A recent article in the Street.com about Introgen (NASDAQ:INGN) and its lead drug candidate, Advexin, focused on a tumor-like mass of evidence demonstrating the apparent worthlessness of Advexin’s clinical trial data and by implication the company itself. The article also peripherally addresses Introgen’s management team.
In many regards, however, this management team is more problematic than Advexin itself. Unfortunately, the pursuit of absurd science by biotech is not a crime. Nonetheless, one has to wonder how a company with two completely dead phase 3 trials that have not enrolled a single patient for the last several years is spending $20 million per year on R&D. In fact, it is inconceivable that the studies referenced in the company’s almost-daily press releases referencing preclinical data are sufficient to burn this amount of capital.
So where’s it going? No one knows.
As mentioned in the article, Introgen appears to have a fiercely dedicated following. While this is laudable in its own right, the absence of institutional involvement on the long side of the Introgen trade should cause a great deal of concern. Small investors can rest assured that the institutions that are short know more than them because that’s their job. And while there are large institutions on the long side, the largest (Fidelity) doesn’t count. Fidelity owns everything and their INGN position represents 0.0007% of their entire portfolio. The bottom line is that it is not a question of if individual investors are going to lose 90% of their investment; it’s a question of when!
In return for waiting out INGN’s and/or Advexin’s death, the stock does nothing as evidenced by investors’ gross return of 0% in five years and the accompanying dilution that comes like clockwork after each incremental $25 million is burned. While not losing a lot of money investing in biotech is a good thing, the number of alternative investments in the sector suggests that leaving money in INGN is a fool’s game.
And Finally . . . The “Data”.
The dead Advexin pivotal monotherapy trial in recurrent, refractory, unresectable SCCHN (‘301) is an approximately 240 patient trial comparing Advexin to standard chemotherapy (methotrexate) after surgery and radiation have been ineffective (the dose in this trial will replicate the high-dose in one of three phase 2 trials). The primary endpoint is survival and the secondary endpoint is time to progression. To date, there has been no disclosure whatsoever with regard to the phase 2 program related to time to progression although the measure was a secondary endpoint in two of the three phase 2 trials and appears to be the only surrogate endpoint viable for the accelerated approval. The ‘301 trial started enrolling around May 2001 and evidence strongly suggests that the trial stopped enrolling a long time ago.
The dead Advexin combination trial (‘302) is an approximately 255 patient trial (although it was originally described as a 288-patient trial) comparing Advexin in combination with standard chemotherapy (cisplatin+5Fu) to standard chemotherapy alone in chemotherapy naïve patients. The primary endpoint is time to progression and the secondary endpoint is survival. The 495-patient phase 3 pivotal trials (expected enrollment as of 5+ years ago) were designed based in part on an analysis of a subset of patients in one of three phase 2 trials (referred to as T-201 – the “high dose” phase 2 trial). This is what phase 2 trials are for.
In contrast, Introgen now believes they are suitable for actual registration. The subset was comprised of the first 88 patients treated and evaluated (107 were treated in total). Reported median survival time for a subset of this subset of 88 patients was 7.5 months (220 days). The subset is apparently characterized primarily as those with refractory disease and a tumor size threshold of <7.5cm but the number of patients in the subset has not been disclosed and the exact characteristics of the subset have not been adequately described for reasons that are entirely unclear to us (INGN, however, cites patent concerns).
This subset apparently didn’t respond as expected so the new subsets are based on some undisclosed biomarker. IINGN has reported that this 7.5 month MST compares to an investigator-estimated expected MST of 4 months, citing an 88% benefit in the subgroup of the subgroup evaluated. As way of comparison, Imclone’s Erbitux has been studied extensively in SCCHN. A phase 2 trial in platinum-refractory recurrent disease demonstrated an MST of 5.5 months as monotherapy as compared to the above-noted historical controls in platinum refractory patients of 3.5 months.
