The Oncologic Drugs Advisory Committee's review of Cell Therapeutics' Pixuvri (pixantrone) - where the treatment for relapsed/refractory non-Hodgkins lymphoma was roundly rejected - was another opportunity for FDA Office of Oncology Drug Products Director Richard Pazdur to urge sponsors to look beyond late-stage settings to test cancer drugs.

Pazdur made the same point about early- vs. late-stage disease settings during an ODAC meeting in June 2009, at which the panel voted in favor of approving GlaxoSmithKline's Arzerra (ofatumumab) (¹ Also see "Oncology Results Can Be Clearer If Trials Target Early Stages – FDA’s Pazdur" - Pink Sheet, 8 June, 2009.).

"[A] major point that I want to emphasize for people who are developing drugs in this disease is you do not have to develop the drug in the most refractory disease population," Pazdur said at ODAC's March 22 meeting. "It does require probably a little more work, but one can develop in combination and take an investigational drug in combination as long as you isolate the effect of the drug in a first-line setting."

"No one is compelled to treat a patient in the most refractory disease setting," Pazdur stressed.

Cell Therapeutics may have courted failure by choosing an advanced setting to test Pixuvri: relapsed or refractory NHL in patients who had seen their disease progress on at least two prior chemotherapy regimens.

The company was able to enroll only 140 of a planned 320 patients in the Phase III PIX301 clinical trial within two years. Cell Therapeutics opted for early termination of the trial and moved forward with analysis of the 140 patients, but FDA found that to have invalidated the special protocol assessment and deemed the trial incomplete.

The treatment setting contributed to the trial tribulations. There is limited availability of patients who meet the criteria, and Cell Therapeutics noted that U.S. and European oncologists prefer to treat patients at this stage of disease with combination therapy, and thus were reluctant to enroll them in the single-agent trial. Other patients at that late stage of disease opted for palliative care instead.

The widespread adoption of frontline treatment with rituximab (Roche's Rituxan ) also was a factor that affected the Pixuvri study.

The company made strenuous efforts to salvage the situation by improving patient recruitment, as CEO Jim Bianco laid out for the advisory committee. These included Web-based investigator outreach, regional education sessions, the initiation of no fewer than 189 sites in 24 countries, and the hiring of six additional regional contract research organizations. But despite those efforts, the company pulled the plug in advance of a planned interim analysis and submitted the data from the 140 patients as its only pivotal trial.

Too Little To Work With, FDA Said

FDA objected to the early termination of the trial - a decision Pazdur maintained was made without consultation with the agency. Bianco maintained the firm had alerted FDA of the possibility in August 2007, about seven months before the unscheduled enrollment halt, and that the briefing documents were the first revelation that the SPA may have been breached. When pushed about the discrepancy after the meeting, the CEO noted that the firm interpreted no comment from FDA as acceptance. "Typically the agency tells you when they disagree with what you are doing as opposed to when they agree," he said.

In the end, FDA interpreted the NDA package to "have less than a single completed trial ... since less than half of the planned patients were accrued," Pazdur said.

Pazdur laid out for the panel the instructions that are given for all sponsors considering using a single supportive trial: "For a single randomized trial to support an application for drug approval, it should be well executed, internally consistent and include statistically persuasive efficacy findings," he said, paraphrasing FDA's guidance on clinical evidence of effectiveness.

He added that the evidence for the primary endpoint has to be statistically persuasive, that there should be corroboration of secondary endpoints and across patient subsets, and that the results should be duplicable in additional trials.

FDA determined that the PIX301 results did not meet the requirements for statistical significance - though Cell Therapeutics attempted to argue that it modified the expectations for the lower patient base and therefore was sufficiently powered (² Also see "Significance Of Evidence For Cell Therapeutics' Pixuvri Is On The Line" - Pink Sheet, 19 March, 2010.). Still, because PIX301 was stopped when only 44 percent of the planned total number of patients had signed up, the primary endpoint of complete response or unconfirmed complete response "did not meet the critical level of significance," Pazdur said.

Pazdur further explained that the secondary endpoint analysis can only be considered as exploratory, as there was not power for statistical interpretation.

Cell Therapeutics did place a heavy emphasis on the secondary endpoints - arguing that the totality of the evidence reinforced the primary endpoint results and collaboratively demonstrated pixantrone's benefit.

The company asserted that pixantrone met the primary endpoint, "with 20 percent of patients achieving a CR/CRu," said Chief Medical Officer Jack Singer. Those responses "achieved irrespective of response to prior therapy were durable and associated with encouraging overall survival. Significant improvements in overall response rate and progression-free survival with a favorable trend in overall survival [the secondary endpoints] corroborate the primary endpoint findings."

"The results were consistent across important subgroups, including in patients confirmed by central histology review," Singer added.

Trouble Within The Data Set

However, after discrediting the value of the trial from the very start, Pazdur also found flaws with the results of the analysis. He noted that five of the 14 responses in the Pixuvri arm "occurred in patients with non-eligible, generally less aggressive histologies," and should have been excluded. In that case, the rate drops to 16.7 percent, compared with 6 percent in the comparator arm.

