[Fasterthanyou]

Yeah, that baseplate isn't what you'd call bang for the buck.
For a cam, Lt4 hotcam will work but why not a zz3/zz4 cam? The Lt4 hotcam would make a bit more power but not by a staggering amount. Really depends on exactly how much compromise. You'll get over 400hp with the zz3/zz4 cam and they can be had for just over $100 on ebay brand new in box. Not to mention the exhaust lift is right at .510" which might (or might not) be the max safe lift out of mild machine work to vortecs.
As for the MAF, it isn't a big restriction but it is one non the less. What's more restrictive is the code in the ecm. I'm assuming since you have MAF and TPI you're using the 1227165 ecm, that runs on $6E code which has been known to give people difficulty. At least the people that drive the cars often and know what it SHOULD drive like... (skidownit doens't know the feeling). The difficulty with $6E code is it's spark advance table. It isn't based on just RPM and Airflow like you would think (or want). Instead it's based on a load variable called LV8. This variable moves all over the place and overlaps with some areas that you want a discrepancy with timing. I could go into more detail but it might get confusing. So my vote is to run $32 speed density code in the 1227165 ecm. It doens't require much effort to run the code, just remove the MAF and wire in the MAP sensor. The only downside to running $32 code is the fact that the tuning needs to be done correctly and the ALDL data-stream rate is much lower than with the $6E mask (mask is another word for code if that's easier to understand). The alternative to the 1227165 ecm is the 1227730 which was used on the last years of TPI (91 and 92). It was also used in lots of v6 applications and this code is far easier to tune than the $6E.
So does that answer your question about MAF?
Your T5 should handle it so long as you don't powershift it on a regular basis. I actually want a T5 to beat around before I go T56.