Type:  Event-driven / Speculative Pharmaceutical

AVEO:  A Seth Klarman small-cap pharma pick

AVEO pharmaceuticals offers what we believe mis-priced wager on clinical trial results expected in Q4 2011 at the earliest, offering up 80% probability for a 50 to 75%+ return on the upside vs a 20% probability for a -60% return on the downside (to net cash) in about 6 months time . Expected value today given our prices and probabilities should equal between 22 and 24 per share.  In per share terms, we believe we're risking $10-12 per share to get $8-10 with an 80% likelihood of success, or about a 5:4 payout with 4:1 odds.  We'll take bets like this all day long, so long as they are sized appropriately.

The opportunity exists because of a) the speculative nature of preclinical biotech b) the small, under followed nature of the company and c) the recent share issuance

AVEO history: buying in to a late-stage VC project at just the right time

Rather than re-create the wheel, just read this article.  Key take-aways are that a) the people are excellent and b) the project has b

AVEO owns 45% - 50% net interest of the next potential Avastin in North America and Europe

Tivozanib (TIVO) is AVEO's special sauce.  TIVO is a small-molecule oral vascular endothelial growth factor (VEGF) inhibitor, currently in Phase III trials for the treatment of advanced renal cell carcinoma (RCC - kidney cancer) and in Phase Ib/II trials for numerous other indications. Very briefly, VEGF inhibitors work by fitting molecules into the 3 types of VEGF receptors in cell walls, preventing key hormones from entering cells and stalling the blood-vessel growth critical for supporting cellular division.  The market for VEGF inhibitors is expected to reach $14 billion by 2014, up from $10 billion today.  Avastin, with ~6 billion in sales, is a VEGF inhibitor.

TIVO is significantly superior to other small-molecule oral VEGF inhibitors because it a) has a much longer half-life in the body and b) has a much better fit to the VEGF receptors without hitting other off-target receptors.  The impact of these benefits can be seen in the dosing quantities.  Patients on TIVO take approximately 1.5 mg per day in a single dose, where patients on other VEGF inhibitors currently in-market take between 50 mg and 800 mg per day.  Axitinib, another small-molecule oral VEGF in Phase III and perhaps TIVO's closest competitor going forward, requires 2 doses of 5 mgs per day.

Physicians frequently say that side-effects are entirely correlated to dosing quantities.  This heuristic rings true with TIVO as well - the tolerability profile is so much better for TIVO than it is for any of the other currently marketed small-molecule oral VEGF inhibitors.  As an example, oncologists currently prescribe Sutent for RCC in advanced patients today.  While the median progression free survival (PFS) for Sutent is around 11 months, meaning that on average tumors don't grow for 11 months while on Sutent, patients frequently discontinue drug usage at around the 7 month mark due to intolerability leading to lack of dosing compliance.

The tolerability of TIVO is a major reason why it could be a big time blockbuster as it allows for combination treatments.  The VEGF cancer-fighting cocktail theory goes like this: prevent further tumor growth with the VEGF inhibitor and simultaneously attack the current tumor mass with chemo/radiation.  This is how a majority of patients on Avastin are treated.  And current results for other indications are promising - Phase Ib clinical trials have shown that TIVO can be combined with targeted therapies and chemo at full dosage - Torisel in RCC, FOLFOX6 in colorectal cancer, Taxol in metastatic breast cancer, and Xeloda in breast and colorectal cancers.

Furthermore, physicians are excited about TIVO, as evidenced by the swift enrollment in the Phase III trial - 6 months ahead of schedule.  Indeed, in Europe, TIVO was granted orphan drug status for RCC treatment by the EMA, which accelerates the regulatory process.  And when physicians are excited, investors should be excited.

TIVO is roughly comparable to Sutent and Nexavar, both oral small-molecule drugs approved for multiple indications, including RCC.  In 2010, Sutent and Nexavar did 1.1 and 0.9 billion in sales worldwide respectively.  On a comparable basis, TIVO is expected to be significantly more potent than Nexavar and somewhat more potent than Sutent in treating RCC populations.  And, as stated above, TIVO's tolerability profile is quite differentiated from both of these drugs.  Note that ONYX pharmaceuticals (ONXX), the company that owns a 50% interest in Nexavar, sports a 2.2 billion market cap (Sutent is owned by PFE).

The TIVO trials - early data hints at extremely positive results

TIVO's Phase II data showed remarkable numbers.  The median PFS for the entire population of Phase II participants (272) was 11.7 months, about 1 month higher than Sutent, the current standard-of-care for treating RCC.  More importantly, in the patient population who had a prior nephrectomy and a clear cell histology (176), PFS was a whopping 14.8 months.  In addition, the tolerability profile was fantastic, with minimal off-target toxicities associated with Sutent and Nexavar (diarrhea, fatigue, stomatitis, and hand-foot syndrome), and the toxicities associated with the mechanism of action (hypertension and hoarse voice) are both very manageable.

