Ziopharm Offers a Positive Risk/Reward Profile

Palifosfamide is a bifunctional alkylator developed by Ziopharm Oncology (NASDAQ:ZIOP) for the treatment of a number of cancers. Soft tissue sarcoma (STS) is the most advanced program with pivotal phase III data expected before the end of the year. Palifosfamide is ifosforamide mustard with tris(hydroxymethyl)aminomethane salt stabilization (earlier versions of the compound had lysine-stabilization). Normally this sort of detail is not critical but in this case it is important because ifosforamide mustard is the alkylaying agent that is created when the body metabolizes the prodrug ifosfamide. Ifosfamide is not new and has been shown to have robust anti-cancer properties but is unfortunately associated with a number of severe toxicities that limit its therapeutic window (see report for a more detailed discussion of this aspect). Even with the severe side effects, the Mayo Clinic lists 14 indications for which ifosfamide is used. As such, it is an attractive idea to create a drug with the effectiveness of ifosfamide without the toxicity.

Ifosfamide, the prodrug, is not necessarily toxic but is metabolized in the body to form the anti-cancer bifunctional alkylator, ifosforamide mustard, as well as a number of other metabolites. Two of the metabolites are quite toxic: acrolein (linked to urotoxicity) and chloroacetaldehyde (CAA- associated with both neurotoxicity and nephrotoxicity). The recently completed STS study by EORTC and presented at ESMO highlights just how much extra toxicity is created by adding ifosfamide to a treatment. A doxorubicin plus ifosfamide regimen greatly increased the adverse event profile compared to the doxorubicin arm. In particular, the rates of grade 3 or higher adverse events were much higher in the combo arm: leucopenia was 43.3% versus 17.9%, anemia was 34.9% versus 4.6%, thrombocytopenia was 33.5% versus 0.4%, and most importantly febrile neutropenia was 45.9% versus 13.5%. The toxicity of the combination lead to 17.6% of the patients discontinuing specifically because of the side effects were too severe and this is compared to 2.6% in the doxorubicin arm.

The toxicity of ifosfamide has created a smaller therapeutic window despite being active against many cancer cell lines. Since palifosfamide is a stabilized form of the active metabolite of ifosfamide, it should have similar efficacy without generating the toxic metabolites (acrolein and CAA). For instance, in early studies with lysine stabilized palifosfamide, it was found that rabbit kidney proximal tubule cells survived exposure to palifosfamide and these are precisely the types of cells that would die if exposed to acrolein or CAA. The larger phase II trial of palifosfamide plus doxorubicin (PICASSO) found that the combination arm did not have significantly more toxicity when compared to doxorubicin. In particular, there was no encephalopathy, hemorrhagic cystitis or Fanconi’s Syndrome and the rates of grade three or higher neutropenia and elevated creatinine was similar between the two arms of the trial.

Not only does palifosfamide have the expected decrease in toxicity but the PICASSO trial also clearly demonstrated efficacy. In terms of the progression free survival (PFS) endpoint palifosfamide plus doxorubicin produced a median PFS of 7.8 months as compared to 4.4 months for doxorubicin alone. The PFS hazard ratio (HR) was 0.43 and statistically significant. While the trial was not powered to show a statistically significant effect on overall survival, the trend was clear. Treatment with the combination had an overall survival HR of 0.78 (trial had crossover which makes it more difficult to discern the true treatment effect on overall survival and biases against the combination arm). In addition, the two year survival of the combination arm was 40% compared to 30% for the doxorubicin arm. In general, then, palifosfamide has shown both efficacy and a significantly better safety profile than ifosfamide and Ziopharm is expected to release PFS data from its pivotal phase III trial in soft tissue sarcoma (STS) by the end of 2012.

Despite the progress of palifosfamide and apparent advantage over ifosfamide, the presentation of the EORTC ifosfamide data at ESMO had a negative effect on the price of Ziopharm stock. This was an odd reaction given that the data actually had a bullish to neutral read-through for palifosfamide. The EORTC trial compared ifosfamide plus doxorubicin to doxorubicin alone in STS. In terms of efficacy, treatment with ifosfamide plus doxorubicin had a median PFS of 7.6 months compared to 4.6 months of the doxorubicin arm and this lead to a PFS hazard ratio of 0.74 (statistically significant). Of course, this efficacy of ifosfamide also came with the severe toxicities that were noted above (leucopenia, anemia, thrombocytopenia, and febrile neutropenia). Most concerning is the febrile neutropenia rate of 45.9% (compared to 13.5% in doxorubicin arm), which is perhaps one of the most dangerous side effects of chemotherapy (see article). Despite the high rates of adverse events compared to doxorubicin (note that the rates of neutropenia in the PICASSO trial were similar between the palifosfamide and doxorubicin combination and doxorubicin alone arms), the ifosfamide combination arm still managed to show an effect of PFS. This should be good news for a drug like palifosfamide that has the same active alkylating agent without the toxic metabolites but the market seemed to focus on the poor survival data from the trial.

