The swift development and approval of Pfizer Inc.’s Xalkori (crizotinib) and the companion diagnostic from Abbott Molecular Inc. resulted from an efficient clinical program shaped by intensive communication and coordination between the two sponsors and both FDA’s Center for Drug Evaluation and Research and its Center for Devices and Radiological Health.

Both FDA centers and the two sponsors put a significant investment of time and effort into coordinating the accelerated approval of the drug with the diagnostic test required by its Aug. 26, 2011 clearance for use in “patients with locally advanced or metastatic non-small cell lung cancer that is anaplastic lymphoma kinase (ALK)-positive as detected by an FDA-approved test.” Abbott’s Vysis ALK Break Apart FISH Probe Kit was approved the same day.

Xalkori achieved a quick approval in large part because it showed notably strong efficacy early in development.

Xalkori is notable not only as a pioneer in personalized medicine, as one of the first drugs approved in tandem with a companion diagnostic, but also for its speed through the development process. FDA’s parallel approval of the two products came only five years after ALK was identified as a target for NSCLC, and FDA spent less than five months reviewing the crizotinib NDA Also see "Pfizer’s Crizotinib Eases Past FDA With Targeted Population" - Pink Sheet, 29 August, 2011.. Pfizer filed the NDA based on early clinical data – a Phase I and a Phase II study – but provided additional reassurance by starting two confirmatory Phase III trials months before the submission.

Xalkori achieved those milestones in large part because it showed notably strong efficacy early in development. FDA’s willingness to put the man-hours – including significant participation by top CDER oncology office and division management as well as their high-level counterparts from CDRH’s Office of In Vitro Diagnostics and Division of Immunology and Hematology Devices – into all the meetings and teleconferences during development reflects its excitement over the early data and confidence in the drug’s efficacy.

The Xalkori decision came just nine days after FDA approved another drug/companion diagnostic pair in record time. Roche’s Zelboraf (vemurafenib) followed a similar trajectory of about five years from the start of IND trials to approval of the drug and its companion diagnostic, Roche’s cobas 4800 BRAF mutation test. Zelboraf’s approval also was driven by compelling objective response rate data in Phase I and II trials. Indeed, with Zelboraf (a candidate for full, not accelerated, approval), FDA intervened in August 2010 to prod Roche to revise its statistical plan so trials could speed up Also see "Roche and FDA in Oncology: Fighting in Public, Collaboarting in Private" - Pink Sheet, 1 November, 2011..

But these sorts of compelling product stories leading to fast, simultaneous approval of targeted drugs and their companion diagnostics may not be sustainable as the field matures and runs up against FDA time and resource constraints.

Johnson & Johnson – a firm with plenty of experience in both therapeutics and diagnostics – predicts that parallel development may not easily become the norm. In comments on FDA’s draft guidance on companion diagnostics, J&J said the draft, with its goal of simultaneous approval of drug and diagnostic, “focuses on an ideal, but largely impractical, model of synchronized development and contemporaneous approval” Also see "Views Differ On How Lab-Developed Tests Fit In FDA Companion Dx Guidance" - Medtech Insight, 24 October, 2011.. J&J argued that drug firms do not typically identify the need for a companion diagnostic until after Phase II and into Phase III.

Early Signs Of Efficacy

Xalkori first made the oncology community sit up and take notice in 2007, when a Phase I trial (A8081001, which would eventually serve as “Study B,” one of the pivotal trials in the NDA) showed a high response rate in patients with ALK-positive NSCLC. The EML4-ALK fusion gene had only been identified as an oncogenic driver of NSCLC earlier that year, in an article in the journal Nature.

Crizotinib was “initially developed against c-Met,” a different genetic target, Office of Clinical Pharmacology Genomics Group reviewer Rosane Charlab Orbach noted. The ‘1001 trial actually started in 2006 as a Phase I all-comers trial for patients with advanced cancers other than leukemia. The company quickly adapted the study to include a focus on EML4-ALK Also see "Crizotinib Holds Promise For Lung Cancer, And For Pfizer Oncology" - Pink Sheet, 14 June, 2010..

