Description

Quick Description:

GHDX is a maker of genomic-based clinical diagnostic tests for cancer.  Since 2004, GHDX has marketed a single product called Oncotype DX, which analyzes the expression levels of 21 genes in tumor tissue samples and helps oncologists decide whether or not to subject breast cancer patients to chemotherapy.  GHDX is also seeking to create and market tests for additional indications and expand its market opportunity.

Quick Thesis:

There are several problems with the GHDX bull thesis.  First of all, GHDX's product is currently marketed under a regulatory "loophole" whereby they didn't have to get FDA approval for their product, but that will likely change soon.  Second, GHDX will find it difficult to grow at market's expectations (38% in 2009 and 25% in 2010) because Oncotype DX has already penetrated more than 50% of its entire market, and GHDX's next product (for stage II colon cancer) will likely disappoint.  Third, Oncotype DX, like other new genomic-based tests, receive a very high reimbursement amount that will likely decline over time.

Product Description:

Oncotype DX is a genomic-based test GHDX designed for breast cancer.  Specifically, GHDX took 250 human genes that are known to be involved with cancer, and then based on further studies narrowed it down to a 21-gene panel.  Oncotype DX takes a tumor sample from a breast cancer victim, checks the expression profile of these 21 genes (ie. how active these genes are inside the tumor cells), and runs a proprietary "black box" algorithm to turn that information into a single number thats called a Recurrence Score.  GHDX has demonstrated that tumors with a high Recurrence Score have a higher rate of recurrence and have a higher likelihood of benefit from chemotherapy.  This predictive capability is key to Oncotype DX's success within a subset of breast cancer victims, because it helps doctors make a decision on whether or not to order chemotherapy.  Conversely, Oncotype DX is not useful for super early-stage patients that almost never need chemotherapy, or for late-stage patients that almost always require chemotherapy.  In those cases, the doctor already knows whether or not he's going to order chemotherapy, so the utility of ordering a test just isn't there.

Regulatory Overview:

Oncotype DX isn't currently regulated by the FDA because GHDX is marketing it under accreditation from CLIA (Clinical Laboratory Improvement Amendments of 1988).  Specifically, GHDX setup its reference lab in CA, where it performs a clinical laboratory service on all the tumor samples it receives.  By qualifying under CLIA, GHDX avoided the much more rigorous and intrusive FDA approval process.  GHDX is by no means the only diagnostics company taking advantage of this loophole, but in any case the party may be over soon.  There is a perception that molecular diagnostics is a real "Wild West" right now, and that numerous companies are selling products and making therapeutic claims that have not been verified by the FDA.

In 2006/2007, the FDA issued draft IVDMIA Guidance that makes it clear that GHDX's Oncotype DX falls squarely in its definition of IVDMIA (in vitro diagnostic multivariate index arrays), and that it will be regulated.  You can find the link to the FDA's draft guidance below: (http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm079148.htm)

In addition, Genentech filed a Citizen Petition with the FDA last December urging the agency to take on greater oversight of diagnostic tests that are intended to guide therapeutic decisions and to regulate all laboratory-developed tests.  Currently, I believe the FDA is very close to coming out with a final guidance (possibly by year end).  My sources indicate that the FDA is intent on defending its turf on IVDMIAs, and the final rule will likely require all new devices to seek FDA approval, and will order all existing non-approved IVDMIAs will need to conduct clinical trials over a 12-to-18 month grace period, or otherwise be pulled from the market.

While I'm not suggesting that GHDX's Oncotype DX doesn't work or that it wouldn't get eventual FDA approval, there are a few interesting points worth discussing here.  First of all, GHDX's Oncotype DX never went through a full prospective clinical trial.  Instead, GHDX picked its 21 genes and did a "prospective study" using retrospective data.  As all good data miners know, any study using primarily retrospective data is never as good as a true "blinded" prospective study.  Given enough time and tries, you and I can also construct a 20-gene panel that can pass a "prospective study" using only retrospective data.  Second, Oncotype DX's original study was based only on patients that were on tamoxifen.  In other words, Oncotype DX's predictive value is in demonstrating that certain patients (with high Recurrence Scores) have better survival on tamoxifen + chemotherapy vs tamoxifen alone.  While tamoxifen is still used today, it's by no means universal and thus, scientifically speaking, I can't even say for sure that Oncotype DX works for non-tamoxifen treated patients.

