) --

Threshold Pharmaceuticals


presented new data Saturday showing how pancreatic cancer patients who "crossed over" in a mid-stage clinical trial of TH-302 may have hampered the drug's ability to demonstrate a stronger improvement on overall survival.

When these crossover patients were excluded from the analysis of the phase II study, the survival benefit attributable to TH-302 improved.

The new analysis was an effort by Threshold to alleviate investor concerns about the future of TH-302. The company's stock price has fallen by 20% over the past two weeks because of worries that TH-302 would not help patients live longer in a larger, phase III trial scheduled to start soon.

Threshold shares closed Friday at $7.24

The phase II study enrolled 214 pancreatic cancer patients and randomized them to treatment with TH-302/gemcitabine or gemcitabine alone. As presented earlier this year, treatment with

TH-302/gemcitabine led to a two-month improvement in progression-free survival

of 5.6 months compared to 3.6 months for gemcitabine alone. The benefit in PFS was statistically significant, although there were some imbalances in patients' baseline criteria that favored TH-302 and may have skewed results.

An analysis of

overall survival from the TH-302 study

was less favorable. TH-302/gemcitabine reduced the relative risk of death by 4.5% compared to gemcitabine alone -- a small survival benefit that was not statistically significant. This survival analysis was made public two weeks and was responsible for the selloff in Threshold shares.

Threshold attributed the modest survival benefit, in part, to patients in the gemcitabine control arm of the study that "crossed over" and began treatment with TH-302 after their tumors began growing again. This made it more difficult for the study to measure the true survival benefit of TH-302 since patients in both arms of the study were treated with the drug.

To add weight to its argument, Threshold on Saturday presented new data from the study that excluded the 38% of patients in the control arm who crossed over to treatment with TH-302. The new data were presented in conjunction with the presentation of the TH-302 phase II data at the European Society for Medical Oncology

ESMO annual meeting.

Without crossover patients, TH-302 plus gemcitabine reduced the relative risk of death by 24-25% over gemcitabine alone. That's an improvement over the 4.5% risk reduction for death when the crossover patients were factored into the analysis.

Two doses of TH-302 were studied.

After six months, 69% and 73% of patients treated with TH-302/gemcitabine were still alive compared to 51% for the gemcitabine patients.

After one year, 37% and 38% of the TH-302/gemcitabine patients were still alive compared to 21% of the gemcitabine patients.

While encouraging, the new survival analysis is not without caveats. Removing crossover patients creates selection bias, meaning the improved survival benefit may not be attributable to TH-302 but instead to patients remaining in the control arm who were sicker and therefore didn't live as long. The analysis was also retrospective, meaning it wasn't conducted as part of the original design of the study -- lessening its significance.

Almost half the pancreatic patients enrolled in the Threshold study were also treated with additional drugs after TH-302 or gemcitabine stopped working. These "subsequent" therapies could have also helped patients live longer, making it even harder to tease out the survival benefit attributable to TH-302 alone.

On the safety side, TH-302 was associated with higher rates of thrombocytopenia (low platelets), neutropenia (low white blood cells) and anemia. TH-302's blood-related toxicities coupled with anticipation for upcoming results from a phase III study of


(CELG) - Get Celgene Corporation Report

Abraxane in pancreatic cancer dampened enthusiasm for TH-302 at ESMO, wrote Pieter Droppert, a cancer consultant who writes the

TheStreet Recommends

Biotech Strategy Blog/

Droppert writes

: "If

Abraxane were to increase overall survival to 10-12 months, this would become the new standard of care that TH-302 would then have to beat. Based on the TH-302 data presented at ESMO 2012, even if you ignore the hematological toxicity, this is hard to imagine from the phase II study, which is why the data failed to impress me."

The final data from the phase II study of TH-302 in pancreatic cancer presented Saturday answer some questions but raise others. Ultimately, the measure of the drug will only come after Threshold and its partner Merck KGaA conduct and complete a phase III study.

For another take on Threshold and the TH-302 data, make sure to read the latest column from


contributor Nate Sadeghi on why the

pancreatic cancer drug may be obsolete before launch


--Written by Adam Feuerstein in Boston.

>To contact the writer of this article, click here:

Adam Feuerstein


>To follow the writer on Twitter, go to


>To submit a news tip, send an email to:






and become a fan on


Adam Feuerstein writes regularly for TheStreet. In keeping with company editorial policy, he doesn't own or short individual stocks, although he owns stock in TheStreet. He also doesn't invest in hedge funds or other private investment partnerships. Feuerstein appreciates your feedback;

click here

to send him an email.