There is no good way to spin a Phase III failure. By then you've made it past the main reasons for a drug to wipe out (PK and total mechanistic failure). A breakdown at this stage is a more subtle affair (well, except for the money involved, which is not subtle at all). For example, a drug might show efficacy in a carefully constructed Phase II trial, but can't perform under the wider (and more realistic) conditions of Phase III.
That's what appears to have happened to Merck's (MRK) MK-7418 (rolofylline, formerly KW-3902). This adenosine A1 antagonist, which Merck picked up by buying NovaCardia a couple of years ago, was being developed for acute heart failure. That's a tough indication, and this isn't going to improve that reputation. (This Forbes piece has a tour of the pile of discards that this area has become over the years. Rolofylline looked as if it might work in Phase II, but (from what I can tell from the press releases) missed every endpoint in Phase III.
On a chemical note, rolofylline is a rather odd-looking molecule. You don't see many noradamantanes hanging off of drug structures. I'm sure this wasn't the reason for the compound's failure (after all, it made it through Phase I and Phase II), but it's sure not something I have on my list of structural fragments to try.