Replies to post #52090 on Avid Bioservices Inc (CDMO)

[cjgaddy]

Ah Yes, EntDoc, the 'ole "Non-Evaluables" dilemma... Recall, early on Peregrine was reporting Bavi interim Ph.2 ORR %'s on an "Evaluables-Only" basis, and then around Summer-Fall'09, switched to an Intent-to-Treat (ITT) basis, I believe driven by RECIST 1.1 recommendations. Non-Evaluables exist, as you say, for a multitude of (non-drug-related) reasons (ex: no-shows), and RECIST 1.1 says (properly so) you must treat them all as trial Failures for ITT stats. There were 9 Non-Evaluables in the Ph.2 Bavi+CP/NSCLC trial, so that the 21 Obj.Responses were divided by 49 (43% ORR, ITT), rather than 21/40 = 53% on an Evaluables-Only basis. Ashamed to have to treat all 9 of the Non-Eval’s as Bavi ITT ‘Failures’, but that’s the way RECIST 1.1 works. At least it works that way for everybody else’s Ph.2 trials too.



Re: Figure1: Note there are 40 BARS shown – that’s one for EVERY “Evaluable” patient in the trial – recall, Stage1/n=21 had 4 non-evaluables, so Stage2/n=28 obviously had 5. They took each of those Patient’s “BEST RESPONSE”, sorted them Best-to-Worse, and showed them in the Fig1 graph.