The only way I see it adding that much HP is if it is purposely mis/under-calibrated in the upper regions where WOT is. In that case, it would act like a performance tune, by leaning out the AFR, which gains some power. But a good portion of power gains from tuning are from timing increases too. And if it does work like that, and you also get a good PCM tune, you'd end up with the WOT AFR way too lean.
So, I'd say it's like the others here are saying, it's probably a rigged test. Like they did several pulls before and after and used the ones that showed the biggest spread between the two. Or, if they changed the air filter before the MAF change, obviously that would make a difference.
We need to think of the MAF sensor for what it is: a sensor. We don't try to "upgrade" the MAP sensor, the IAT sensor, or anything else. It's supposed to sense something, it's calibrated. We wouldn't port out our thermostats so that they could flow better, would we?
Now, if you have access to data-logging/wideband and can do a lot of fine-tuning, then you may see some ultimate airflow/HP gains by significantly porting the MAF and recalibrating it completely across its spectrum.
Just my 2c!