We will conclude our five part series on Exact Sciences (EXAS) (or "the Company") with an expose on the Company's history of product failures and the ways in which the Company has misled investors. In response to several criticisms of the analysis presented in this series, we may also write an appendix later this week to give our perspective on these points of contention. We continue to believe Exact Sciences is a compelling short and is worth under $3 per share.
Company History: This Is Not The First Rodeo For Exact Sciences
Investors may not be aware that Exact Sciences has a history of failures with other stool based, non-invasive, colorectal cancer screening tests. Cologuard is actually the third iteration of Exact Sciences' failed stool DNA testing kits. Neither of its previous products, PreGen-Plus and ColoSure, generated significant revenues and both failed to receive FDA approval as well as CMS reimbursement. We believe the third time is not the charm, and will instead be their last strike. Simply put, Exact Sciences is in the business of raising money and marketing hope, and not in the business of bringing valuable cancer screening tools to market. We are getting to watch history repeat itself right before our eyes.
Failure 1: PreGen-Plus 2003-2006: Only 12,500 Tests Sold And $559,000 In Test Royalty Revenues
This is a picture of Exact's first failed test, PreGen-Plus. It looks surprisingly like Cologuard.
Exact Sciences commercialized its first product, PreGen-Plus, in 2002 in collaboration with Laboratory Corp. of America Holdings (LH) (or "LabCorp"). As seen here in this fact sheet, Exact touted PreGen-Plus as a "breakthrough colorectal cancer screening option" and highlighted that it was "four times more sensitive than the fecal occult blood test, currently the only other non-invasive screening method for colorectal cancer." Like Cologuard, PreGen-Plus was a take-home test with a similarly high price point of $350.
When PreGen-Plus was released, sell-side analysts were uniformly bullish, much as they are today. According to analysts at the time, Exact Sciences had come up with a test that not only was statistically superior to other existing screening tools at the time, but would also be the solution to low patient compliance in colon cancer screening. PreGen-Plus adoption was expected to be strong. Merrill Lynch, which initiated with a Buy rating and a $19 price target and later upgraded to a Strong Buy, stated "EXACT's genomics-based test 1) improves sensitivity/specificity, and 2) greatly enhances ease of use in order to increase early detection and bolster compliance." Merrill Lynch estimated revenues of $125 million for Exact Sciences in 2006, which assumed just 1.9 million tests or 8%-9% penetration of the potential patient population.
Thomas Weisel Partners echoed Merrill Lynch's sentiments in its initiation report, and stated, "We believe that EXACT's technology is non-invasive, greatly enhances detection and will dramatically improve compliance rates." Thomas Weisel launched coverage with a $17 price target and later increased it to $20. These positive recommendations (along with an aggressive price target of $22 from the now out of business Fulcrum Global) drove the stock price to its peak closing price of $17.55 on September 11, 2003.
So how did this "breakthrough" test work? PreGen-Plus could not get past its high cost, its low bang-for-the-buck, and its inconvenient stool collection process. Within a year, the stock price plunged to $4.49, a drop of 74%! Per the 2007 10-K, in the three and a half years between the commercial launch of PreGen-Plus in August 2003 and December 31, 2006, LabCorp received approximately 12,500 patient samples for testing from physicians across the country. Exact Sciences' product royalties over this time frame totaled just $559,000 and the Company never earned an operating profit.
Finally, in 2007, the FDA issued a warning letter to Exact Sciences regarding "serious regulatory problems," as seen here, and in 2008, LabCorp stopped offering PreGen-Plus. As could be expected, the stock price completely collapsed, falling to $0.35 by November 24, 2008, a drop of 98% from its high!
Failure 2: ColoSure 2009-2011: $71,000 in Test Royalty Revenues
Undeterred by the failure of PreGen-Plus and the resulting stock price of $0.35, Exact Sciences tried again in 2008, and again took LabCorp along for a ride. Like PreGen-Plus, ColoSure was a non-invasive take-home test, but it had better performance statistics. As can be seen here, ColoSure tested at a sensitivity range of 72%-77% and a specificity range of 83%-94% in average risk individuals. It also promised to be more cost-effective at $200 per test, as per Exact's 2Q 2008 conference call here.
Just like the PreGen-Plus before it, the ColoSure failed to gain testing traction, doomed by its less than convincing sensitivity and specificity, unpleasant collection process, and high cost relative to existing tests on the market. It also never received FDA approval or CMS reimbursement. From 2009 to 2011, Exact Sciences earned just $71,000 in royalty revenues from ColoSure. Over that same time, Exact lost $49.4 million in net income and burned $53.9 million in cash flow from operations. By 2012, ColoSure was no longer generating any revenue for the Company.
