Replies to post #78189 on Biotech Values

[DewDiligence]

NVGN – In other words, they are unblinding prematurely without any agreement with the FDA to alter the statistical analysis plan:

The Company has decided to assess these data from the Ovature trial at this time, as the current downturn in the global financial markets makes raising further equity or debt in the near term to fund the trial through to completion most unlikely. According to the agreement reached with the FDA in a Special Protocol Assessment, the planned interim analysis of the data can be scheduled only after 95 events of disease progression and full recruitment of 340 patients.

The comments by the ex-CEO in the link you posted appear to confuse the number of patients who completed therapy at the Nov 2008 DSMB analysis — 78 — and the number of progression events at the time of the Nov 2008 analysis, which is surely a number less than 78. If the number of progression events in the Nov 2008 analysis was much less than 78, it’s hard to see how the number of events could now be as high as the 95 events the FDA wants for conducting the final analysis.

Thus, even if the FDA were to overlook the severe reduction in total enrollment relative to the trial size specified in the SPA, it would take a miracle for the PFS numbers to work out satisfactorily.

Why the heck has the share price been rising?

[iwfal]

So, what realistically are the odds of that being the case? To answer this, we need to return to the 13 November 2008 announcement which stated that the Independent Data Monitoring committee (IDMC) had reviewed the data from 78 patients who had “completed the study”, and had recommended continuation of the study. The IDMC is composed of oncologists, statisticians and radiologists that operates at arms-length from the Sponsor who remains blinded to the data. The primary task of the IDMC is to recommend to the Sponsor on the appropriateness or otherwise of continuing the study. This ‘STOP-GO’ power in the hands of the committee relates particularly to the ethics of continuing the study. A ‘STOP’ recommendation could arise in the event of serious safety concerns or a strong likelihood that the study was unlikely to reach its clinical objectives. And it is this last factor that is particularly relevant to our interpretation of what we might expect in 3 months or so.

We don’t know how many of the 78 patients were reviewed by the IDMC for disease progression, but given the Company statement about 78 ‘data sets’ being reviewed, it is a fair bet that most had their clinical status reviewed. And given the key role of this committee in advising the Company on the appropriateness of continuing to recruit patients to a study that may offer no benefit, and on the appropriateness of continuing to spend large sums of money on the study that may be pointless, it would seem highly likely that they would have carried out some degree of analysis of clinical benefit.

The reference point for the IDMC in terms of whether phenoxodiol was providing a benefit or not would be the 95 event data-set pre-agreed under the Accelerated Approval program. The basis of that data-set would be a difference between the two treatment groups that carried a certain level of statistical significance. If the IDMC believed on the basis of 78 events that the required level of benefit was unlikely to be reached after 95 events, then it would have been their responsibility to have alerted the Company to that situation and to have recommended discontinuation of the study.

The last sentence, which is the key sentence, is not only wrong but assininely so:

a) IDMC's generally don't do futility analyses unless asked to by the sponsor. And clearly, based upon the tempered words he used to describe the IDMC's analysis, the sponsor didn't ask for one.

b) Even if the IDMC choose to do a futility analysis they would have done it against the overall chance of success - not the interim chance of success (95 events is the interim as per elsewhere in the speech). And for most trials the vast majority of the power is in the final look, not the interim look.

c) Futility is often recommended to be judged as 20% or less chance of success of obtaining a successful trial if the efficacy assumed in designing the trial applies to the remaining patients. I.e. you can have only a 25% chance of success and the trial will go on. Combining this with #b and EVEN IF THEY RAN A FUTILITY AT 78 EVENTS THE TRIAL COULD SHOWN A NULL RESULT AT THE 78 EVENTS AND STILL HAVE PASSED.

It is distressing that an ex-CEO should make this kind of gaffe - which is, in fact, more extreme than that seen on most rah, rah message boards.

[DewDiligence]

GNVC released the TNFerade KM curves for the first interim look:



The complete poster presentation is at:
http://www.genvec.com/download/2009%205.29%20ASCO%20Chang%20et%20al%20PACT%20Interim%20Anal%20Poster%20FINAL%20OUT.pdf