GTx announced top-line results from the enobosarm phase IIb study in June 2009. Note at that time, Merck ( MRK) was collaborating with GTx on enobosarm's development. (The drug was called Ostarine back then.) In March 2010 (and after having the phase IIb cachexia data in hand) Merck decided to abandon Ostarine/enbosarm, handing rights to the drug back to GTx. GTx has discussed and presented enobosarm data from the phase IIb study since 2009, but the full study was only published this past April in Lancet Oncology. (Adrian S. Dobs et al, Effects of enobosarm on muscle wasting and physical function in patients with cancer: a double-blind randomised controlled phase 2 trial.) Read the Lancet Oncology paper very closely and what you'll discover is that the study was widely skewed in favor of enobosarm over placebo. Patients with more advanced disease or less muscle function at baseline were randomized disproportionately to the control (placebo) arm, which artificially elevated the effect of enobosarm. A close read of the study also reveals that a good number of patients were outliers, meaning they performed either really well or really poorly on muscle function tests. Coincidentally or not, the study's endpoints were assessed using mean (average) values, not the median, which in this case, worked out incredibly well for enobosarm relative to placebo. But when the effect of outlier patients is minimized by analyzing data using median values -- the more widely accepted method -- enobosarm's benefit over placebo disappears. Take a look at this slide, which comes from GTx and depicts results from the phase IIb study isolated only for the non-small cell lung cancer patients. Remember, this is the group of cancer patients in which GTx says enobosarm had the most benefit, which is why the ongoing phase III studies are only enrolling lung cancer patients. Sixty-one lung cancer patients were enrolled in this study out of 159 total, or 38 percent. But notice only half the enrolled lung cancer patients -- 31 to be exact -- were included in the efficacy analysis. According to the Lancet Oncology article, there were 10 patients and 13 patients with lung cancer randomized to the placebo and 3 mg enobosarm arms of the study, respectively. This means GTx claims positive results for enobosarm in lung cancer based on data from 15% of patients in the phase IIb study. Why would enobosarm work better against cachexia in lung cancer patients than in other forms of cancer? GTx hasn't offered an explanation but the Lancet Oncology paper shows clearly that the lung cancer patient subgroup was skewed in enobosarm's favor.