The latter is still inferior to the results reported by Introgen but were based on a prospectively designed trial as opposed to being the result of combing through data until something of interest was found. Besides this trial, two other trials of Erbitux in combination with platinum compounds in platinum-refractory SCCHN patients reported objective responses in 12% and 15% of patients and an MST of approximately 6 months.
Iressa has also been studied in refractory recurrent SCCHN. In a phase 2 study, median survival was extended over the expected 6 months for that patient population to a reported 11 months. As noted, INGN designed and initiated its pivotal phase 3 trials in part based on data scoured from the above referenced subset of 88 patients from the T-201 trial. As things stand today, the T-201 trial is retrospectively defined subset of the three-trial phase 2 program to the extent that INGN has regularly referred to the aggregate data from the phase 2 program as essentially one registration trial in terms of its data analysis. INGN has been cherry-picking this data in order to determine what questions to ask based on the answers generated and surprisingly some of the most important answers, specifically those related to ORR, have actually changed.
The primary efficacy measure in two of the phase 2 trials (T-201 and T-202) was objective response rate yet ORR for the intent-to-treat group has been reported for neither of these trials (it was not a primary efficacy variable in the third trial (T-207). Nevertheless, INGN reported in March 2004 that the “primary endpoint results” from the three phase 2 trials confirm its “previously reported” findings regarding the activity of Advexin. The primary endpoint results that INGN “previously reported” and appears to be referring to (specifically, tumor growth control) have not been previously reported by Introgen from any of these three phase 2 trials except for preliminary analyses and/or selected post hoc subgroups – there has been no disclosure, ever, with regard to intent to treat populations. Having said that, a National Cancer Institute website states that the objective response rate in the three phase 2 trials
was 4% and stable disease was observed in 22% of patients (145 evaluable patients).
This is pathetic and is consistent with the 8% ORR observed in the first 88 patients treated and evaluable in the T-201 trial and the 25% stable disease rate reported for these patients. However, it is inconsistent with the 15% ORR that INGN has previously reported for a select group of unidentified patients that Advexin luckily appeared to benefit that has recently and inexplicably been increased to 20% and even at the new, higher figure is inferior to the reported 23% ORR in a prospectively designed Erbitux phase 3 trial that was analyzed on the basis of a prespecified statistical plan on an intent-to-treat or prespecified per protocol basis.
Perhaps “inexplicably” is an inaccurate description inasmuch as INGN management declared on a May 10, 2005 conference call that the “analysis” which resulted in an increase in the ORR to 20% from 15% is something that they are “conducting” and “defining” and the figure is “approximate”. There is clearly a possibility that when INGN is done “conducting” and “defining” that analysis that ORR will be even higher.
Also of interest is the fact that tumor growth control was not the primary endpoint of any of the three phase 2 trials. In T-201 and T-202, ORR of the treated lesions was the primary endpoint (the difference appears to be that that tumor growth control also incorporates stable disease). As for T-207, which was a biodistribution and transmission study, no data has been reported other than the MST of an unidentified subgroup of unidentified scope that received chemotherapy after the fact and likely received it outside of the scope of the protocol.
MST's for both T-201 and T-202 were also reported for this de facto patient population in March 2004. In something I have never seen in the annals of reporting clinical trial results, INGN refers to the already preposterous interpretation of the data from these three trials as a “multi-center, multi-national, analysis” as if the global scope of the analysis validates the selective and highly biased post hoc conclusions. Aside from the fact that tumor growth control was not the primary objective of these studies, the clinical import of these studies based on “tumor growth control” is particularly suspicious and the endpoint is not an approvable one anyway, particularly on an accelerated basis.
In fact, the protocol summary for T-301 itself speaks at length about the questionable correlation between clinical benefit and “stable disease” (i.e., tumor growth control). Although the summary does not entirely dismiss the notion that tumor growth control is indicative of clinical benefit, it clearly states that traditional tumor size measurements are inefficient in reflecting anti-tumor activity and refers to the measure as a “not fully validated surrogate for clinical benefit” for the following reasons: (a) measurements do not consistently reproduce the actual tumor changes and (b) tumor size measurements [in the Advexin phase 2 trials] were inaccurate because of the nature of the tumor and the fact that prior surgical resection and irradiation sequelae commonly rendered tumor limits imprecise.
