Another tough week for the biotechnology industry overshadowed impressive clinical results from several companies. Alkermes (ALKS 0.99%) scored a big win in a difficult indication. Puma Biotechnology (PBYI -5.85%) announced a high probability of success with its single candidate. The stock received a short bump before the market punished it for the rest of the week. Eli Lilly (LLY -1.68%) added another oncology therapy to its stable of late-stage candidates. As usual, the lumbering giant just entered a race its competitors have nearly finished.

A win-win
Alkermes hit the bullseye with aripiprazole lauroxil. The schizophrenia therapy is basically a once-monthly injectable version of the successful daily, oral therapy Abilify.

Patients with neurological conditions have an unfortunate tendency to respond to placebos, making this sort of test a real nail-biter for drug developers. Luckily for Alkermes, patients receiving aripiprazole lauroxil scored significantly better on two separate scales.

If the drug is approved, Alkermes could effectively take market share away from itself. The company's delivery technology is in Johnson & Johnson's (JNJ -0.71%) blockbuster schizophrenia therapy Invega Sustenna. Sales of the one-month formulation reached an impressive $1.25 billion in 2013. Johnson & Johnson recently won big in the clinic with a three-month formulation that might do even better, if approved.

Probability problem
In a mid-stage breast cancer trial, Puma Biotechnology placed its investigational neratinib head-to-head with Roche's Herceptin. The adaptive trial design employed some non-orthodox methods to predict the likelihood of neratinib's success in a larger, longer, and far more expensive trial.

Investigators used pathological complete response data from just 115 patients to estimate the probability neratinib will significantly improve overall survival in a 300-patient trial. Then the study went a step further and estimated probabilities of neratinib beating herceptin in several genetically defined populations.

The results were pretty good. Across three different patient populations, the probability of Puma's neratinib outperforming Herceptin are between 71.8% and 79.1%. The market responded well on Monday morning, sending shares up about 10%.

Then fear crept in.

PBYI Chart

Source: Puma Biotechnology data by YCharts.

As a pill neratinib would likely take a big share of injectible Herceptin's $6.9 billion market, but only if it clearly outperforms Roche's blockbuster. Puma's market cap of more than $3 billion going into the announcement suggested the win was already at least somewhat priced in.

With only one candidate in development, fancy statistics weren't enough to comfort Puma investors in the midst of a biotech sell-off. They did just the opposite. The stock crashed, ending the week down 28%.

Another CDK inhibitor
At a recent scientific meeting Eli Lilly entered the CDK inhibitor race with results from an unusually large phase 1 trial. 132 patients with five different tumor types received Lilly's bemaciclib as a monotherapy every 12 hours for 28 days or until their disease progressed. The disease control rate across all five categories was an impressive 70%. Of course, some patient groups showed a stronger response than others.

There were 47 heavily pretreated patients with metastatic breast cancer. Despite a median of seven previous therapies, 19% showed a partial response, and 51% had stable disease. Results were even better for hormone receptor positive breast cancer patients.

Lilly seems prepared to go all the way with bemaciclib. The company has already registered -- not started recruiting -- a phase 3 trial for bemaciclib in patients with hormone receptor positive, HER2 negative advanced or metastatic breast cancer.

Pursuing this indication will put Lilly's CDK inhibitor on a path with Pfizer's (PFE -0.10%) palbociclib. At the same meeting, Pfizer presented data from a phase 2 trial. Advanced breast cancer patients receiving palbociclib plus letrozole showed progression free survival of 20.2 months -- nearly twice as long as patients in the control arm.

As usual for Lilly, it's well behind in this race. Pfizer has already started recruiting patients into four separate phase 3 trials involving palbociclib.