The enrichment strategy in the early phase clinical studies led to a restricted entry Phase II trial – ALK+ only – to support accelerated approval, Orbach summarized.

Pfizer also reached out to FDA as it moved toward restricting the population. “The applicant, noting a partial response in 7/14 ALK positive patients, met with the agency to discuss their registration strategy,” clinical reviewer Shakun Malik recalled in an Aug. 12, 2011 review. At that April 23, 2009 meeting, Pfizer raised the potential for accelerated approval based on a single-arm study. Within a month, Pfizer submitted a request for a Special Protocol Assessment for a Phase III trial comparing crizotinib with standard of care chemotherapy. No agreement was reached on an SPA, but with confidence that adequate Phase III studies would happen in the near term, FDA did not dig in its heels about the design of the trials supporting the accelerated approval.

And aside from the novel patient stratification, with the accelerated approval application Pfizer was bidding to become the first NSCLC therapy cleared on the basis of response rate data from single-arm trials. “All of the approved therapies for unresectable, locally advanced, recurrent or metastatic disease NSCLC have been based on improvement in overall survival compared to a comparator,” Malik stated.

When Pfizer initially set forth its accelerated approval strategy, FDA stressed its desire for randomized trial data. At the April 2009 meeting, the agency “recommended that the applicant conduct a randomized trial of crizotinib vs. conventional therapy,” Cross-Disciplinary Team Leader V. Ellen Maher reported in a July 12, 2011 review. “The agency did suggest that if the applicant chose to pursue accelerated approval, ‘that you entertain a randomized study with an interim analysis of a surrogate endpoint in a larger population,’” she said.

Pfizer vigorously defended its proposed pivotal single-arm trial over FDA’s randomized study preference during the April 2009 meeting, the minutes show. The company conceded that “for second-line treatment in unselected patient populations, a randomized trial comparing the new drug to established standard of care is completely reasonable.” However, the company continued, “in the situation with [crizotinib] treatment of patients with EML4-ALK fusions though, there are several potential issues with a randomized study.”

“Investigators may be reluctant to enroll patients on a study where they may be randomized to a more toxic, non-targeted therapy, such as pemetrexed or docetaxel, that does not directly target the molecular defect driving their tumors,” Pfizer argued. Additionally, “a randomized trial in this very rare patient population would be expected to take longer than a single-arm study, thereby delaying the availability of this potentially effective treatment.”

A year later, Pfizer returned to FDA with a proposal based on submission of two single-arm studies of crizotinib in ALK+ NSCLC. FDA conceded at the April 14, 2010 meeting that “such a strategy may be acceptable for accelerated approval.” The agency was reassured that “the overall registration strategy” included two randomized Phase III trials: Study 1007, comparing crizotinib with pemetrexed or docetaxel in ALK+ advanced NSCLC patients who had received a course of platinum chemotherapy, and Study 1014, testing the drug against pemetrexed in combination with either carboplatin or cisplatin in previously untreated subjects (see sidebar).

[Sidebar#06120301039]

FDA continued to reiterate its preference of overall survival as the primary endpoint for Phase III, but Pfizer argued that “progression-free survival should be the primary endpoint and overall survival a secondary endpoint given the expected subsequent treatment with crizotinib and other investigational ALK inhibitors, which will be in clinical trials and would likely confound the assessment of survival,” according to the April 14, 2010 minutes. Pfizer reminded FDA that during the April 2009 meeting “there was agreement that PFS is an acceptable endpoint for targeted therapy such as crizotinib for full approval given a large effect size and an acceptable risk to benefit ratio.”

FDA allowed that “a large difference in PFS combined with an acceptable safety profile may be acceptable.”

Confidence In Confirmatory Trial Outcomes

Xalkori’s swift trip down the accelerated approval pathway was smoothed by the review team’s confidence that the efficacy data from single-arm Phase I and Phase II studies that formed the basis of the NDA would be confirmed in Phase III. Two randomized Phase III studies were already underway by mid-2010 after plentiful discussion of the protocols by the sponsor and FDA, grounding the reviewers’ confidence.

