At the third annual Blueprint for Drug/Diagnostic Co-Development forum, cohosted by Friends of Cancer Research in Washington, DC, and the Alexandria Center for Life Science in New York, two panels tackled considerations in simultaneous development of drugs and companion diagnostics.
Friends of Cancer Research Chair and Founder Ellen V. Sigal, PhD, opened the forum by noting the rapid improvement in next-generation sequencing. “It is being increasingly used to determine treatment options in oncology. Therefore, we need a regulatory pathway to ensure that next-generation sequencing panels are reviewed by [the U.S. Food and Drug Administration (FDA)] and can be amended to incorporate new markers into clinical practice.”
As an example, Dr. Sigal referenced the Lung-MAP clinical trial, in which several companies are collaborating to create a multidrug, multiarm, biomarker-driven trial for patients with advanced lung cancer. Genomic profiling via a comprehensive next-generation sequencing panel matches patients with one of several investigational treatments designed to target specific genomic alterations. FDA has supported Lung-MAP with creative approaches to rapidly review the drugs and their companion diagnostics.
Challenges Enumerated
Mya Thomae, RAC, CQA, Vice President of Regulatory Affairs, Illumina, Inc, outlined challenges facing next-generation sequencing:
The one-drug/one-test approach is inefficient and unlikely to be clinically acceptable for much longer because the quantity and quality of biopsy material are so diverse.
Each next-generation sequencing panel is developed independently by separate organizations with little or no standardization of markers or performance characteristics.
Different orthogonal reference methods for validation exist for each test, making it difficult to draw comparisons.
There is little published validation data for determining reference standards. [To this point, Elizabeth Mansfield, PhD, Director, Personalized Medicine Staff, FDA Office of in Vitro Diagnostics and Radiological Health, said, “FDA realizes that not everything will have reference material, which is important but not critical.”]
It is difficult to determine the fraction of malignant cells in DNA.
“These and other issues need to be in the public domain,” said Karen Gutekunst, PhD, Vice President for Development, Illumina, Inc. She added that a community-wide approach is the best way to establish regulatory pathways, implement standardization of panel content and performance protocols, and make reimbursement decisions.
Five Proposals
The NGS Working Group, established by the Friends of Cancer Research, developed five proposals to address these challenges and facilitate the development and use of next-generation sequencing panels.
1. Define a regulatory pathway for cancer panels (a selection of multimarker gene assays) intended to identify actionable oncogenic alterations (those with supporting data to create risk-benefit assessment of treatment choice) that allow flexibility in the appropriate FDA medical device pathway—for instance, one based on risk classification of different panel components depending on the specific marker.
A regulatory pathway must be flexible enough to account for the rapidly changing science and drug development pipeline, said Roman Yelensky, PhD, Senior Director, Biomarker and Companion Diagnostics Development, Foundation Medicine, Inc. “A key consideration in this flexibility should naturally be potential risk to patients, which depends on how specifically the markers are intended to be used,” he noted. Thus, the pathway needs to stratify based on risk.
Mickey Williams, PhD, Molecular Characterization Laboratory, Leidos Biomedical Research, Inc, added that the intended use of an assay is critical to any regulatory proposal. For example, certain genomic alterations may require more rigorous review when they are related to safe and effective use of certain treatments.
“Moreover, if a drug was targeted against a marker not on the panel, then a separate companion diagnostic would be needed and classification assigned based on intended use of the marker,” said Dr. Williams.
Dr. Mansfield said that FDA would be concerned about special controls put into place for clinical trials. She explained the current device classification reference guide:
- An investigational device exemption allows a device to be used in a clinical study to collect safety and efficacy data required to support a premarket approval application or a premarket notification submission. A study with devices that hold significant risk must be approved by FDA or an institutional review board before it can begin.
- Devices that pose significant risk require premarket approval.
- Premarket notification is made to FDA to demonstrate that a device is at least as safe and effective as an already-marketed one not subject to premarket approval. The data required to support equivalence vary significantly and may not require a clinical trial.
2. Approaches to validation studies should be based on the types of alterations measured by the assay rather than on every alteration individually.
Dr. Williams reiterated the need for specificity of next-generation sequencing for analytic accuracy, and the absolute requirement for variant standardization.
Dr. Mansfield added that FDA has been innovative in its approach to analytic studies for multiplex microbiology panels in order to decrease study requirements while still ensuring safety—for example, approving a study that compares well-characterized composite reference information derived from the combination of multiple sequencing methodologies, publicly available data, and hereditary information.
Classifying types of genomic alterations and performing analytical studies based on those classifications are also important, given that reference specimens never bear all possible genomic alterations, added Dr. Williams.
FDA requires validation from the tissue type at which the assay is directed, or from blood, which provides assurance of assay performance within each tissue.
The panel agreed that handling massive datasets, developing tools to ensure sequence quality, conducting sequence alignment and assembly, and interpreting and drawing inferences from data are increasingly vital to the success of next-generation sequencing.
3. Determine the contents of a cancer panel by classifying potential markers based on current utility in clinical care and clinical trials and peer-reviewed publications, as well as recognized clinical guidelines. Draw upon various sources to determine the recommended marker set for an actionable cancer panel.
The NGS Working Group, said Omar Perez, PhD, Director of Diagnostics Oncology, Biotechnology Clinical Development, Pfizer, Inc, suggested a process to determine the contents of an actionable panel by classifying genomic alterations. Performance standards for each marker or group should be established, and those that are FDA-approved should be available to the public. Laboratories complying with the clinical laboratory improvement amendments (CLIA) developed by the Centers for Medicare & Medicaid Services also should publish performance data so that all cancer panels, regardless of where they were developed, can be compared.
Dr. Mansfield applauded this suggestion and said, “Standardization is critical, not just helpful.”
4. Promote standardization of cancer panels through development and use of a common set of samples to ensure reproducibility on each platform.
The Working Group encourages public communications so that decisions can be based on hard data, as well as on how materials will be replenished. Panels, based on well-characterized tissue specimens, are difficult to prepare and replenish.
The question remains about who would prepare these samples and whether work that is underway at the National Institute of Standards and Technology and National Cancer Institute should be considered. And will standardization of assays result in fairer reimbursement. That is, does a particular set of data make reimbursement decisions easier?
5. Establish a framework for determining an appropriate reference method rather than relying on any single method for all studies.
Elias Ketchum, Manager, Regulatory Affairs, Roche Sequencing, said that finding agreement about a reference method is central to all other decisions. A number of factors are required: the analyte being measured, sensitivity of the assay system, regulatory status of the proposed method, and what to do if there is no reference method available.
Excited and Scared
Dr. Sigal said she was excited by what she had heard from the panels—but also “scared to death.”
She wondered how accurate these new tests and diagnostics are. “Are they really safe? What if the wrong test is used, and thus the wrong drug? Not only would individual patients be affected, but so would the whole industry.”
- Which studies should be required to define an extraction method, and which would be acceptable to FDA?
- What level of sensitivity is sufficient, and what are the critical components?
- Because there is so much variance within genes, which types of evaluation demonstrate clinical evidence?
Steven D. Averbuch, MD, Vice President, Translational Clinical Development and Pharmacodiagnostics, Bristol-Myers Squibb, said that best practices need to be delineated. “They must depend on the quality of the data, and they require high standards of reliability. Moreover, we must maximize patient access to reliable tools.” ■
Disclosure: Drs. Sigal, Yelensky, and Averbuch reported no potential conflicts of interest.
