It's always flattering to be called out by a company for an article you've written, so we were pleased to see Celsion (NASDAQ:CLSN) describe our last article (which can be read here) as "misleading" and that our statements "require clarification" in a press release that can be read here. If anything, we're pleased to report that a close reading of Celsion's press release suggests that our statements were spot on. Our key takeaways from the press release, none of which investors previously knew, were:
- A change from Celsion claiming that they would meet with the FDA to now claiming that an FDA meeting would be dependent on a maturation of OS data.
- An admission from Celsion that they may need to pursue additional animal testing before they can restart clinical tests in humans. This is an incremental delay and therefore negative.
- A new concept from Celsion of pursuing additional trials in enriched populations, which we believe is a further negative due to the potential for a smaller target market and a longer time to market.
We believe that all of this new information is incrementally negative. Our price target remains $0.73 or roughly -50% from today's price.
With those as the highlights, let's analyze Celsion's recent press release, which oddly enough never says what was misleading about our last article:
Celsion: As is common in the biotechnology and pharmaceutical industry, Celsion has conducted a comprehensive post hoc analysis of the data from its Phase III HEAT Study of ThermoDox® in hepatocellular carcinoma (HCC) with its key principal investigators, data experts and liver cancer experts.
Alpha Exposure: We agree. It's entirely normal for companies to data mine after failed trials. In our opinion, this is typically fruitless because if you look in enough different places you can generally find something that looks statistically promising simply due to random chance.
Celsion: This analysis followed the announcement on January 31, 2013, that ThermoDox® in combination with radiofrequency ablation (RFA) did not meet the study's primary endpoint.
Alpha Exposure: Exactly! The HEAT study was a failure. In fact, by the company's own admission, it wasn't even close.
Celsion: Emerging findings from the HEAT Study post hoc analysis, which has provided important insights regarding the use of ThermoDox® in conjunction with RFA to treat HCC, suggests that ThermoDox® markedly improves progression-free survival (PFS) and overall survival (OS) in patients if their lesions undergo RFA for 45 minutes or more.
Alpha Exposure: We think the conclusion that Thermodox plus RFA for 45 minutes or more improves PFS and OS is unlikely to be true. The conclusions are being drawn from data mining and there is no way to know whether it's due to randomness or due to a treatment effect. In fact, this is why the FDA approves drugs based on meeting pre-specified endpoints in trials and not on post hoc analysis after data mining. Lastly, it is important to note that regardless of whether or not Celsion thinks that there was an improvement in PFS and OS, any benefit seen was NOT statistically significant. As a result, even if the FDA would give weight to these results (and we are confident they would not), the results still would not be good enough for approval of Thermodox.
Celsion: These findings apply to HCC lesions from both size cohorts of the HEAT Study (3-5 cm and 5-7 cm) and represent a sizable subgroup of approximately 300 patients.
Alpha Exposure: 300 Patients is less than half the trial size, which enrolled 701 patients (see the boilerplate here). This means that more than half of the patients enrolled in the trial are not even part of this subgroup. In fact, if you look at the slide for the patients undergoing RFA for less than 45 minutes, the hazard ratio for Thermodox treated patients was 1.34, suggesting that there was a trend of patients in this subgroup who had been treated with Thermodox dying faster than the placebo patients. Of course, this patient group isn't statistically significant either, so unlike Celsion, we won't try to mislead people into believing that there is something there. More importantly, as previously stressed, even in their chosen subgroup of patients receiving RFA for more than 45 minutes, Celsion was unable to find a statistically significant result. In our opinion, this speaks to the lack of efficacy of Thermodox.
Celsion: In the patient subgroup treated in the ThermoDox arm whose RFA procedure lasted longer than 45 minutes (63% of single lesion patients), overall survival improved by 53% (Hazard Ratio of 0.65 and a Pvalue = 0.105) when compared to the control arm of RFA treatment only.
Alpha Exposure: Once again, Celsion does our work for us by pointing out that even after extensive data mining (which they refer to as "post-hoc analysis"); the company is unable to find a patient group that has a statistically significant benefit from Thermodox.
Celsion: While the Hazard Ratio reported above should be viewed with caution since the HEAT Study has not reached its median point for overall survival analysis, there is a strong signal which our investigators and data experts consider to be encouraging and sufficient to warrant additional clinical investigation.
