When life hands you lemons, you make lemonade. In biotech, when a study fails, you go on a mining expedition into the bad data to find something positive to say. This is what Keryx did unsuccessfully with perifosine, and this this is exactly what Apthera, the original developer of NeuVax, and Galena, which later bought Apthera for just $7 million, have also done. Pouring through the data from the failed trials, Apthera/Galena discovered that breast cancer tumors with low and intermediate levels of HER2 expression appeared to respond better to NeuVax. At 36 months, the cancer relapse rate for a subgroup of 50 NeuVax-treated women with breast cancer that expressed low (1+) and intermediate (2+) levels of HER2 had a relapse rate of 8% compared to a 23.5% relapse rate for a subgroup of 34 women in the control arm with tumors that also expressed relatively low levels of HER2. This difference in relapse rate was statistically significant with a p value of 0.045, claims Galena. Naturally, Galena is conducting its phase III study of NeuVax in breast cancer patients with tumors that express low and intermediate HER2. And of course, this NeuVax study will fail because the premise upon which it is designed -- culled from a retrospective data-mining adventure through the phase II study -- is false. How do I know it's false? For starters, it makes no sense for a HER2-targeted cancer vaccine to work better in tumors that express low levels of HER2, or no HER2 expression at all. Breast cancers in the phase II study were analyzed for HER2 expression using a staining process called immunohistochemistry, or IHC. It's not the most accurate or reliable way of determining HER2 status -- flourescence in situ hybridization (FISH) testing is better. Moreoever, tumors that come back HER2 1+ via IHC are usually categorized as HER2 negative. How is a HER2 cancer vaccine like NeuVax supposed to work on HER2-negative tumors? Galena doesn't explain. Again, logic and science dictates that NeuVax should work really well in breast cancers that express high levels of HER2, the so-called HER 3+ group. But some mathematical extrapolation on the data made available by Galena from the phase II study reveals that NeuVax actually performs worse in the HER 3+ group. The 36-month relapse rate for NeuVax-treated patients with HER 3+ tumors is 9% compared to a 7% relapse rate for the control group, also with HER 3+ tumors. Galena cherrypicks a subgroup of patients where it claims NeuVax works but ignores another equally large subgroup of patients where NeuVax actually hurts patients, even though the vaccine should work for better these latter patients. None of this bodes well for NeuVax. The phase III study is going to fail, all that's necessary is patience since we may not see results for three years. NeuVax isn't going to work -- the phase II data don't lie. Before wrapping up this dissection of Galena, let's do away with a couple more canards that seem to have retail investors excited and optimistic. First, Roche/Genentech is not interested in acquiring Galena, nor is the company a partner in a phase II study testing the combination of NeuVax and Herceptin in breast cancer. Galena has issued many press releases claiming Roche/Genentech's involvement in this small phase II study, but that's misleading because Roche/Genentech's participation is limited to a small amount of funding and the donation of Herceptin supplies to the investigator so he can conduct the study, says company spokesman Ed Lang.