While the ifosfamide combination produced an effect on PFS, the effect on median overall survival was quite modest (14.3 months compared to 12.8 months). In addition, one year survival rate for the combination was 60% compared to 51% for doxorubicin alone and this difference got worse over time with two year survival rates of 31% (ifosfamide and doxorubicin) compared to 28% (doxorubicin alone). Of course, focusing on the inability of adding ifosfamide to significantly improve survival rates is misleading for two key reasons. First, the adverse event profile of ifosfamide will hinder its ability to produce an overall survival effect as individuals drop out and do not get a complete course of treatment. In addition, adverse events such as febrile neutropenia are often life threatening and likely lead to early deaths. As such, there is little reason to expect that the difficulty ifosfamide had in increasing survival will directly translate to palifosfamide.

The second, and perhaps more immediately relevant, reason not to read too much into the EORTC overall survival data is that overall survival is not likely to be the endpoint that the FDA will use in evaluating palifosfamide. In 2011, four products were approved with a PFS endpoint and, more importantly, in 2012 pazopanib (trade name Votrient® and marketed by GSK [NYSE:GSK]) was approved with a PFS endpoint for the treatment of STS. While pazopanib was approved for second-line STS and palifosfamide is being evaluated in first-line, management at Ziopharm remains quite confident that PFS will be sufficient for approval (assuming a similar clinical benefit as seen in PICASSO 2). As such, there is little reason to see the EORTC ifosfamide trial results as a negative for palifosfamide and it is more likely that if there is a read through it is positive.

This note was drafted, however, before a Forbes article on Friday October 19 highlighted some potential concerns about Ziopharm. While this is not meant to be a response, it seems that the issues raised in the article are relevant for the arguments put forth in this article. The Forbes article focused on the reported protocol violations of Dr. Chawla and intimated that these put the PICASSO 2 trial results in question. This is a logical leap that does not seem justified. First, these reports, while more serious than a Form 438, are not rare (search FDA website to see number of warning letters) and given that Dr. Chawla is working in PICASSO 3, and additional STS drug programs, one could infer that the FDA was satisfied with the remedial measures. Second, and perhaps more importantly, none of the violations had anything to do with efficacy measures. It seems a stretch to take delayed reporting on non-efficacy measures from one clinical site and conclude that all efficacy measures from all sites are inaccurate. In fact, a JMP report issued October 22nd noted that an external review of the data by an ex-FDA monitor concluded that the protocol violations had no impact on the results. Finally, and perhaps most convincingly, Ziopharm management noted that even removing Dr. Chawla’s patients from the PICASSO 2 analysis would not materially alter the efficacy results.

Even if one still wants to argue that these violations affected the efficacy results of PICASSO 2, they also then have to explain the EORTC study. The implied claim in the Forbes article is that the efficacy data are not to be trusted. The EORTC study was completely separate from Ziopharm and PICASSO 2 but is in a similar population (EORTC patients are slightly younger because of ifosfamide side effects) with the same active agent (both palifosfamide and ifosfamide generate ifosforamide mustard). If the supposed clinical trial issues compromised PICASSO 2 results, then we would expect them to differ markedly from the EORTC study (as this would be free from the bias). Instead of that huge difference, the EORTC PFS results are consistent with the PICASSO 2 PFS results. This makes the bearish case more difficult to justify as they would not only need to show why PICASSO 2 is biased (which does not seem justified by the evidence) but also why the EORTC are incorrect as well. In addition, the bearish case would also need to explain away the overall survival trend seen in the PICASSO 2 results, which was not even addressed. As such, the bearish argument about efficacy does not seem supported by the evidence, especially in light of the EORTC results, the fact that removing Dr. Chawla’s treated patients does not affect efficacy in the PICASSO 2 results, and the OS trend favoring palifosfamide.

The other major issue raised was the lack of stability of palifosfamide and how this would translate into a poor commercial launch. Note that once again this argument has nothing to do with efficacy or why the upcoming data will not hit the endpoint. In many ways, the company has sufficiently addressed this in noting that over 300 clinical sites worldwide (ranging from academic centers to community hospitals) have used palifosfamide and there has not be an issue with the stability of the drug. These are the same sort of sites that would be using palifosfamide if approved, so if stability were a significant issue, they would have already been noticed. Of course, this is beside the point of this article, which is meant to examine the odds of palifosfamide hitting its PFS endpoint in its phase III trial. Overall, the risk/reward profile of Ziopharm was biased towards the bullish trade before the stock was hit by the Forbes article. The bearish efficacy argument explicated in that article does not seem consistent with either the PICASSO 2 or EORTC studies. As such, odds favor palifosfamide hitting its PFS endpoint when top line data are released by the end of 2012.