Wouldn't injector PW be after fuel trims are added? If so, then I don't see why you couldn't do it in closed loop. You still know the fuel mass (inj pw * flow rate) and you know actual lambda/AFR for that fuel, so with that you can calculate how much air there must have been. From that, GMVE.
I could be wrong about the reported PW being the "trimmed" value but I don't believe I am