The Institute of Medicine's National Cancer Policy Forum recently convened a public workshop, "Facilitating Collaborations to Develop Combination Investigational Cancer Therapies," to address the promises and challenges involved in the development of combination oncologic drug therapies. In the September 1 issue ofThe ASCO Post,we explored the concept that combining investigational products early in their development may be a promising strategy for identifying effective therapies, especially when a combination targets multiple pathways, or more than one step in a pathway, conferring greater benefit than a therapy directed at a single target.

In the concluding installment of this report, we consider scientific challenges and opportunities in the codevelopment of investigational therapies; the regulatory environment for codevelopment, including the recent FDA draft guidance on this topic; and legal issues that influence collaboration.

Challenges and Opportunities in Clinical Development

Drug developers will be faced with the need to change their mindset from competition to collaboration if these new combinations are to go forward. They won't like it, but itcanbe done. For example, the Biomarkers Consortium, a public-private partnership was established in 2010. It comprises NIH, FDA, patient groups, and pharmaceutical and biotech companies and initiated a groundbreaking trial known as I-SPY 2 that predicted drug responsiveness in breast cancer based on the presence or absence of genetic and biologic markers. The trial continues to evaluate tumor response to a variety of investigational drugs, albeit not in combination.

Stuart Lutzker, MD, PhD, Vice President for Oncology Early Development, Genentech, presented an industry perspective. "Most conventional oncology drug development follows the paradigm of first identifying a biologic feature unique to cancer cells as compared to normal ones, identifying a drug-able target, and generating candidates with acceptable preclinical pharmacology and toxicology. Then we assess safety, pharmacokinetics, and efficacy as a single agent or in combination with standard of care, which is most often chemotherapy," Dr. Lutzker said.

"This is a relatively straightforward process but has several limitations: First, there is a high failure rate due to inability to combine with standard of care therapy. Second, efficacy may be limited due to failure to fully suppress the pathway or biologic process, the pathways may be resistant from the outset, or resistance may be rapidly induced," he continued.

With rational combinations of two targeted therapies, many of these problems may be overcome. However, these combinations require both a strong scientific rationale and pharmacologically compatible molecules. It is often difficult and time-consuming to identify an optimal dose and dose schedule for these rational combinations, and many combinations turn out to be more toxic than thought.

"An early emphasis on pharmacodynamics markers is seen to be key in making correct phase II dose and schedule decisions," Dr. Lutzker said.

Clinical Trial Considerations

Pharmaceutical companies are talking about collaboration, but the challenges appear daunting. Especially in phase I trials, said Patricia LoRusso, DO, Director of Experimental Medicine, Barbara Ann Karmanos Cancer Institute. She asked whether, based on our current knowledge, we may be moving into drug combinations too quickly. Since often we lack appropriate molecular signatures, we don't know that best in class is important when selecting agents in combination.

"Where will these new combinations have the greatest impact: in metastatic disease or in the adjuvant or neoadjuvant setting?" she asked. "And what are the risks of studying both the individual agents and drug combinations at the wrong clinical stage?"

Clinicians' challenges, she said, are to find the best drugs to use in combination. But how many combinations of similar targets need to be tested, and what are the moral obligations of identifying and using the most appropriate of each class selected for combination therapy? And can the use of many of these combinations as proof of concept-despite not having the clinically best in class of each agent combined-slow or even negatively impact clinical drug development of selected targeted combinations?

And if the drugs are not optimal or if tumor types with appropriate targets are not selected, then what? "Does this also limit targeting similar combinations, should we use them as proof of concept only, and does it slow clinical development?"

Dr. LoRusso said that patient selection is critical but difficult because selection tools are currently less than effective and very expensive. However, is investing dollars in the development of these selection tools more time-efficient and financially efficient? Despite all these obstacles and unanswered questions, she continues to conduct phase I studies.

"Our expectations are different from traditional monotherapy in terms of acceptable response and use of assays to define tumor effects. Often we focus on surrogates, which, although identifying target effect in select tissues, may actually lead to false hope when it comes to target effect at the tumor site-the true tissue target for the agents under investigation."

Regulatory Concerns

There has been a good deal of concern in industry and academia about how FDA would regulate testing and approval of targeted combinations.

The agency panel was chaired by Richard Pazdur, MD, Director, Office of Oncology Drug Products. He was joined by Rachel Sherman, MD, Associate Director for Medical Policy, Center for Drug Evaluation and Research, Robert Temple, MD, Deputy Director for Clinical Science, CDER, and others.

In response to the concern, FDA has developed a draft guidance for development of combination therapy. In general, basic requirements of safety and efficacy apply, but the draft guidance addresses several unique issues:

• Whether codevelopment is appropriate. It usually provides less information about safety and efficacy, with resultant increased risk. Therefore it should be undertaken under only limited circumstances, that is, for difficult-to-treat disease.
• Whether the biologic rationale is plausible (clearly advantageous over monotherapy) and whether the model developed should be compared to its individual components.
• The need to demonstrate that each component of the combination contributes to the treatment effect. This could require a large clinical trial in a multiarm design, unless such a study is unethical because it might promote resistance and render the new therapy ineffective.
• Preclinical testing for proof of concept and safety.
• Criteria for clinical safety and efficacy, first in healthy volunteers if possible, and if not, in early phase I studies.
• The need for phase II studies to demonstrate the contribution of each component of the combination as to efficacy and dosing. The amount and types of clinical data needed will vary depending on the nature of the combination, the disease, and other factors.
• The need for phase III studies to demonstrate the contribution of each component, compare the combination to standard of care, and fine-tune the dosing.
• Assurance that drugs in a combination be used only together.

Legal Concerns

This new enterprise carries potential antitrust concerns, said Robert F. Leibenluft, Partner in the Washington law firm Hogan Lovells. "It is possible that if pharmaceutical companies, who are otherwise competitors, coordinate their activities to develop drugs, the result will be higher prices, lower quality, and reduced innovation, which could in turn have an effect on the future market for products, services, technology, and innovation," he said.

"On the other hand, research and development collaborations among competitors can be efficient and are not uncommon. Most such efforts are undertaken without serious antitrust scrutiny."

Mr. Leibenluft said that the Federal Trade Commission and the Department of Justice have few antitrust concerns where the following issues are accounted for:

• There is no intent to harm competition by raising prices, reducing output or quality, or thwarting innovation.
• There are other comparable research and development efforts in addition to those of the collaboration that would ultimately compete with them-in other words, competition is preserved.
• Collaborators do not already have entrenched products similar to the type to be developed.
• Collaboration is limited to core research, with collaborators free to commercialize independently.
• The benefits of collaboration are very convincing and likely could not be achieved independently. ■

Disclosure: Dr. Lutzker is an employee of Genentech.