In other words, the 59% tumor growth control rate evidenced in the combined phase 2 trials that will apparently be used in part as support for accelerated approval based on its surrogacy for clinical benefit were dismissed by whoever wrote the T-301 protocol sometime during the summer of 1999. Moreover, the 59% tumor growth control rate cited appears refer!
s to treated and responsive lesions as opposed to treated patients.
The company’s 2002 10-K reveals that the preliminary analysis of 88 patients in the T-201 included patients who had multiple tumors treated. If efficacy was measured as the number of patients responding the figure appears to be 25%. When efficacy was measured as the number of patients who had at least one tumor responding the figure appears to be twice as large at 60%. Again, the company appears to have created the opportunity to take multiple shots on goal on a de facto basis. It is not unreasonable to ask what the clinical value of a patient who has one tumor responding while one or more of the patient’s other tumors don’t respond?
The answer is zero.In an FDA review of an experimental drug for SCCHN in 2001 (a formulation of collagen, cisplatin, and epinephrine in the form of a gel directly injected in head and neck lesions) the FDA noted the following:
“Although a lesion may decrease in size or completely disappear, clinical benefit from such shrinkage cannot necessarily be assumed. The findings have implications for future trials of local therapy in the refractory/relapsed head and neck squamous cell carcinoma setting.”
This comment was made notwithstanding the fact that the FDA’s preliminary analysis of the data indicated that objective response rates would ultimately work out to be between 19% and 27% in one of the phase 3 trials and 16% to 21% in the other phase 3 trial. These figures were actually the result of the FDA’s dissection of the sponsor’s reported response rates which were 34% and 25%, respectively, for each trial compared to the ranges above, which is a type of dissection I suspect could likewise be applied to both the phase 2 and phase 3 Advexin data.
Nonetheless, even the dissected and reduced figures cited above are significantly higher th!
an the 10% ORR claimed for Advexin in a presentation at the 2004 AACR of a meta-analysis and higher than both the 15% claimed by the company in a 2004 press release describing a subset of the phase 2 trial and the more recently enhanced 20% ORR. At the risk over-analyzing the FDA’s position on clinical benefit, I think the following comments from the FDA on this NDA are critical:
“In the mid-1990s, FDA and the Applicant held numerous meetings to discuss how one might design a trial to document the clinical benefit associated with local treatment of incurable head and neck cancer. FDA's position was that shrinkage of tumors from local injection of chemotherapy was not sufficient evidence of clinical benefit. Any treatment, including local treatment, was likely to produce some side effects, and nobody had proved that temporary shrinking of head and neck cancer yielded any benefit that the patient could perceive. One of the difficulties encountered was that head and neck cancer patients suffer many different problems, so that it would be impossible to accrue enough patients to evaluate treatment of each problem separately.”
The likelihood of FDA approval based on this data or even the acceptance of a rolling BLA based on the data verge on the ludicrous. In another historical oddity, INGN announced that it had submitted the first portion of what it intends to be a rolling BLA in December 2004 even though approval of the company’s request for a rolling BLA has not been granted by the FDA. There have been suggestions that INGN has dredged enough data from the phase 2 studies to isolate a link between response rate and survival and between stable disease and survival. There is not a shred of evidence that this is the case and the FDA comments cited above argue against this possibility.
INGN frequently cites the Code of Federal Regulations when describing the potential regulatory pathway available for Advexin citing, for example, Subpart E (21 CFR 601.31), which permits accelerated approval based on drugs and biologics that meet unmet medical needs and based upon surrogate endpoints that are reasonably likely to demonstrate clinical benefit. What INGN doesn’t mention is the requirement under Subpart E that marketing approval even under these relaxed standards is on the basis of adequate and well-controlled clinical trials establishing an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit. Regardless of whether a surrogate endpoint is deemed sufficient for approval of a drug or biologic, the FDA always operates under the premise that studies relied on to support effectiveness meet the requirements for adequate and well-controlled studies in 21 CFR 314.126, regardless of whether those studies are the basis for approval on a surrogate endpoint.