The level of efficacy was another influence. “The response rate seen in these single-arm trials is much superior to any chemotherapeutic regimen even in the first-line setting,” clinical reviewer Malik declared. “The magnitude of crizotinib benefits, i.e. response rates of 50% in Study A and 61.2% in Study B, was superior to all the drug regimens approved for the treatment of advanced metastatic NSCLC.” While the NSCLC armamentarium features many options, Malik noted that five-year survival remains “dismal” at around 15%.

“Compared to conventional chemotherapy, there was a marked elevation in response rate and duration of response observed in two single-arm studies with crizotinib. It is thought that this will translate into an improvement in overall survival in this patient population,” Office Director Richard Pazdur concluded in an Aug. 26, 2011 decision memo, noting that the two Phase III trials of crizotinib vs. conventional chemotherapy were still necessary to “verify clinical benefit.”

Division Director Robert Justice, in an Aug. 26, 2011 memo, pointed out that the two efficacy studies in the NDA “enrolled patients who had received multiple prior systemic therapies. Twenty-eight to 36% of patients had received four or more treatment regimens. Despite this, the objective response rates in these studies (50% and 61%) were higher than would be expected with standard chemotherapy. In addition, the median durations of response were 41.9 and 48.1 weeks, respectively.” Justice also concluded that “it is likely that these responses will translate into an improvement in progression-free survival and overall survival in the ongoing confirmatory trials.”

The confidence in Xalkori’s efficacy expressed by FDA review staff was echoed by two outside consultants. The National Cancer Institute’s Udayan Guha said that “an ORR of more than 40% is a significant improvement to current therapies,” according to minutes of an Aug. 22, 2011 teleconference. NIH’s Wyndham Wilson, who also serves as chair of FDA’s Oncologic Drugs Advisory Committee, found the results “clinically meaningful,” according to the minutes of an Aug. 19, 2011 session.

Safety Profile Appears Promising

It wasn’t just the promising efficacy that helped smooth the way – the initial evidence indicating that crizotinib had a favorable safety profile, possibly better than the alternative treatments for NSCLC, caused further enthusiasm among the reviewers.

Malik noted that safety “could not be adequately assessed due to the single-arm trials,” but added that the adverse event profile seemed favorable compared to chemotherapeutic agents approved for NSCLC.

Pazdur’s stance was that “although crizotinib appears to be less toxic than conventional chemotherapy, further follow up and examination of the adverse event profile of crizotinib in a randomized trial will be necessary to fully define the safety signals associated with crizotinib.”

In his consulting role, Wilson expressed some misgivings about Xalkori’s safety profile. “The toxicity spectrum is a bit worrisome especially pneumonitis, vision disorders and hepatic toxicities,” he told FDA. “However, it is not possible to differentiate if pneumonitis was caused by the underlying lung cancer or if it was drug related. Most of the liver and visual toxicities noted were reversible.”

The vision effects of crizotinib received another consultant review, from FDA Division of Transplant and Ophthalmology Products Deputy Director Wiley Chambers. He noted that crizotinib patients had a 53% incidence of a grade 1 visual disorder, and that 44% of events resolved as patients continued drug therapy. The oncology division adopted Chambers’ recommendation for a post-marketing study, including it as a post-marketing requirement in the approval.

From Lab-Developed Tests To Abbott’s Vysis

At Pfizer’s first major meeting with FDA to set the stage for potential accelerated approval – the April 2009 EOP2 meeting – FDA pointed out that “a means for detecting ALK inversions is not clearly described in your briefing package,” and asked Pfizer how the company planned to “detect such inversions exclusively.” Pfizer proposed an ALK break-apart FISH assay be used as the primary screening assay to determine patient eligibility. The agency stressed that “the test used for the study should be the same as that for registration and marketing.”

In the Phase I Study ‘1001, five different lab-developed FISH tests were used to determine patients’ ALK status for inclusion in the ALK-positive NSCLC cohort. A lab-developed test from Massachusetts General Hospital served as the reference test.

