Positive results in a Phase II trial of its breast cancer drug neratinib sent stock in Puma Biotechnology Inc. skyrocketing 68 percent Thursday. In the trial, a combination of neratinib and Taxol unexpectedly outperformed a Herceptin/Taxol combination by 18 percent to 19 percent, according to analyst estimates.

With the positive Phase II results, neratinib has now “graduated” and is eligible for Puma’s upcoming I-SPY Phase III trial. Upside suggested by strong superiority to Herceptin has investors jumping on the bandwagon, with a share volume of 3.59 million, nearly 19 times the average of 186,000. Puma’s stock (NYSE:PBYI) gained $31.49 to close at $77.70.

The Los Angeles-based biotech acquired neratinib from Pfizer Inc. in 2011. The oral, irreversible tyrosine kinase inhibitor was in development at the Phase II stage for HER2-positive metastatic breast cancer. Pfizer had encouraging data for neratinib, but licensed it out in the course of prioritizing its pipeline. (See BioWorld Today, Dec. 15, 2009.)

The drug is in the same family as Tykerb (lapatinib, Glaxosmithkline plc) and afatinib, a Phase III compound in development by Boehringer Ingelheim GmbH. It is designed to inhibit the HER2 and epidermal growth factor receptors.

Herceptin (trastuzumab) was developed by Genentech Inc., of South San Francisco, a subsidiary of Roche AG, to target the HER2 receptor, and it has been a powerhouse in breast cancer since it was approved in 1998.

The I-SPY 2 trial is a randomized Phase II trial for women with newly diagnosed Stage II or higher breast cancer that evaluates the effect of adding investigational drugs to standard chemotherapy in the neoadjuvant setting, or before the primary treatment is administered. The primary endpoint of the trial is complete pathologic response in the breast and lymph nodes at the time of surgery. The trial matches investigational regimens with patient subsets on the basis of biomarker signatures to determine benefit of the regimens.

The results are analyzed using Bayesian predictive probability to determine whether a regimen will be shown to be statistically superior to standard therapy in a future randomized 300-patient confirmatory trial. Those that yield a high Bayesian probability of success graduate, and those that show a low probability of success are dropped. A maximum of 120 patients are assigned to each regimen, and a regimen can theoretically graduate any time after having 60 patients assigned.

Alan Auerbach, CEO of Puma, said in an investor conference that the trial began in 2010, and at the time it began it included neratinib and “a bunch of other agents.” Since then, most of the other agents have dropped out, and newer agents have been added. However, there are no immediate prospects for other graduating regimens from the trial.

“In terms of newer agents for HER2-positive disease, the only newer ones that came in entered in July of this year. It took neratinib three years to graduate. We are anticipating the same for those newly added. We don’t expect them to graduate any time soon,” Auerbach said.

Detailed results of the trial will be released at a future medical meeting. Puma did not provide detailed information about safety results. Auerbach, however, addressed the company’s plan for managing diarrhea in its Phase III trial. Historically, diarrhea has been the primary adverse event associated with the drug, with some level of diarrhea present in 90 percent of patients, according to trials carried out by Pfizer prior to Puma’s acquisition of the drug.

Puma expects diarrhea rates to be in line with that previously published figure, so it has been evaluating antidiarrheal agents for prophylactic use. Auerbach said tests of low and high doses of loperamide showed that when given at a high dose in the first three days of therapy, then tapered to a low dose, loperamide produced good results, reducing rates of all-grade diarrhea to 10 percent.

The unfamiliar statistical analysis threw some analysts for a loop, and was the subject of a number of questions during the question and answer period of Puma’s press conference. According to Puma, the Bayesian probability of superiority for its neratinib-containing regimen was 95.3 percent, and that result is analogous to a “p” value of 0.047.

Puma is not releasing detailed results, but the firm did offer some hints. When a ballpark estimate of 15 percent for the absolute improvement in pathologic response rate was suggested, Auerbach said, “I am completely comfortable with that estimate.”

Leerink Swann analyst Howard Liang wrote, “Neratinib’s improvement over Herceptin appears to be approaching 18-19 percent based on management comments,” noting that management indicated that Herceptin did not underperform historical data and that the improvement of the neratinib arm over the Herceptin arm was similar to, although slightly lower, than the improvement of combination arms over Herceptin, which was 18 percent to 19 percent.

The good news could extend to use of neratinib in the adjuvant setting, where it is being evaluated in a Pfizer Inc. study following completion of one year of Herceptin therapy in HER2-positive breast cancer. The trial could show benefit of changing treatment from Herceptin to neratinib in the second year.

The positive results for neratinib in I-SPY 2 have had naysayers. The nontraditional trial design has drawn fire from observes who doubt the Bayesian analysis. However, in Puma’s investor conference, Auerbach affirmed that the correlation of the more familiar “p” value measure with the Bayesian analysis was indeed completely valid.