The FDA further notes that reliance on substantiation of efficacy from related trial data leaves little room for study imperfections or non-supportive information and the results must always reflect a clearly established prior hypothesis documented in a protocol. Investors must ask themselves whether the Advexin phase 2 trials meet this standard. They don’t.
These regulations are no secret in the world of biotech but may be instructive for investors to also consider INGN’s strategy in light of ICH guidelines relating to principles for clinical trials. These principles are supportive of the view that the likelihood of a BLA approval based on the phase 2 program for Advexin is unlikely. Ironically, a number of these comments were quoted in 2001 by FDA reviewers of Matrix Pharmaceutical’s NDA for approval of IntraDose for squamous cell carcinoma of the head and neck, which was supported by two pivotal trials and two phase 2 trials.
As noted above, IntraDose was a combination of two already approved drugs, cisplatin and epinephrine, and cisplatin has well-documented cytotoxic activity in SCCHN. The drug was nonetheless rejected partially due to safety issues but primarily due to efficacy issues and questionable trial design and endpoints. So ask yourself: What is the hypothesis upon which the Advexin BLA is based? It’s not pre-defined inasmuch as INGN management openly stated in its recent May 2005 conference call that “part of the value of these subpopulation analyses of course is to better educate us on asking the most appropriate question as databases are open.” INGN itself has indicated that it finds value in the opportunity to basically collect data first and ask questions later. The only p-value ever mentioned in reference to the phase 2 programs relates to the difference in MST of the high-dose and low-dose trials.
Aside from the fact that these trials were not distinct, the p-value supports the null hypothesis – i.e., there is no effect. The meta-analysis of the three phase 2 trials, notwithstanding the fact that it was a “multi-national analysis” did not have its own prospectively written protocol. INGN’s comments confirm this fact inasmuch as they frequently refer to evolving analyses. Aside from these guidelines, the FDA also noted in the above-referenced NDA review that it had told the sponsor that “Clinical benefit and response analyses should be based on the intent to treat population with all other analyses playing a supportive role, subset analysis for patients who are considered incurable with surgery should be done, and prior therapies must be extensively described to document refractoriness of disease.” Pivotal Trial Enrollment INGN has not revealed and has indicated its intention not to reveal in the future the status of enrollment in either of the pivotal phase 3 trials, allegedly due to competitive concerns.
I do not buy into this argument and have heard it many times in the past from early stage biotech companies. The T-301 monotherapy trial was announced in March 2000 while the intent to initiate the T-302 trial was announced in late July 2002. At the time, Introgen announced that expected completion of enrollment in both trials would occur simultaneously at the end of 2003. Introgen also stated that interim analyses would be conducted for both trials at the 50% accrual point.
However, other comments have suggested that the ‘301 trial could be analyzed on an interim basis at 109 deaths, i.e., the interim look would be event driven as opposed to accrual driven. No interim data has been announced or released to date raising the possibility that neither trial has reached 50% accrual. In fact, in June 2004 INGN disclosed that it expected the interim analysis triggering event for the T-301 trial to occur over the next year – the next year is now here and there is no word on an interim analysis.
In other words, the ‘301 trial, even though commencement was announced in March 2000, has likely not reached the 50% mark (this assumption is fairly safe given MST’s of 7.5 months seen in the phase 2 trials in an apparently very tiny subset of patients). The facts appear consistent with all attempts to estimate patient accrual and consistent with the company’s late 2004 decision to file based on phase 2 data. While it is not possible to accurately back into the enrollment status in the SCCHN pivotal trials anyone’s best efforts to do so raise disturbing possibilities.
INGN 1-yr chart
Disclosure: Author has no position in INGN.