Nonetheless, since GHDX has delivered 100,000 tests at this point and the FDA has approved competing products with only retrospective data, I will assume that GHDX will eventually get a pass from the FDA, even if its operations are temporarily disrupted (and not to mention bad publicity, since most physicians today probably don't realize that Oncotype DX isn't FDA-approved).

GHDX's management also believes that if necessary, they can get FDA approval for their product.  But that begs the question - why haven't they done so already?  They have plenty of data already, and surely having FDA approval prevents future headache and can be used as a good marketing tool today?  I believe a reason for GHDX to pick the CLIA path is because seeking FDA approval would force GHDX to open up its "black box".  Currently, Oncotype DX uses a 21-gene panel and an algorithm that it doesn't fully reveal.  I have heard speculation via industry sources that while Oncotype DX has 21-genes in its panel, it's heavily weighted toward only 5 genes.  Why is this significant?  Because we already know that Oncotype DX measures the activity level of some well-known cancer-related genes like ER, PR and HER2.  If the test predominantly relies on only 5 genes, there may only be 1-2 "extra" genes that it accounts for.  If accurate and widely known, it's unlikely that insurance companies will be happy about paying $3000 for a test that's primarily measuring activity levels of just a few well-known genes.

Reimbursement Overview:

Currently Oncotype DX is reimbursed by Medicare and private insurance that represent 90% of the insured lives in the US, and I'm not suggesting that Oncotype DX's coverage decision is at risk.  I was somewhat surprised that the original Contractor Medical Director for the California Medicare Part B program who approved Oncotype DX's reimbursement level has since left the agency and is now a paid lobbyist for GHDX.  I can only hope his gig with GHDX started after he left Medicare.

The average reimbursement is about $3000 per test (vs. $3900 list price).  While I don't have any indications that will change anytime soon, I have heard from field sources that $3000 for an mRNA expression level test is probably unsustainably high and will sooner or later come under scrutiny.  I'm not counting on, but will hope for, a situation like CardioNet (BEAT) in the remote cardiac monitoring space - where they also found a friendly jurisdiction (Highmark) with a friendly medical policy director, set up shop, and received very generous reimbursement rates until that medical policy director finally left, a new guy took over, and the ax came down 2 years later (along with commercial reimbursement as well).

While GHDX believes eventually they can market their product overseas, to date they've had little success obtaining reimbursement and thus international sales are minimal.

Market Size - Breast Cancer:

Currently there are approximately 200k new breast cancer victims each year in the US.  Oncotype DX is primarily used for ER+/N- patients (80-100k/year).  GHDX is forecasting 50-53k treatments in 2009, which means they've more than half penetrated the ER+/N- patient pool.  While Oncotype DX recently received clearance to market to ER+ patients with 1-3 positive lymph nodes (20-30k/year?), I believe most oncologists will err on the side of caution and offer the patient chemotherapy treatment regardless of what Oncotype DX says (ie. rendering it pointless).  Imagine if you or your wife had breast cancer and found positive lymph nodes in your breast, wouldn't you want to do chemotherapy "just in case"?  Never mind the fact that most oncologists are paid on a fee-for-service basis, and make much more money from doing chemo treatments than the diagnosis/consultation.

Potential Markets - Ductal Carcinoma in Situ (DCIS):

DCIS is a very early stage of breast cancer that impacts 60k/year in the US.  While this is a market GHDX is actively targeting with Oncotype DX, the feedback I've received is that recurrence is so low in this population pool that it's unclear if Oncotype DX is necessary or if the uptake will ever be significant.

Potential Markets - Stage II Colon Cancer:

GHDX is also working on a test similar to Oncotype DX for the colon cancer market.  The main problem they're facing is that the results from their validation study (QUASAR) only met the prognostic but not the predictive endpoint.  In plain English, it means GHDX's colon cancer test can tell you how likely it is that you will get a recurrence of colon cancer, but it failed to predict if getting chemotherapy is going to make you live longer.  In the words of an oncologist I spoke to, "why would I order a test if I don't even know what to do with the information it gives me?  So what if you tell me I have a 40% chance of getting my cancer back, if your test can't prove that I will live longer with treatment?"  In my opinion, GHDX's colon test reveals a problem with tests only validated with retrospective data.  If you data mine hard enough, you will probably find a panel of genes that correlate to cancer recurrence, but predicting clinical outcome (not to mention in a prospective setting) is much harder.  I think this is another example of why the FDA needs to step in and regulate these tests and their predictive/therapeutic claims.