Alpha Exposure: Celsion further helps investors understand that its data mining is of minimal value, because the data mining itself is based on an immature data set. Furthermore, it is our opinion that "investigators and data experts" almost always want additional clinical investigations. These investigations are both lucrative and prestigious for them, so it's unsurprising that they're encouraging the company to spend more money on pursuing Thermodox, which has already failed a Phase III trial.
Celsion: These data were presented at a recent medical meeting of interventional oncology by two of Celsion's lead investigators. Furthermore, the detailed data has been submitted for peer review at upcoming international meetings in liver oncology and interventional oncology. HEAT Study investigators have agreed to the submission of the HEAT Study results and the post hoc data for peer review.
Alpha Exposure: The presentation of failed trials at conferences and in scientific journals happens all the time. This is one of the ways that scientists learn and share data. It's no reason for Celsion to celebrate and it's in fact the data that they've presented that has been so poorly received by the experts we have consulted.
Celsion: Celsion is refining the appropriate theoretical models to establish quantitative predictions for the influence of RFA heating time on local tissue concentrations of ThermoDox®. Empirical studies will then be used to verify the model predictions using appropriate animal studies which are targeted for completion over the summer.
Alpha Exposure: Celsion is basically admitting that it needs to go all the way back to animal models with Thermodox to design a new trial. In fact, the prior animal models suggested an optimal release of Dox within about 30 minutes, so the 45 minute cut-off Celsion has been using seems arbitrary and most likely chosen to put the failed HEAT trial in the best light possible. In our opinion, if Celsion needs to go back to animal models before it can even design a new trial, the timeline to new results is likely to be even longer than we previously estimated. Shareholders should be ready for massive additional dilution in the meantime since we estimate that Celsion will need at least twice as much cash as it currently has to run another trial.
Celsion: Assuming that our hypothesis is supported by non-clinical studies as well as continued follow-up of patients in the HEAT Study to the overall survival endpoint, we will meet with the U.S. Food and Drug Administration to discuss these findings.
Alpha Exposure: Celsion takes another step backwards with this statement. In the past, investors were led to believe that Celsion would meet with the FDA to plan next steps when the CEO said in the April 23 business update "Based on the strength of these findings, the Company will request a meeting with regulatory authorities for guidance on moving forward with our HCC development program." Now, investors know that such a meeting with the FDA will only take place if both the animal models Celsion needs to go back to as well as the still immature OS data are positive. This is an even riskier proposal than investors were previously led to believe. Lastly, as investors wait for this data to be generated, Celsion will continue to burn through cash and the timeline to starting (much less completing) another Phase III trial will be even further extended. This is a clear negative revision to investor expectations.
Celsion: We would plan to propose a regulatory path forward which may include additional clinical studies based on FDA's predictive enrichment guidance.
Alpha Exposure: We agree that Celsion will need to conduct additional clinical trials. As regards "predictive enrichment guidance," we believe that Celsion may be referring to choosing a "study population in which detection of a drug effect (if one is in fact present) is more likely than it would be in an unselected population" - which is described in this document. This is an interesting idea by Celsion, but it carries with it two significant drawbacks. First, in the unlikely event that the company ever runs a successful Phase III, any drug approved on a trial that succeeds based on an enriched patient population is likely to receive a label where the use of the drug is indicated for the population in which it was studied. As a result, this may cause the hypothetical future market size to be much smaller than it otherwise would be. A second drawback is that enrolling a large trial in an enriched population is likely to take much longer than it would to enroll a general population of HCC patients who need RFA. Hence the timeline to study completion would be even longer.
The single most interesting thing about Celsion's press release is that it claims our prior article was "misleading" and that our statements "require clarification," yet there was not a single quote or piece of analysis from our article that was refuted by Celsion. Instead, our careful analysis of Celsion's desperate press release reveals many more reasons that the company's future is bleak. We continue to see Thermodox as an abject failure and believe that the Company's post hoc data mining analysis was both misleading and useless. In our opinion, Thermodox has no future in HCC and the fair value for Celsion remains $0.73 per share or nearly 50% below today's price.
For our next article, we will provide an explanation of post hoc analysis and data mining along with a mathematical example that shows why this type of analysis is more likely to be random (or even misleading) than meaningful.
Disclosure: I am short CLSN. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.