There were problems that extended beyond the small sample size. "The results are clearly not generalizable to the U.S. population, as the U.S. had especially poor accrual of only eight patients, U.S. patients were generally more heavily pre-treated, and none of the CR/CRus occurred in the U.S.," Ian Waxman, a medical officer with FDA's Division of Drug Oncology Products, said.

This argument found resonance among ODAC members. "I think the evidence would speak against this being applicable for patients treated in the West; 97 percent of all CRs and CRus occurred in patients treated outside Europe and the U.S.," Wyndham Wilson, National Cancer Institute, said.

FDA also was concerned that the study's independent assessment panel "upgraded" two patients in the active arm and one in the comparator arm from partial to complete response, but made no such changes in the other direction. These apparently subjective decisions were based on re-reviews of the patients' CAT scans and not on any new radiological, clinical or biochemical evidence, Waxman said.

ODAC members agreed that the trial results were less than impressive, statistically speaking. "I think certainly with the small sample sizes that we're seeing here, and the fact that the results may not be robust in terms of the responses ... the results of the study come into question, given the data questions we have about the responses [and] the issues with whether the histologies are aggressive or not," Brent Logan, Medical College of Wisconsin, said.

Safety Could Be Another Stumbling Block

Aside from the concerns about the efficacy side of the data set, FDA identified safety issues as a potential problem for Pixuvri. It also was another point where FDA and the company did not see eye to eye. For example, Pazdur noted, 7.4 percent of the Pixuvri patients experienced Grade 3-4 febrile neutropenia, two-and-a-half times the 3 percent rate seen in the comparator arm. In addition, 25 percent of the patients in the pixantrone arm developed cardiac dysfunction, more than twice the rate suffered by the patients in the comparator arm.

Looking at the same numbers, Singer called the rate of Grade 3-4 neutropenia "acceptable," and judged the rate of febrile neutropenia and neutropenic infections to be low. The rate of Grade 3 or greater cardiac events in the Pixuvri arm also was low, according to Singer, "especially considering their prior anthracycline exposure." Still, the higher incidence for the pixantrone arm over the control arm raised flags for FDA.

Assessing the safety profile probably will be a matter left for further clinical evaluation of pixantrone.

New Trial Likely Will Be Next Step

"I think the drug does have biological activity and I think that it's going to be very important to work on what the specific disease indication and patient population is in subsequent studies, and whether in fact it would be an agent best tested in combination," S. Gail Eckhardt, University of Colorado at Denver, said after ODAC unanimously recommended against approval. "Because it is a cytotoxic agent, I do think it's important we obtain an adequate risk-benefit ratio for the patient."

Cell Therapeutics is prepared to conduct an additional trial. During a March 25 call with investors, Bianco said that "While we await the FDA recommendations or decision" - which the firm expects in the form of an action letter on or before the April 23 user fee date - "we have been working on one potential path with our clinical experts, in developing a combination study." The possibility of conducting another study of Pixuvri as part of a combination regimen was mentioned by Pazdur at the panel review (³ Also see "Cell Therapeutics Should Consider Different Trial Design For Pixuvri, FDA's Pazdur Says" - Pink Sheet, 22 March, 2010.).

Cell Therapeutics has submitted a draft clinical development plan to FDA, which it is "hopeful could serve either as a post-marketing requirement or an additional pivotal study if the FDA should require such an approach," Bianco said. Based on their consultation with lymphoma experts in the U.S., the firm expects a combination trial would enroll "at a very quick pace."

The company is also pursuing another approach to salvaging its approval prospects - by submitting the safety results of a Phase II trial in the first-line setting, a trial using combination regimens set in the U.S. and Western Europe. "This data will help answer questions surrounding whether or not pixantrone has a better cardiotoxicity profile than the standard doxorubicin-containing regimen," Bianco noted. Plus the company could scrounge up additional data from other trials.

"We plan to discuss with FDA whether the PIX203 safety data coupled with the 70 patients treated on the PIX301 [original pivotal trial] and supplemented with the 33 patients treated in our Phase II single-arm study [AZA201] could constitute an amendment to the NDA seeking accelerated approval,"

Aside from the fact that FDA has been explicit that it deems PIX301 to be a failed, incomplete trial and now has a unanimous rejection by ODAC, that would be the best option for the company. Bianco confirmed CTI has roughly $51 million in cash, though he stressed that the burn rate is between $4 million and $4.5 million a month (if there isn't a launch in the second half).

"Should we need an additional trial to support approval, we do not believe the costs would be overly burdensome or highly dilutive, especially if we were able to obtain a co-development from a pharma partner like Novartis," Bianco said. On the investor call, he announced that CTI has initiated a dialogue with Novartis regarding "a more traditional co-development program for pixantrone."

- Martin Berman-Gorvine (⁴ m.berman-gorvine@elsevier.com)

Learning From Pixuvri: Pazdur Urges Sponsors To Test Cancer Drugs In Earlier Stages

More from Archive

More from Pink Sheet