The Phase III trial is set up as a head-to-head against Nexavar, where about 260 patients will take TIVO and 260 patients will take Nexavar.  Patients in this trial must have had a) a prior nephrectomy and b) a clear cell histology, so these patients are identical to the 176 in which TIVO performed fantastically.  Initial results will be available at the 310 patient event mark (an event is defined as noted tumor growth based on CT scans independently reviewed).  The trial was 'powered' with the assumption that the median PFS for the Nexavar trial would be 6.7 months - significantly above the 5.1 to 6.5 months Nexavar has shown in all independently-reviewed trials - and the median PFS for TIVO would be 9.7 months - significantly below TIVO's shown PFS in this patient population.  The trial completed enrollment in August 2010.

Initial results were expected mid-2011, about 12 months after the trial had begun.  By then, it would be extremely likely that all the patients on Nexavar would have shown progression and that a good bit more than half of the patients on TIVO would have shown progression.  However, management re-guided initial results until Q4 2011 at the earliest and only because of the event rate.

There are three possibilities for this: 1) Events are slower than expected in the Nexavar arm.  We find this unlikely considering historical Nexavar trials, which have hardly ever done better than 6.5 months PFS.  Admittedly though, this outcome would be bad for AVEO.  2) Events are slower in the TIVO arm.  We find this very likely considering what we know about the Phase II trials in the similar patient population.  In fact, we believe that a shift from 9.7 to 15 months median PFS in the TIVO-1 arm can almost entirely explain the 6 month delay in event expectations.  This would be incredible results for AVEO. P3) Some combination of the two.  Again, we think this is unlikely because of #1 above.  Unfortunately, even management cannot tell which arm the events are coming from, only that they are coming significantly slower. 

So then, how do we come up with some meaningful probabilities for the Phase III trial?  To begin, we take a seminal study from Deloitte about preclinical drugs, specifically the oncology group.  Out of the 46 compounds which advanced to Phase III trials, 28 filed for an NDA, meaning 18 did not.  Thats roughly a 39% attrition rate for the industry average.  Of this 18, 13 were abandoned because they lacked efficacy in their trial, 3 were abandoned for pipeline prioritization, 1 was abandoned for adverse effects, and 1 didn't provide a reason.  We believe the evidence clearly shows efficacy.  We also know that TIVO will not be abandoned for pipeline prioritization.  However, to be conservative, we leave 33% of the normal efficacy risk and 100% of the adverse event risk and other risk.  This leaves ~34 drugs in our relevant bucket, with ~6 of them being abandoned, or an 18% attrition rate.

So what exactly are we buying?

AVEO owns the following:

• 45 to 50% of the economics of TIVO in North America and Europe as well as a double digit royalty on TIVO in non-Asia rest-of-world.
• Up to 1.3 billion in milestone payments associated with TIVO
  • 570 million in clinical milestones
  • 780 in commercial milestones
• Wholly-owned ficlatuzumab, which is still in early development although the recent equity raise allows for an acceleration of profits
• Moslty wholly-owned AV-203, in development with BIIB
• Royalty rights and milestones (up to 540 million) associated with anti-RON antibodies, after partnering with JNJ recently
• A proprietary drug discovery platform that has produced at least three viable compounds
• 340 million in adjusted cash, burning approximately 100 million annually.

All of this today costs approximately 820 million. 

 

 

Good question - and somewhat of the tricky part.  The honest, simple answer is that we don't know.  It's an amazingly ridiculous exercise to come up with probability-weighted DCF models for each drug and each indication and for each step of the approval process while simultaneously guessing how much R&D expense will be for each step because almost indefinitely we'll be unbelievably off.  We like to think about the small cap pharma companies as giant options with imbedded giant options, due to the high failure rate of trials but immense profitability upon success.  In the case of AVEO, cash flow generated from TIVO for RCC (assuming success) will be re-invested into the same product for other indications, and each other indication will be another option - the cost of the option being the development/trial cost and the payout being access to incredibly profitable markets.  Furthermore, the additional compounds in the pipeline carry similar optionality profiles.  Indeed, a discovery platform is an option on an option generator!  Unfortunately, this makes a hard value relatively tough to pin down.

