The only problem is
#1 the housing is not very thick, so tapping the threads will not be easy. You will be lucky to get maybe one or 2 threads.
#2 the maf works by using a simple ratio. the air that passes through the sample tube is x percent of the entire housing. With the sensor installed, it will change the entire housing's flow. Essentially, making the maf think it's getting more air than it is, and thusly a rich condition will occur.
#3 Since the computer doesn't calculate a/f at WOT and/or open loop, it will be affected. How much? I don't have a clue.
#4 You would really have to make a standoff of sorts, since the sensor has 1/2" of threads and most of the body of the threads would be inside the maf, creating a lot more turbulence.
All is possible when you set your mind to it. If you were serious about it, you need to make a standoff and tap threads onto it, and glue it to the maf; Only the "V" and the element need be in the airflow path. They make glue specifically for plastic, one that actually eats away at the plastic very slightly, making a very strong bond.
After it's installed, you need to get hooked up to a wideband and test to see how much it has affected your a/f. I can only guess, but if you put the sensor at the back of the maf, behind the sample tube but across from it, I would guess it would have the least effect there. Then the wake left over from the sensor would not affect the sample tube (so I think anyway).
Another idea would be to extend the pipe that the filter attaches to, so you have some area of pipe (non-visible) that you can attach the sensor to instead. I think that idea is a little better.
Scott