Avid Bioservices Inc (CDMO)

[Protector]

nh, thanks. The rd part of the text shows that PPHM did start from the assumption that Docetaxel would be expected at 10.4 months (which was the max from Studies at that point, Herbst et al 2010 at 9.9 months in second place). So my 10.4 months was not speculation but an educated assumption and the statement of CEO King confirms it, I forgot it was even in a transcription.

Steve King: I can start off and maybe Joe can add in, but when we designed the SUNRISE study we did take into account the variability that had been seen in the docetaxel studies that had been reported up to that point. I don’t think we’ve seen any really significant variations from that as over time as more readouts have come. So, that’s one thing that we did take into account, that we were toward the upper end of what have been previously reported for docetaxel. Secondly, when we powered the study, in the Ph2 study we saw about a 4mos. difference in median OS, and the SUNRISE study was designed to really show a statistical difference even at 2mos; we built in some powering assumptions we think will give us some opportunity there should be arms behave in a way that was little bit unexpected. So, we’ve tried to hedge against that going into the stage design and we think we’re in good shape and now it’s just a matter of getting the readouts - we’re little under 2yrs right now from when we started to study, so at this point there’s really nothing more we can learn from the ongoing operations of the SUNRISE trial itself.

Add to that that in a next CC PPHM has said they included some MARGIN in the CTRL arm expectations. So assuming 1.6 months (that would be a 6 week margin) the expectations for the control arm Docetaxel alone would be 12 months. And as CEO King says above, SUNRISE was designed for bavituximab to show STATISTICAL DIFFERENCE at 2 months (that means not only a difference - improvement - by two months but also in a statistical significant way) 12+2=14 months.

Therefore for CEO King to say in the special SUNRISE PR/CC that Bavituximab performed according expectations it must have been AT LEAST at 14 months otherwise that statement would be a lie in the above setting And this supports that to be a realistic expectation even with much sicker patients.

For the control arm to DRAMATICALLY outperform the expectations it must have been 12+x months where x must have been so large that the Independent Data Monitoring Commission (IDMC) could NOT give any other advice as to stop SUNRISE because they expected it to be impossible for the Bavi arm to ADD 2 months to 12+x. In PII the

Joe Shan: George, obviously it's a very dynamic field right now. As Steve mentioned, we tried to make the SUNRISE design in such a way that it's as homogenous as practical. We obviously have some stratification criteria built in and some preplanned, subgroup analyses that are going to be pre-specified. Probably the biggest change in the landscape since we started enrollment is the approval of checkpoint inhibitors in this space, and we probably have patients that have received checkpoint therapy, and if there's an imbalance between the 2 arms of the study the subgroup analysis should account for that. But, it’s something that’s really impossible to predict how the changing SOC is going to effect the results of the control arms - this is why we have to run double-blind. Probably one predictor would be if we reach the number of interim events necessary to trigger the interim analyses, and the general projections that can give us some a little bit of confidence that these aren't too different than what we have previously seen.

IMPORTANT
There is a BIG difference between an early stopped trial for futility or efficacy (not including safety) and a failed trial. A failed trial is one were the FDA refuses approval based on the data that the sponsor provides in the BLA or where the sponsor decides to stop all further activity. This is NOT the case here, PPHM clearly claimed they were going to look into things and TEMPORARY stop other CHEMO trials.

Shan's statement (red text) about sub groups and statistics and his explicit example being the approval of checkpoint inhibitors (just say Opdivo/Keytruda) go hand in hand with Dr. Garnick statement that they had some guarantees related to the approvals that would impact SUNRISE after it started.

SUNRISE satisfies all the conditions to file for BLA:

1) All, even MORE, patients as required by the FDA for final unblinding were enrolled already at least on NOV 2015 because PPHM reported on it DEC 8th 2015 during the Q/CC.

2) The # patients involved in the first look-in is 77 ctrl+77 bavi=154 which is MORE then the PII that had 3 arms. We know Bavi 1mg+Bavi 3mg combined where statistical significant on 41+40+40=121 patients and that was INCLUDING the PII sabotage in the DISADVANTAGE of Bavituximab.

3) The results of the SUNRISE FIRST look-in are obtained FULLY "double-blinded" which means PPHM has 154 VALID results and in an early stop they can use them for a BLA (just as if the trial would have been stopped for efficacy, there also one would file the BLA with only 154 patients results obtained in a double blinded way).

4) Since Bavituximab performed as EXPECTED (that means it showed a statistical difference with the CONTROL ARM expectations (not with the CURRENT control arm results) and since proactive provisions where taken to correct exceptional performances of the control arm by SUB-GROUP analysis and hence PPHM can bring the dramatically outperforming control arm to the correct proportions, PPHM could have statistical significant data showing bavi outperformed the ctrl arm as needed which would be accepted by the FDA since it was PRE-DEFINED.

5) Remember that the PII showed a p-value of 0.113. p-value for the combined CONTROL arm which is not statistical significant (the 1 + 3 mg combined were). To reach p=0.02 for the STUDY a p-value of 0.176 would suffice for SUNRISE to make the study (not this individual trial) statistically significant.