By this time, EXAS rose from its low of $0.35 in 2008 to $10.98 the day the 2012 10-K was released. The statement that ColoSure provided $0 revenue for the company in 2012 dropped the price to $8.14 by April 24, 2013, but a complete collapse was avoided due to newfound hype behind Cologuard.
Cologuard Will Be The Third Strike
Exact Sciences is currently at it again with its third generation colon cancer screen. As we have shown in parts one through four of our series, we believe Cologuard will fail for the same reasons as PreGen-Plus and ColoSure. It is not as effective as existing tests, it is too expensive, and it is a really unpleasant test to administer.
Let's first elaborate on the controversy regarding bleeding polyps and cumulative testing results. CMS clearly states its opinion on the bottom of page 44/top of page 45 here.
In summary, the authors state that the assumption of the conditional independence of repeat FIT/FOBT screens is reasonable due to the fact that a particular lesion may be bleeding during one test, but not another. However, they then state that this assumption does not hold for DNA stool tests because regardless of whether or not a particular lesion is bleeding, the DNA of a polyp or cancerous lesion is unlikely to change. Therefore the outcome of successive stool DNA tests will also not change, eliminating much of the benefit of repeat testing with DNA based tests such as Cologuard. This actually strengthens the point that we have been attempting to make: not only are the cumulative results of repeat FIT testing incredibly important, but their benefit to DNA based testing is largely absent.
Given the importance of cumulative testing to FIT sensitivity, it is appropriate to compare one year of Cologuard results with three years of cumulative FIT results.
Having said that, it is clear that over any three year period Cologuard is ineffective relative to its competition. Further, even if the test were perfect (100% sensitivity and specificity), CMS states it would still only be economically justifiable at $237-$302 as we showed in part four, much less than the $400 that the company is planning on charging.
LabCorp Is Not Fooled Again
With the Company's share price close to 52-weeks highs, investors clearly believe the hype behind Cologuard. However, we think it's important to point out that the Company will commercialize Cologuard on its own. LabCorp has been fooled by Exact Sciences already and appears to have learned its lesson after two previous failed tests. Quest Diagnostics (DGX), the other major player in the diagnostic testing industry, has also avoided Exact Sciences altogether.
In our opinion, the fact that neither Quest nor LabCorp is partnering with Exact Sciences to provide Cologuard indicates that they believe Cologuard has minimal value and is a failure waiting to happen. Complex, high margin tests, such as Cologuard, are a major driver of growth and profits in the diagnostic testing industry, so it's rare for the large testing companies to pass on partnering with a new molecular diagnostic test if there's even a reasonable chance the test will be successful.
With no partners and forced to commercialize Cologuard on its own, Exact Sciences is likely to burn through its cash. The benefits of partnering with a diagnostic testing company are obvious: a higher chance of success at a lower cost to the test manufacturer. The probability of success is higher because the big testing companies have existing relationships with doctors and hospitals. Likewise, risks to the test manufacturer are lower because it will incur lower out of pocket costs. As a new, unproven entrant into this field, we believe that Exact Sciences will have trouble developing sales channels and convincing doctors to use its expensive and inferior test, especially given that it will be competing with Quest, the giant in the industry who owns the InSure FIT, Cologuard's chief competition. Indeed, we have repeatedly seen independent single-product companies fail to commercialize products with questionable efficacy, such as Amarin and Vivus.
A Pipeline Of Better Competitors
Perhaps companies like LabCorp and Quest passed on selling Cologuard because they can see the coming competition. In our analysis, we have identified at least three non-invasive potentially disruptive technologies that may revolutionize colorectal cancer screening.
Epi proColon Blood Assay Test
The first is a blood test called Epi proColon Early Detection Assay that can be run from the same blood sample patients give at their doctor's office for tests such as cholesterol and triglycerides. On December 19, 2012, Epigenomics released data from a head-to-head comparison of Epi proColon to the FIT, which showed Epi proColon had a sensitivity of ~71%-74% and a specificity of 81%, as found here. While the sensitivity and specificity are not as high as some FITs, the breakthrough for Epi proColon is its ease of administration. It can truly be done as part of a normal, annual checkup using a sample of a patient's blood, which can significantly improve the stubbornly low compliance rates associated with take-home colorectal cancer tests. On February 25, 2013, the FDA accepted Epigenomics's PMA and gave it priority review status, which means that if it is approved it is likely to reach the market before Cologuard.
Another potentially disruptive technology is the PillCam, which involves swallowing a small, pill-sized camera that rotates taking thousands of pictures as it moves through the gastrointestinal track. The leader in this type of testing is the PillCam, produced by Given Imaging (GIVN) (or "GI").
