Last month, Threshold Pharmaceuticals (THLD) surprised the market with positive data for its lead program (TH-302) in pancreatic cancer. Although actual results were not published and despite several issues with the trial design, TH-302 generated what is possibly the best pancreatic cancer data set in over a decade.
Phase II data
Based on top line results from a large randomized Phase II, adding TH-302 (two different doses) to standard chemotherapy led a statistically significant difference in progression-free survival (PFS). PFS is the time patients live without experiencing disease progression. Patients who received TH-302 with Gemzar had a median PFS of 5.6 months compared to 3.6 months in patients who received Gemzar alone (p = 0.005).
The response rate in the combination arms was also superior to that in the Gemzar arm (22% vs. 12%). Interestingly, patients who received the higher dose of TH-302 appeared to do better than those receiving the lower dose, which is an important indication of the drug's activity. According to the press release, TH-302's safety profile was consistent with a previous Phase I trial for TH-302 and Gemzar, where skin and mucosal toxicities were the most significant side effects.
To put things in perspective, a statistically difference in PFS in pancreatic cancer is a very rare event. Even Amgen's (AMGN) ganitumab (AMG-479), currently in Phase III in pancreatic cancer, could not demonstrate a statistically significant PFS or OS benefit in its randomized Phase II. Ganitumab+Gemzar had a PFS of 5.1 vs. 2.1 months with Gemzar alone in an 80-patient trial. Unlike Amgen, Threshold evaluated TH-302 in a very large trial (214 patients), enabling the drug to reach statistical significance and making it more predictable of future Phase III studies.
The most impressive results to date in pancreatic cancer were generated using the aggressive chemotherapy regimen FOLFIRINOX. In a recently published Phase III trial, FOLFIRINOX was superior to Gemzar in both PFS (6.4 vs. 3.3 months) and overall survival (11.1 vs. 6.8 months), with both differences being highly statistically significant. Despite the clear benefit, FOLFIRINOX is associated with severe side effects, which preclude its use in the broad patient population. If proven effective, TH-302+ Gemzar could be widely used due to a favorable safety profile.
There are still several open questions pertaining TH-302's data. As a general comment, available information is based on a short press release, so caution is obviously needed when evaluating the data. Three main issues are physician's bias, the cross over protocol chosen for the trial and potential safety issues.
The trial's biggest issue was the fact it was an open label study, so physicians knew which patient was assigned to each arm. This is problematic since PFS was assessed by the treating physician, which might create a bias favoring the TH-302 arms. Bearing in mind that PFS was assessed every 8 weeks, identifying progression one scan later could have a major impact of almost 2 months, which is the alleged benefit for TH-302 in the trial. Still, it is hard to believe the highly statistical difference is driven purely by this investigator-bias. Pancreatic cancer is a very aggressive disease where progression in the form of tumor growth or new metastases is often associated with clinical symptoms physicians cannot ignore.
A second issue has to do with TH-302's effect on overall survival, the most important clinical endpoint that will probably be required for approval. The company did not provide any overall survival data, probably because the data set is not mature enough. Even when survival results are reported, they might be confounded by the trial's cross-over design that allows patients on the Gemzar arm to receive Gemzar+TH-302 upon progression on Gemzar alone. If TH-302 is indeed efficacious, any benefit in the active groups might be masked by giving the drug to patients on the control group.
In fact, the cross-over design creates two control arms: (i) patients who receive Gemzar alone followed by Gemzar+TH-302 upon progression (ii) patients who receive Gemzar alone and get off the trial upon progression. Ideally, the first group should do better, however, these two groups are not likely to be balanced as patients who are too frail and sick will not be kept on a regimen of active chemotherapy. As a result, patients who cross over have better prognosis inherently. It will also be interesting to see whether patients who cross over from Gemzar alone to Gemzar+TH-302 experience tumor shrinkage or prolonged disease stabilization.
Lastly, there is the issue of side effects, which could be crucial when two agents have overlapping toxicities. Side effects could be a major barrier even for an effective regimen in a deadly indication such as pancreatic cancer (see FOLFIRINOX's case). Therefore, the incidence and degree of side effects could have a major impact on TH-302's ultimate market share. According to the press release, the safety profile of TH-302 is consistent with that observed in an earlier Phase I trial- primarily rash and mucosal toxicities. These groups of toxicities are not common with Gemzar, which is encouraging. It is also encouraging that no hematologic or GI toxicities were mentioned in the PR, since these are toxicities frequently associated with Gemzar.
Although pancreatic cancer is not one of the most common it still has an annual incidence of ~100 thousand cases in US and Western Europe. Since the disease is often diagnosed at an advanced stage with such dismal prognosis, a drug with a survival benefit and a reasonable safety profile could attain a substantial market share. Using other next-gen chemotherapy drugs such as Abraxane and Doxil as benchmarks and assuming 5-6 month of treatment duration, TH-302 could generate $35k-$45k per patient. Even with a conservative penetration of 30%, TH-302 could generate $1B in sales every year in the US and Western Europe alone.
Earlier this year, Threshold sold worldwide rights for TH-302 to Merck-Serono (NYSE:MRK) for $25M upfront, $525M in milestones and double digit royalties on sales. Merck will pay 70% of worldwide development costs for TH-302 and will have marketing rights outside of the U.S. In the U.S., Threshold has the right to co-promote TH-302 with its own sales force.
Needless to say, had Threshold waited for the pancreatic cancer data, it could have gotten a much more lucrative deal. Therefore, the deal is a great achievement for Merck-Serono, which now has a promising asset with high chances of success in its pipeline.
Biotech portfolio updates
We are initiating a position in Threshold based on the positive pancreatic cancer data in anticipation for full data and initiation of a Phase III trial later this year. TH-302 is being evaluated in additional indications including a Phase III trial in soft-tissue sarcoma and several other Phase II studies. Even when assuming Threshold will not co-market TH-302 in the US, it could still generate peak royalties of ~$150 every year assuming a 10-15% royalty rate and limited market penetration in pancreatic cancer alone.
We are selling all of three Micromet (MITI) positions at a profit of 270%, 194% and 81%, respectively, following its acquisition by Amgen. Micromet was not only the largest holding in the portfolio but also one of my favorite companies. All that is left is congratulate the company and its management team for a remarkable achievement and hope that blinatumomab finds its way to the market as soon as possible.
We are also selling our position in Pharmacyclics (PCYC) at a profit of 413%. Pharmacyclics has one of the best drugs in development with high likelihood of being a multi-billion dollar franchise. Nevertheless, with a market cap of $1.8B and no substantial near term catalysts except initiation of Phase III trials, the company looks fairly priced.