It’s a good idea and would likely work very well over flat terrain or the sea (where the CFIT chance is less) but I have no idea how accurate, precise or consistent RA readings are over hilly/forested/city areas, covering the range of aircraft that are going to be flying these approaches and the fact that there will be a lateral splay even on a precision approach, let alone a RNP one. Within the boundaries of the airfield, which is where we use RA to seriously measure stuff and make decisions based on the results, we’re over a level, solid surface. At the FAF and/or 3NM point there could be almost anything underneath that varies in height by hundreds of feet depending on your precise track.