As found here, the PillCam's sensitivity and specificity are stunning. It showed an ability to detect pre-cancerous 1cm (10mm) adenomas or larger at a sensitivity of 92% and a specificity of 95%, which are comparable to a colonoscopy. For smaller pre-cancerous adenomas at least 6mm (0.6cm) in size, the sensitivity was 88%. The benefits of the PillCam (highly accurate, non-invasive, convenient, lack of bowel prep, etc.) are pretty obvious, as shown here.
Given Imaging filed for FDA approval in November 2012. As can be seen here, GI expects the US Food and Drug Administration (FDA) to approve the PillCam Colon by the end of September. It received EU CE Mark in September 2009 and is already commercially available throughout Europe, Latin America, Canada, and parts of Asia.
CT (Virtual) Colonoscopy
The last emerging competitor we'd like to highlight is the CT (Virtual) Colonoscopy. As seen here, this non-invasive procedure takes a computed tomography (CT) scan of the entire rectum and colon. Although it requires a highly skilled physician and bowel preparation is necessary, there is no need for sedation and there is no rectal penetration. It is incredibly effective and can recognize ~90%-95% of cancers, as documented in this study.
The reason the Virtual Colonoscopy has not gained greater acceptance is its high cost and lack of reimbursement. However, there has been a building groundswell of support for reimbursement, as documented here in the American Journal of Roentgenology. With endorsements already from the American Cancer Society, the US Multi-Society Task Force, and the American College of Radiology, the Virtual Colonoscopy could gain wide acceptance if it is approved for reimbursement.
All three competitors referenced above have significant advantages relative to Cologuard. They address the problems of compliance and efficacy that existing colorectal screening tests are plagued by, and may do so at equivalent or even lower costs when compared to Cologuard. We believe these tests are the future, not repackaged failed products such as Cologuard.
Ignorance or Fraud?
Through our five part series, we believe we clearly illustrated that Cologuard will likely fail. In order to shield investors from this reality, we believe Exact has used deceit, data manipulation, dishonesty, and post-hoc data mining in order to distract investors from the low quality of Cologuard and the small market for it.
Exact Sciences' Latest Misrepresentation
When evaluating Exact Sciences' claim that Cologuard can detect 66% of pre-cancerous lesions greater than or equal to 2cm, it is important to note Cologuard's original intent. In addition to screening for colorectal cancer, Exact Sciences wanted Cologuard to screen for pre-cancerous polyps and had a precise target for detection of greater than 50% sensitivity for all pre-cancerous polyps, as stated by CEO Kevin Conroy on the company's Q4 2012 conference call (a copy of the transcript is here).
On April 18, 2013, the Company announced that Cologuard exhibited a 42% sensitivity for pre-cancerous polyps, which clearly missed their 50% target. Instead of recognizing Cologuard's inability to screen for pre-cancerous polyps, the Company boasted that Cologuard detected 66% of pre-cancerous lesions greater than or equal to 2cm in size.
While this may be true, the 66% sensitivity heralded by the Company is a meaningless statistic that resulted from the exact same type of post-hoc data mining that we have warned against on Seeking Alpha in the past here. In fact, prior to the disappointing result of the DeepC trial, Exact never once discussed the importance of detecting polyps larger than 2cm.
We examined other colorectal screening studies and could not find one single study that highlighted any test's sensitivity to precancerous polyps greater than 2cm. In fact the American Cancer Society article that we used to source the InSure FIT statistics, found here, examines almost a dozen different endpoints, not one of which even alludes to 2cm polyps. The 2cm polyp sensitivity is a completely arbitrary endpoint and without having been prespecified by the Company is more likely to be random noise than a true breakthrough.
It's clear to us that Exact Sciences' data-mined for this result in order to distract investors from Cologuard's actual 42% sensitivity for precancerous lesions. The fact that the Company is trying to mask the disappointing Cologuard results for pre-cancerous polyps with irrelevant statistics obtained through post hoc data-mining is alarming. We believe Exact Sciences is doing everything it can to shield investors from the truth, which is that Cologuard is an ineffective and soon to fail test, an all too familiar outcome for Exact Sciences.
The Truth: Cologuard Is Inferior, Too Expensive, And Unlikely To Gain Reimbursement
The 66% sensitivity is the just one example of Exact Sciences' misrepresentations. While we covered these points in detail in parts one through four, it is helpful to see all the instances of deceit and data manipulation associated with what we know about Cologuard in one place. In combination it paints an overwhelming picture of deceit and dishonesty.
