Because an area has been"dug out" and replaced you effectively have two joints in your substrate. Whilst I would expect no vertical movement I would expect some horizontal movement due to thermal expansion and contraction. The joints are supposed to be reflected through the tiles. Clearly this would be impracticable not to mention look awful. I think this is one where a decoupling mat is appropriate. It will take account of the "different" substrates and then you would not have to wait for the repair to dry either so straight ofroward sand cement mixed in a bucket would do the trick. G-mat says it is sutiable on asphalt - I guess the other are as well. It would also help avoid all that tiresome preparation
Spoke to both Gemini (G-Mat) and Schluter. Both basically same idea but Schluter were slightly more informative. The question of suitability over bitumen is really down to the adhesive/bitumen contact. (I decided to lay a slurry coat after reading advice on here, Mapei were particulary negative on using their products on it.). Schluter say ditra should be tiled onto directly and any movement/cracking underneath will be absorbed so I am following this order:
1/ Slurry coat.
2/ Insulation on rapidset.
3/ heat mat
4/ slc/rapidset adhesive covering
5/uncoupling membrane on rapidset
6/adhesive
7/tiles
(everything flexible of course)