Competition:

Currently there is very little competition for Oncotype DX in the US.  Agendia's MammaPrint has received FDA approval, but like GHDX's colon test, it's a prognostic test and not a predictive test (ie. it cannot tell if the patient will benefit from chemo treatment).  Therefore, I don't expect Agendia to gain significant traction here.  Nonetheless, a number of other companies are working on similar products that may eventually take market share away from Oncotype DX (and at least pressure reimbursement levels).

By the way, it's important to pause here and think about what really is the core competency here.  While GHDX would like people to believe that they're developing a platform that can launch Oncotype DX-like products for other types of cancers, it's important to understand that their "core competency" and intellectual property is primarily around the 21-gene panel itself, which is only for ER+/N- breast cancer.  The "platform" or assay technology is for the most part commodity and based mostly on standard laboratory techniques that are part of the public domain.  What does that mean?  GHDX's expertise doesn't transfer well from one type of cancer to another.  Even if Oncotype DX really works for breast cancer, those genes are probably not involved in colon cancer, so GHDX is back to square one along with everybody else looking for useful genes involved with colon cancer.  With many other companies looking to get involved with this space (Celera, Clarient, Applied Genomics, bioTheranostics, Exagen, University Genomics, Genzyme, LabCorp, Quest, Siemens, Roche, JNJ), it's unlikely that GHDX will get the head start in other cancers like the one it got in breast cancer.

Consensus Revenues and Valuation:

Finally, I believe GHDX will have a very difficult time meeting inflated street estimates.  As shown above, I believe GHDX has already penetrated >50% of its core market, yet consensus still has GHDX growing revenues by 25.4% and 22.5% over the next two years.  If shipments grew by the same rate, GHDX will have either fully penetrated its core breast cancer market, or found significant new opportunities elsewhere.

Another thing to note is that GHDX uses an accounting methodology that can be confusing to outsiders.  Specifically, they recognize under accrual basis all tests delivered to customers with firm pricing, and they recognize under cash accounting all tests delivered without firm pricing.  Historically they delivered most of their tests under cash accounting (because they were still in the process of setting up contracts), but over time they've moved more of their business to accrual accounting.  This has the impact of "understating" revenues in the early years and "overstating" revenues during the "catch-up" years.  (Please note that I use the phrase "understating" and "overstating" in quotes because I'm not implying improper accounting methodologies here.  They're following the correct GAAP rules, but GAAP accounting here isn't a good reflection of underlying business).  Thus in 2009, I'm expecting revenues to grow by nearly 40% while underlying shipments will only grow by 25-30%.  It's possible that revenues may continue to grow faster than shipments during the rest of the "catch-up period", but that only means revenues will continue to appear "inflated" for longer, and that the revenue growth rate will collapse even more once the "catch-up" is over.  Currently, I expect shipments to grow by 15% and 10% in 2010-2011, and revenues to growth by 25% and 5%.

As for margins, it's more difficult to say because GHDX has never made money before.  Consensus has them doing just under $0.50 in 2011, which puts the stock at 40x 2011E, which I believe is a stretch because by then they would've completely penetrated their core market.  GHDX currently trades at 3x 2010 sales.  I think a good comparison is with BEAT, which trades at 0.8x sales.  Like BEAT, GHDX operates a straight forward service business out of a central location that was cherry-picked for regulatory/reimbursement reasons.  Like BEAT, GHDX tries to justify its pricing not on a reasonable gross profit level, but on a dubious comparison to "status quo" (false positive and false negative errors).  Unlike BEAT, GHDX's core market is more than half penetrated and shipment growth is slowing quickly.  Of course, BEAT is getting its reimbursement cut while GHDX isn't going through that right now.  Thus I use a 20x 2011 PE for a base case ($10 stock price), and a BEAT-like multiple ($7 stock price) in a bear case scenario.

Disclaimer:

This posting is solely for the evaluation of club members and is not a recommendation to buy or sell this stock.  The views expressed are those of the author individually and should not be attributed to any affiliated investment firm, which may or may not hold positions consistent with the views expressed herein and may buy or sell shares at any time.  

Catalyst

- FDA requires all IVDMIAs to get pre-marketing approval.

- Shipment slowdown due to high penetration of core market.

- Lack of uptake in DCIS and Colon Cancer.

- Exaggerated revenue slowdown due to accounting methodology.

- Competitive pressure and potential reimbursement cut further down the road.