I haven't used it yet, but I've been dieing to try DUY's Magic Spectrum. I've heard great things about it. Might be worth a look.

If you can't do that, I've found that EQ'ing in individual bands of frequencies at a time can make the job of matching dialog MUCH easier. Use high and low pass filters to make a band pass on the original and the re-record, then EQ the re-record to to match the sound in that band. Slide the filters to the next frequency range you need to work on and repeat. Build the bands around the fundamentals for the vowels and the consonants to speed up the process.

Obviously, remove the band pass filtering when you're done and review.

Get the timing right first using time compression and expansion.

Then match the levels using gain and compression/expansion.

Next match the timbre using EQ.

Finally hide the edits using cross fades.

As in ADR, i think the most important thing to match is your talent's performance. Also, they need to say the whole sentences which contained the mispronounced words - not just the words themselves - or your rhythm will be off. If you can get your talent to match their original tone (or pitch) and rhythm, then dropping in the new words will be easy, and shouldn't need too much processing (unless you're using a drastically different mic).

Sorry if you're already on top of the performance side of things; it's just my experience that, when recording ADR, performance is 95%. Unfortunately, they don't have a plug in for that yet.