- In trying to show that the Cologuard is superior to the FIT, management never provides a true apples-to-apples comparison or any side-by-side comparison. Any comparison to the FIT test should be done over a three-year time horizon due to the different administration frequencies for each test. When the proper comparison is done, it is clear that Cologuard identifies fewer cancers and misses more incidences of cancer than the FIT. We covered this in detail in part one of our series. Further, as CMS proved in the research above, cumulative results of FITs are incredibly relevant due to the random frequency with which polyps bleed, while the same cumulative results are of questionable value for stool DNA testing, because the tests rely partially on DNA, which may be unlikely to change from one test to the next.
- We also know that Exact Sciences compared Cologuard to old FIT test statistics from 2005. The outdated FIT study that the Company references puts the FIT in the worst light. The Company chose the single worst FIT study we could find from the last 10-years. Further, Exact Sciences had more appropriate data to use, or at least to average in with the 2005 Morikawa study that it completely ignored, as evidenced in slide 16 of its May 2013 presentation, which can be found here. As we illustrated in our first article, this slide looks similar to slide 18 in the June 2013 presentation except Exact provided a range in May (with two sources). In June, the Company dropped the second source, and simply reported the lower end of that range. Further, the averages of more up to date studies of FIT sensitivity and specificity are significantly better, and make the Cologuard results look even worse.
- We believe the Company is hiding data from investors. The Company used a FIT as a comparison in its DeepC clinical trial but it has not released this data. Exact Sciences clearly states that a FIT was administered as a comparison here. However, the Company never published the results. We believe this is because directing investors to the worst available historic control makes Cologuard appear better. Put simply, nothing in the Company's highly promotional history is consistent with holding back good data. Instead, the Company has historically been aggressive at publicizing good data and has delayed and distracted investors when presenting poor data. We have no reason to believe it is any different now.
- Another impropriety was the Company's use of the OC FIT-CHEK test in DeepC rather than the leading take-home FIT, the InSure FIT. At the end of the day, it is the InSure FIT that the Cologuard will ultimately be competing against, so it is odd that the Company chose a FIT used primarily in doctor's offices in its attempt to gain FDA approval and CMS reimbursement. When compared to the InSure FIT over the same time horizon, Cologuard underperforms in every single metric, regardless of which set of reported sensitivities and specificities one uses for Insure FIT. We covered this in detail in part one of our series.
- Not only is Cologuard inferior to the InSure FIT, but it is also significantly more difficult for patients to administer, an incredibly underrated aspect of Cologuard that many analysts on the sell-side fail to recognize. Compared to InSure, Cologuard requires patients to handle more excrement more frequently, which will likely make compliance significantly lower. We covered this in detail in part three. As a result, we don't believe Cologuard is likely to gain wide adoption.
- Lastly, Exact Sciences wants investors to believe that CMS reimbursement for Cologuard is likely and at a high price per test. While it is true that the Company has done everything that it can to earn reimbursement, we know that CMS has studied the Company's previous tests in detail and concluded that they are not cost-effective, as evidenced by its efficient frontier of tests displaying cost vs. quality of information found here on pages 35-36. As a result, we believe CMS reimbursement at any profit producing price point is unlikely.
Over the course of this five part series, we have explained why we feel that EXAS is worth less than $3 per share and that the stock should be aggressively sold or shorted.
We have done so by looking deep into the six different reasons for Cologuard's impending failure.
- The Cologuard test is actually worse than what is currently commercially available
- The economics behind the test don't make sense
- Nobody will use Cologuard even if it is approved by the FDA because Cologuard is gross
- The bull case, which stresses that the test will become as widely accepted as the pap smear, doesn't make sense
- Cologuard faces a number of new non-invasive competitors that are either easier to administer or have better sensitivities and specificities
- The Company has misrepresented its data to fool investors
Given what we know about Exact Sciences' history of product failures and the Company's past misrepresentations of data, we do not understand how anyone can give Exact Sciences the benefit of the doubt any longer. Investors should now consider themselves sufficiently warned about the dangers of investing in Exact Sciences.
1. The DeepC test comparison makes the following assumptions:
- 10,000 patients over the age of 50 are tested
- Based on the company's DeepC trial, the expectations for positive tests for cancerous and precancerous polyps for this age group is approximately 0.64% and 7.52% respectively
- The Cologuard test is given once every three years, while the FIT is given every year
2. Below are links to FIT data that Exact Sciences failed to mention in its June 2013 Presentation:
- HemeSelect (Fujirebio, Inc., Tokyo, Japan), Hemoccult-ICT (Beckman Coulter, Inc., Primary Care Diagnostics, Los Angeles, CA), and Magstream (Tokyo, Japan)
- American Cancer Society (InSure Data used in Comparison)
- Graeme P. Young
- Wjikerslooth-found in May 2013 presentation above