In spite of the barrier of the mountain, is it possible the uplink continued well into the approximate 20 second time frame (until impact) of the last trajectory, and simply did not have sufficient time to effect RTH? Time to impact is based on last known speed and distance to the mountain side you cite.
I've actually never seen a documented case of continued uplink after downlink was lost, so I don't know if it ever happens. But even if it does, in this case the sudden loss of downlink implies a pretty solid barrier, and that is supported by the flight track and DEM. Uplink would have gone down very shortly afterwards if not at the same time.