Why This Fund Manager Is Short Ziopharm

NEW YORK ( TheStreet) -- Remember my fund manager source who nailed short calls on Elan ( ELN) (bapineuzumab), Clovis Oncology ( CLVS) (CO-101) and Celsion ( CLSN) (Thermodox), among other drug blow ups? Well, at the risk of instilling panic in many of you, he's short Ziopharm Oncology ( ZIOP) and expects to notch another victory when the palifosfamide sarcoma study comes up lame next week.

Before I explain the Ziopharm short thesis of my investor source, let's get a few things straight. I'm not telling you to sell your Ziopharm shares (if you're long today) nor am I advising you to short the stock. I'm not urging you to believe the short thesis. I'm not even endorsing the short thesis. In fact, I'm already on record predicting a positive result from the palifosfamide study and I'm sticking with that call. Although I'm suffering from a bout of the doubts.

What I am doing is giving you a head's up. A smart hedge fund investor (he's an MD, too) looms out there with a proven track record of correctly calling the failure of phase III clinical trials. He is prohibited by his fund's rule from being quoted by name. Now he has Ziopharm in his crosshairs. No one gets 100% of the biotech stock calls right, but it's always a good idea to know what the other side of the trade is thinking.

With that said, let's discuss his Ziopharm short thesis.

First, the obvious: He thinks the palifosfamide study will fail, meaning the combination of palifosfamide plus doxorubicin will not prolong progression-free survival compared to doxorubicin alone in first-line sarcoma patients.

Why will the study fall short on the primary endpoint? The answer lies in the nitty-gritty details of a previously conducted phase II study, which also pitted palifosfamide/doxorubicin against doxorubicin but in a mix of first- and second-line sarcoma patients. Results from this study were presented in 2010 and looked positive, with a 57% reduction in the risk of tumor progression (PFS) and a median PFS of 7.8 months for the palifosfamide combination versus 4.4 months for doxorubicin alone. The response rate for the treatment arm was 23% versus 9% for the control.

Those data are a mirage, my fund manager source believes. In reality, the palifosfamide/doxorubicin arm performed more like doxorubicin usually does on its own in previous sarcoma studies, while the doxorubicin arm under-performed historically. The phase III study, with more patients and more stringent measurements of response and tumor progression, is likely to come back with data showing the palifosfamide combination only marginally better than doxorubicin alone, if at all.

Are there data to support the bear thesis? As proof, my fund manager source points to the slide reprinted below, which depicts PFS from both arms of the phase II study. This slide can be found on page 17 of the protocol for the palifosfamide phase III study.

The blue line tracks PFS for the sarcoma patients in the control arm treated with doxorubicin alone. Focus on the sudden and steep drop in the curve between one and two months. Each red asterisk represents a patient who progressed. Between months one and two, eight patients in this control arm progressed very rapidly. That's unusual. Without those eight early and rapid tumor progressions, the two PFS curves would be much closer together. In fact, they might even overlap, which would mean no significant difference in PFS between the two treatment arms and no positive data for palifosfamide.

What might explain the eight early and rapid tumor progressions between months one and two in the trial? Well, it could be just the way the randomization worked out in the trial. Or, it could be a result of physician bias. The phase II study was open label, which means patients were aware if they were chosen to receive palifosfamide or not. The doctors of those patients knew, too. Obviously, no sarcoma patients entered this trial hoping to be placed in the control arm receiving doxorubicin alone. Doctors likely felt the same way for their patients, so when it came time to analyze scans, they could have been quick to diagnose a tumor progression so patients could switch to new treatments, including palifosfamide.

Unlike the phase II study, the phase III study is blinded to patients and doctors. No one will know who is receiving palifosfamide and who is not. This eliminates the potential for bias. Moreover, independent experts -- not the doctors treating patients -- will be reading and confirmed tumor progression scans.

That explains why the doxorubicin arm of the phase III study will likely perform better than what was observed in the phase II study. The palifosfamide-doxorubicin arm is also expected to perform worse.

Again, take a look at the slide representing PFS data from the phase II study. The red curve is the patients treated with the palifosfamide combination. The blue circles above the red line represent "censored" patients. These are patients who are excluded from the PFS analysis because they were either lost to follow up or had not completed enough treatment cycles to be included in the analysis. It's a bit hard to count, but there are either 15 or 16 blue circles (censored patients) in the palifosfamide-doxorubicin arm of the study before median (50%) PFS was reached. Thirty sarcoma patients were enrolled in the treatment arm, so 15-16 blue circles means half the patients were excluded from the analysis before median PFS was reached.

A little bit of patient censoring is relatively common in clinical trials. Censoring, or excluding, half the patients early in the trial (before median treatment effect is reached) is unusual and raises a red flag about the reliability of the data being reported. If those censored patients were included, the PFS reported for the palifosfamide-doxorubicin arm might have been significantly lower. In a larger trial conducted more rigorously, like Ziopharm's phase III study, it's reasonable to expect far fewer censored patients.

Early and frequent patient censoring was also a big problem for Peregrine Pharmaceuticals and the bavituximab second-line lung cancer study.

My fund manager source is further convinced the phase II study results aren't real because if you dig deep into the history of palifosfamide, which stretches all the way back to 1983 and its inventor Robert Struck, you find the drug's chemical structure is inherently unstable. Extrapolating data from ifosfamide papers like Santoro (Journal of Clinical Oncology, 1995) Palifosfamide doesn't penetrate cells well at all, so in order to have any tumor-killing effect, it must be administered at very high doses. Unfortunately, high doses can't be used because the drug causes kidney toxicity.

The palifosfamide dose used by Ziopharm in the phase II and phase III trials is ineffective. Proof? While supposedly increasing tumor response rate from 9% to 23% in the phase II trial, palifosfamide had no effect on other dividing cells such as leukocytes (white blood cells.) The incidence of leukopenia (a side effect of decreased white blood cells) was 17% in the palifosfamide-doxorubicin arm versus 16% in the doxorubicin arm. If palifosfamide was an active drug, the rate of hematological side effects such as leukopenia would be much higher.

Chemically and clinically, palifosfamide makes no sense. Therefore, my fund manager source believes the phase III study in sarcoma fails.

We'll know if he's right next week.

-- Reported by Adam Feuerstein in Boston.

Adam Feuerstein writes regularly for TheStreet. In keeping with company editorial policy, he doesn't own or short individual stocks, although he owns stock in TheStreet. He also doesn't invest in hedge funds or other private investment partnerships. Feuerstein appreciates your feedback; click here to send him an email.