 

We think the most appropriate method for valuing companies like this is via comparables, and we think ONXX is a fair one considering the 50% lead drug on-market candidate's mechanism of action and wholly-owned secondary compound with good potential.  In fact, if anything, TIVO is much more valuable than Nexavar in RCC, and our guess is that it is even in other indications.  However, ONXX is about 3 to 5 years ahead of AVEO in its approved drugs and pipeline but its profile is roughly comparable.  We believe that, should AVEO be successful in its RCC treatment, it too will garner a similar if not better marketable and clinical profile in 3 to 5 years. 

 

(A quick note: we realize that there are many differences between these two companies, not least being a) Nexavar rights in Asia, which is an area driving sales growth currently b) depth of Nexavar trials for other indications and c) larger cash cushion but we believe these to be fairly balanced by the i) TIVO milestones ii) superiority of the TIVO drug candidate and iii) earn out owed by ONXX should its secondary compound do well commercially.)

 

So, to simplistically compare, we discount ONXX's market cap by a high discount rate of 15% over a 3 to 5 year period:

 

Upside value	2200	Time
		3	4	5
	10.0%	1,653	1,503	1,366
	12.5%	1,545	1,373	1,221
Discount Rate:	15.0%	1,447	1,258	1,094
	17.5%	1,356	1,154	982
	20.0%	1,273	1,061	884

 

At 46 million fully diluted shares outstanding, that looks like:

 

 

Upside per share		3	4	5
	10.0%	35	32	29
	12.5%	33	29	26
	15.0%	31	27	23
	17.5%	29	25	21
	20.0%	27	23	19

 

And given our probabilities, expected value looks like:

 

Expected value		3	4	5
Up prob	80%	30	27	25
Down prob	20%	28	25	22
Cash per share(downside)	6	26	23	20
		24	21	18
		23	19	16

 

Clearly there's quite a bit of sensitivity to the variables here, but we're comfortable with a 15% discount rate given the success of the recent equity raise at 17.50.  One could even argue thats too considering the WACC according to bloomberg is only 10%!

 

 

There are approximately 150k patients annually diagnosed with RCC in North America and EU and the incidence is growing as the population ages.  Of those, approximately 25% have metastatic cancers which will be treated with targeted therapeutics.  Of those, approximately 80% have a clear cell histology and have had a prior nephrectomy.  So we're talking about a sweet spot of 30k patients annually and approximately 7.5k additional patients treatable.   I'm relatively comfortable saying that, should TIVO show excellent results, it will eventually become the market leader, so approximately 33% share of this market is fair.  Should median PFS for TIVO be in the sweet spot of 15 months, its plausible that, on average, each patient is treated for 15 months.  We'll say of the sweet spot, they're on average treated for 13 months, and that of the 7.5k patients, they're on average treated for 10 months.  So, on average, that's approximately 156k patient months per year.  (Yes, we know that you can't treat a patient for more than 12 months in a year but that patient month just adds to the following year).  Now, a big question is on average, what will this drug cost?  I don't think its unreasonable to compare to Sutent in this regard, as it will be more potent (again assuming more.  Sutent currently costs on average about 4k per month, and prices are increasing.  We think TIVO can sell for at least 4.4k a month when it reaches market.

 

Given these assumptions, TIVO can do about 730m in gross revenues.  At AVEO's share of approximately 45% (less than 50% due to royalty owed on NA sales), TIVO for RCC alone should generate about 320m in net revenue for AVEO before factoring in market (incidence) growth.  That prices the company at about 2.5x future revenues for RCC alone today.  On a cash flow basis, we think operating margins of 50% are reasonable for the drug and should increase as the commercial infrastructure builds scale.  Of course, this doesn't include R&D spend as that's an investment and not an expense.  At these margins, TIVO should generate 160m in pre-tax cash flow for AVEO, which will promptly be used to fund a) other indications b) other compounds or c) the discovery program.  So in other words, in the upside scenario, it's possible we're paying 5x forward (very forward, mind you) pre-tax cash flow for 50% of the most potent VEGF inhibitor found to-date, a solid pipeline with promising wholly owned drugs, total milestone payouts greater than 2x the market cap, enough cash currently to fund itself through at least 2013, and an excellent team of people running the show.  Not bad, we guess.  Our upside valuation from the ONXX comp values the company at about 9x forward pre-tax cash flow, which we believe still significantly undervalues the potential.

 

We're calling adjusted cash per share at year end the downside value (est 270m) even though we believe the pipeline has real value.  The company will likely be sold should the Phase III trials come in poorly.

 

 

 

Weak top-line data released between 4Q11 and 1Q12.

Price pressures in health care force pharma to lower prices of targeted therapies.

New drugs show more potential than TIVO.

Poor results out of other trials reduce value of pipeline

 

Top line data released between 4Q11 and 1Q12