Top-line results for Aastrom Biosciences' (Nasdaq: ASTM) bone marrow-derived cellular therapy, ixmyelocel-T, look good. I'm not sure they were let's-announce-the-conference-call-31.5-hours-in-advance-and-get-investors-excited good. But good nonetheless.

Ixmyelocel-T met its primary endpoint of delaying the time to first occurrence of treatment failure in patients with critical limb ischemia, an artery disease in which poor circulation can lead to amputations.

Treatment failure is a composite endpoint of major amputation of the treated leg, all-cause mortality, doubling of wound size from baseline, and de novo gangrene. The first two are obviously clinically relevant. Who wants to lose a leg or die? The second two are important because the tissue loss is a predictor of eventual amputation.

Aastrom is saving the full data for a medical meeting, so we don't know by how much the treatment failure was delayed. On one hand, it doesn't matter too much; these are patients without any treatment options, so the fact that there is any kind of delay is what's most important. But this is a phase 2 trial, so the magnitude of the delay will help determine how easy it will be for Aastrom to meet its endpoint in phase 3 development.

There are plenty of examples where companies have taken drugs that looked like they were working in phase 2 and weren't able to duplicate the results in phase 3. It happens a lot in oncology -- Novartis' (NYSE: NVS) lung-cancer treatment ASA404 comes to mind -- but it's not specific to cancer. Medivation (Nasdaq: MDVN) crashed hard after its Alzheimer's disease drug, dimebon, failed to live up to expectations. Ditto for Inspire Pharmaceuticals' cystic fibrosis drug, denufosol, which failed a phase 3 trial. In the latter case, denufosol had actually passed one phase 3 trial, proving that nothing is certain in drug development.

If ixmyelocel-T does meet its endpoint in phase 3, the Food and Drug Administration is likely to approve the treatment because Aastrom is getting a special protocol assessment, or SPA, from the agency. The SPA allows companies to get input on the trial design from the FDA, so the company knows that the agency is content with the endpoint, trial size, and statistical analysis. Aastrom thinks it has reached an agreement with the agency, but it's waiting for the final approval before starting the phase 3 trial.

Getting started with the phase 3 trial quickly is important, but investors will want to be certain they check the full phase 2 data when it's presented in the fourth quarter.