The answer to your question is called "subgroup analysis." In designing the trial Mgmt was already confident that Bavi works-- so they included PS 2 patients in order to get a reading on just how well it works.
In the pancreatic trial, the randomization skewed against the treatment arm and affected the overall result. In hindsight, perhaps that was too high a cost to pay for this additional level of knowledge on how well Bavi works (or doesn't work) in extremely sick death-row PC patients.
In the 2nd line NSCLC trial, the randomization seems to have worked quite well. 75.6% of the patients were PS 0 or 1. That gives them a nice sub-group of 30 patients to see how well Bavi works in these healthier patients.
This is why the Company emphasized in the opening paragraph of today's PR that "Peregrine plans to report additional data from the trial, including updated subgroup analysis and safety data, at an upcoming scientific meeting.
The subgroup data reported at AACR in early April or t ASCO in late May should be very interesting.