Detecting Set Over Set

I'm a big believer that NL HE players should sometimes be able to lay down sets in full ring games when set-over-set is a strong possibility. But, having been known for seeing monsters under the bed, I figured I should ask.

NL HE $200-buy-in $1/$2 blinds online: Limped pot with five players including big blind. I have $225, Unknown Player has just joined and bought in for $200 and has the big blind. I limp in cutoff with 4 4.

Flop is K T 4. I lead $5 into $9.80 when it is checked to me., I am check-raised to $25 by the Unknown Player. I make it $50 to go. At the time, I was really thinking about getting away from the hand if he came back over the top. He did, for all his chips, and I eventually called, thinking that I didn't know the player that well and sometimes players go crazy with top two. I figured he'd have raised preflop almost all the time with KK so his range is only KT and TT (most players where I play don't semi-bluff with the nut flush draw, but I guess I could throw specifically A Q to the mix). Also, the average player (which I have to declare him since he just joined) will sometimes raise from the big blind with TT, so that contributes a little bit to the odds he has that in the big blind. The statistics I could compute in the 15 seconds I had (no time bank on this site) seemed to indicate that even if he is twice as less likely to make the play with KT/A Q than he is with TT, I should probably call for roughly 1.5-to-1. Of course, he had TT, or I wouldn't be telling this story.

I can't really take a turn from his check-raise due to the heart draw, so I think the reraise was right most of the time. Maybe I should have reraised more on the flop, in which case it would have been an auto-call due to odds. His over-the-top for all his chips made it possible for me to fold, but I just couldn't do it. Should have I?

My game selection has gotten so good that I basically never get stacked anymore drawing this thin, so I'm hyper-aware when I do and want to be sure I did it right.