to the patient with the data available to us.

What’s the lesson in all of this? Apart from the lack of utility of the COLOGIC test specifically, clinicians need to be incredibly sceptical with new tests being marketed to them. Scott Gavura from Science-Based Medicine has a great article here about the allergy testing that is being heavily marketed by local laboratories. Take their marketing materials, and if you are considering ordering the test, you owe it to your patients to do at least a little bit of digging into the quality of the evidence being presented. Most of us don’t take all of the claims of pharmaceutical reps at face value, and it’s time we applied the same skepticism to the claims of local labs.

  1. Cochrane systematic review of colorectal cancer screening using the fecal occult blood test (hemoccult): an update. Am J Gastroenterol. 2008 Jun;103(6):1541-9.
  1. Collins, JF, et al., Accuracy fo screening for fecal occult flood on a single stool sample obtained by digital rectal examination:  a comparison with recommended sampling practice.  Ann Intern Med 2005; 142: 81-85
  1. Lieberman, DA, et al., One time screening for colorectal cancer with combined fecal occult-blood testing and examination of the distal colon.  N Engl J Med 2001; 345: 555-60
  1. Greenberg PD, et al., A prospective multicenter evaluation of new fecal occult blood tests in patients undergoing colonoscopy. Am J Gastroenterol 2000; 95:1331-8
  1. Mandel, JS, et al., Reducing mortality from colorectal cancer by screening for fecal occult blood.  Minnesota Colon Cancer Control Study. N Engl J Med 1993; 328: 1365-71
  1. Ritchie, S.A, Ahiahony, P.W.K, et al. Reduced levels of hydroxylated, polyunsaturated ultra long-chain fatty acids in the serum of colorectal cancer patients: implications for early screening and detection. BMC Med. 2010; 8: 13.
  2. Ritchie, S.A, Heath, D, et al. Reduction of novel circulating long-chain fatty acids in colorectal cancer patients is independent of tumor burden and correlates with age. BMC Gastroenterol. 2010 Nov 29;10:140.
  3. Ritchie, S.A, Tonita, J, et al. Low-serum GTA-446 anti-inflammatory fatty acid levels as a new risk factor for colon cancer. Int. J. Cancer. 2013, 132:355-362.
Advertisements

2 thoughts on “An analysis of the COLOGIC test for colorectal cancer

  1. Eddy

    Hi Dr Elia

    I asked my family doctor for a colonoscopy test (I’m over 50) but he recommended a cologic test. He said it is more effective and is a better predictor of cancer. Based on your article, this does not seem to be true. Should I insist on getting the colonoscopy?

    Regards

    Eddy

    Like

    Reply
    1. supermarioelia Post author

      Hi Eddy,

      I would definitely not consider the COLOGIC test a useful test at all. For colon cancer screening, current guidelines suggest either an FOBT or FIT test, or a colonoscopy.

      Like

      Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s