If you follow the references you finally end up with 'non-classical Polonoski Reactions'.
Note that the N-cyclopropylmethyl homologue produces similar yields so naltrexone-->noroxymorphone is facile.
I feel that those who understand the possibilities this work possible make it an interesting read. The 14-hydroxy morphinans appear to follow the QSAR of homologues with the 4.5-epoxy bridge so one would imagine that the N-phenylethyl or better. the N-(2-(2-furanylethyl)) products will be QUITE potent (480-640x M is my best guess).