“We will engage a diagnostics company … for the development of a FISH-based RUO [research-use only] diagnostic assay, with the intent to convert this into an IUO [investigational-use only] test for patient enrollment in our upcoming trials,” Pfizer told the agency.

But FDA informed the company that it did not think an investigational-use only assay was sufficient for accelerated approval of the drug. “The sponsor would need approval of the marketable test in concert with approval of the drug,” FDA stated. The agency urged Pfizer to meet with CDRH’s Office of In Vitro Diagnostics. OIVD already was aware of the potential for a parallel approval process; meeting minutes show that OIVD Division of Immunology and Hematology Devices Associate Director Reena Phillip attended the EOP2 meeting.

Speed was important, FDA emphasized. The sponsor “should bring the assay to IUO (investigational-use only) status, including the design controls, in time for use in the pivotal trial,” the agency instructed. Pfizer noted that it planned to “immediately start” working on the first step, a research-use only assay, “with a diagnostic partner for use in the pivotal trial.”

Only a few months later, on July 24, 2009, Pfizer and Abbott Molecular met with the Office of In Vitro Diagnostics to discuss requirements for an investigational-use only assay, laying the groundwork for testing the companion diagnostic in the pivotal studies for the drug. The companies announced a deal for a test to detect ALK translocations to select patients for crizotinib therapy on Aug. 27, 2009 [See Deal].

Over the next several months, Pfizer, Abbott Molecular, CDRH and CDER were in frequent contact, review documents show (see[A#06120301011]). Pfizer appears to have taken an energetic role to facilitate coordination between the drug and diagnostics offices. The April 14, 2010 EOP3 meeting minutes depict Pfizer announcing that “we have initiated the scheduling of a meeting with CDRH/OIVD to discuss FDA’s expectations for a diagnostic in conjunction with accelerated approval,” and requesting that “members from CDER attend the meeting (in person).” The sponsors’ facilitation of contact between CDER and CDRH staff may have helped overcome an institutional divide; a Pfizer R&D exec recalled the CDER oncology and CDRH diagnostics staff meeting for the first time at the April 2010 session Also see "Xalkori And The Art Of Modern Drug Development" - In Vivo, 24 February, 2012..

Abbott Molecular continued in a leadership mode, instigating a May 24, 2010 teleconference with the sponsors and the drug and device reviewing divisions “to ensure that all entities involved have a complete understanding regarding the data that Pfizer plans on submitting” in the NDA’s 60-day update.

No Equivalent Of Accelerated Approval For Devices

The lack of a formal expedited approval mechanism for diagnostics was a minor stumbling block. The sponsor’s proposed path forward for “an expedited PMA review status submission for the companion device” required further consultation within the agency, according to minutes of a May 11, 2010 meeting between OIVD, Pfizer and Abbott Molecular. “CDRH does not have a method for conditional approval of the device, such as CDER’s Accelerated Approval option, and as such we will need to discuss with CDER whether or not there are potential concerns and timelines.”

CDRH cautioned during a June 24, 2010 meeting that expedited review status for the Abbott Molecular Vysis ALK Break Apart FFPE FISH Probe Kit “to align with the accelerated approval status for the drug crizotinib” could not be assured “definitely” because “FDA cannot predict if there will be review issues.”

“We will need to coordinate and discuss with CDER” about the path for an expedited PMA review status submission for the companion diagnostic, OIVD reiterated during the June meeting with Abbott Molecular and Pfizer; also present were CDER Division of Drug Oncology Products medical officer Ian Waxman and Office of Oncology Drug Products Acting Deputy Director Anthony Murgo.

On April 21, 2011, CDRH granted the Vysis diagnostic kit expedited review status “because it was believed through literature support that the Vysis Anaplastic Lymphoma Kinase (ALK) Break Apart FISH Probe Kit would aid in the selection of previously treated non-small cell lung cancer patients for treatment with the Pfizer drug Xalkori (crizotinib), which may result in a life-prolonging survival benefit and is therefore in the best interest of those patients,” according to the summary of safety and effectiveness data for the Vysis assay.

“Because no other legally marketed device is available, the FDA decided to grant expedited review,” the summary document states.

