Considering MyVax needed a p value of 0.0002 to get stopped for efficacy on the first interim look and 0.0026 for the second interim look, it's not surprising that the trial would be continued to the end, even if it were efficacious. What's interesting is that the stat sig p value for the final look isn't
0.05 minus 0.0026 minus 0.0002 = 0.0472.
The final p value needed for stat sig is supposedly 0.019, which means that the total allocated over the three unblindings is 0.0218. I guess the reason it's not 0.05 is the fact that the primary endpoint is not overall survival, but some sort of progression measurement like PFS or TTP.