Beyond FISH

While preparing for parallel approval of Xalkori and the Abbott Molecular FISH diagnostic, Pfizer also sounded out the agency on how to expand to other companion diagnostics in the future. Again, Pfizer was moving quickly: the company announced a deal with Roche subsidiary Ventana Medical Systems Inc. for a ALK diagnostic based on a different technology in January 2012 Also see "Pfizer Turns To Ventana To Expand Companion Dx Options For Xalkori" - Pink Sheet, 18 January, 2012..

Ventana will develop a fully automated, standardized immunohistochemistry companion (IHC) diagnostic test for ALK mutations. IHC tests are protein-based, and are generally considered less accurate than nucleic-acid based FISH assays – but a fully automated IHC test would be less expensive than the FISH alternative.

At the April 14, 2010 EOP3 meeting, Pfizer asked if “retrospective use of these clinical specimens for potential follow-on PMA applications for additional diagnostic tests” would be acceptable. FDA said retrospective use of specimens “could be feasible if test results can be ascertained on a high percentage of trial patients and if the number of samples is sufficiently large for precise estimation of concordance.”

“Please note that an evaluation of concordance only on patients prescreened for ALK positive events is insufficient,” FDA continued. “Concordance between the current test and a future additional test is evaluated not only on patients who are ALK positive on the current test (positive percent agreement or PPA) but also on patients who are ALK negative on the current test (negative percent agreement or NPA),” FDA explained.

“To evaluate NPA, a subset of samples of patients who are ALK negative on the current test is needed. For the future test to be considered equivalent to the current test, both PPA and NPA should be very high, e.g., significantly greater than 95%. The numbers of ALK negative and ALK positive samples should be large enough to power the study accordingly and we believe that this will require no fewer, or possibly more, than 100 ALK negative samples.”

‘”Please note that any discordance between a future test and the current test leads to uncertainty in the treatment effect among patients who are ALK positive on the future test,” FDA emphasized.

Lightning In A Bottle

The Xalkori review in many ways looks like a model for accelerated drug development in the era of targeted therapies with companion diagnostics – fast and efficient, with cooperation and communication between four key groups (the two sponsors and the two FDA review centers). And, combined with the similarly swift Zelboraf approval, the agency certainly deserves credit for nimbly navigating these early tests of a new paradigm for drug approval and development.

But the Xalkori story resonates more deeply with an older theme: FDA’s oncologics review division will go to great and creative lengths to bring to market drug candidates that show exceptional promise in early trials, especially if a survival benefit seems likely. It is a rare drug that shows a notable benefit over existing therapy, coupled with indications of better tolerability than existing therapy, that can be used with a diagnostic test that is already available through local labs, and so only needs to show concordance to support a branded companion version.

Future drug/companion diagnostic pairs probably also cannot count on the sustained participation of division management as clinical trial plans are hammered out in the same way that, for example, CDER Office of Hematology and Oncology Products Associate Director for Regulatory Science Anthony Murgo (who also served as acting deputy director of the Division of Drug Oncology Products during the review) or CDRH’s Office of In Vitro Diagnostics Director of Personalized Medicine Elizabeth Mansfield or Division of Immunology and Hematology Devices Associate Director Reena Phillips and Director Maria Chan were consistently present for detailed meetings with the sponsors.

Pfizer saw the potential in Xalkori early and had the institutional flexibility and capability to move the drug forward rapidly. The deep regulatory experience of Pfizer on the drugs side and Abbott Molecular on the diagnostics end helped avoid trouble spots and make credible arguments in defense of choices like pursuing approval with single-arm trials when FDA repeatedly called for randomized trials comparing crizotinib to conventional chemotherapy.

The need to coordinate the simultaneous approvals of the drug and the companion diagnostic added a new level of complexity to the Xalkori NDA, but ultimately the story is an old one: FDA will work creatively with sponsors to bring breakthrough medicines to advanced cancer patients.

Synchronized Regulation: Tandem Approvals Of Xalkori, Diagnostic Required Considerable Coordination

More from Clinical Trials

More from R&D