Over the last decade, the cost of sequencing an individual’s genome has plunged from about $10 million to about $1,000. Simultaneously, in an effort to shore up aging blockbuster and me-too models of drug development, the pharmaceutical industry has rebalanced its pipeline to focus on unmet medical needs. The era of genomic (aka precision or “personalized”) medicine has truly arrived.
Sequencing the genome has given rise to other “omic” technologies, such as proteomics (the large-scale study of proteins), pharmacogenomics (how genes affect a person’s response to drugs), and metabolomics (the metabolic responses to pathophysiological stimuli or genetic variation), which also have become more affordable. These new tools can tailor diagnostic and clinical decision-making to a patient’s unique DNA. In addition, clinical trials can become smaller and faster to conduct thanks to more targeted and genetically-defined entry criteria.
- Consequently, the proportion of new drug approvals for targeted medicine has risen from 20% of FDA approvals in 2014 to 28% in 2015. As of mid-2015, the FDA listed more than 150 approved medicines with pharmacogenomic information in their labeling, and over the next five years, the proportion of personalized medicines in clinical development is expected to increase to nearly 70%, according to the Tufts Center for the Study of Drug Development (TCSDD).
Unfortunately, companies trying to create commercially viable products using genomic medicine have encountered substantial obstacles. Among these are the challenge of producing products at an affordable price, a lack of expertise required to stratify patients for clinical trials, the absence of harmonized global ethical and regulatory requirements governing the handling of patients’ genetic data, and the many difficulties associated with developing drugs that require a companion genomic test (i.e., one company may hold rights to the drug while another has rights to the diagnostic).
Three new skills to realize the promise and avoid the perils
Many companies are not yet equipped to function effectively in the personalized medicine environment. At PAREXEL, we believe this transformation requires honing three critical skills:
- Collaborating with external partners
- Integrating genomics into the development process
- Maintaining ethical and privacy standards
Collaborating with external partners
The rise of consortia and multifactorial trial designs has removed taboos against industry data sharing. Today, collaboration, even with competitors, both directly (via deals to co-develop combination products) and indirectly (via private-public-academic alliances), is central to success in genomic development.
For example, the Biomarkers Consortium (BC), launched in 2006, is a public-private biomedical research partnership that seeks to discover which biomarkers can accurately predict clinical benefit. Managed by the US Foundation for the National Institutes of Health, the Consortium’s partners include government agencies such as the NIH, FDA, and the Centers for Medicare & Medicaid Services (CMS), and more than 30 private companies, trade associations, and non-profit disease-specific advocacy groups.
One recent BC project, the I-SPY 2 breast cancer trial, uses an adaptive randomization design to match patients with newly diagnosed Stage 2 or 3 breast cancer to the experimental agent that best suits their molecular subtype (the genetic signature of each patient’s tumor is profiled at entry). The Phase II study subs in and out up to a dozen different drugs from different companies, comparing their efficacy to standard chemotherapy, and dropping and adding arms in response to the data. Drugs that produce promising results graduate to Phase III trials (or other types of confirmative testing) which are anticipated to be smaller, more focused and targeted to patients most likely to respond.
The potential synergy offered by combination therapies (that is, two or more drugs delivered simultaneously or in a predetermined sequence) could boost clinical efficacy, but it requires competitors to work together in bringing products to market. Even companies developing single-agent personalized medicines may need to collaborate with a diagnostics developer on a companion diagnostic (CDx) test. When the test and the medicine are owned by different companies, things get complex. Selecting and qualifying a partner, providing clinical laboratory oversight and standardization, and obtaining regulatory approval for a diagnostic represent uncharted waters for companies that have heretofore focused exclusively on therapeutic drugs.
And what if there are multiple approved diagnostics for the same class of products? How will hospitals afford to stock them all? How will doctors know which one to use? In 2015, developers of PD-1/PD-L1 immunotherapy products for non-small cell lung cancer (NSCLC) decided to attack that problem with the Blueprint PD-L1 Assay Comparison Project. The initiative is a first-of-its-kind collaboration between four pharmaceutical companies (AstraZeneca, Roche/Genentech, Bristol-Myers Squibb, Merck) and two diagnostics makers (Dako, Ventana Medical Systems), all engaged in developing anti-PDL1 immunotherapy/CDx products. The goal is to clarify how well investigational PD-1/PD-L1 tests work to identify which NSCLC patients will respond to treatment.
As they expand their collaboration skill sets, companies will need to be mindful of:
- Choosing quality partners, including CROs, experienced in conducting multifactorial trials and managing complex collaborations;
- Initiating multi-party alliances that can accelerate development;
- Managing the mix of relationships and commitments to deliver cost-effective drugs to patients while protecting shareholder value and intellectual property.
Integrating genomics into the development process
Designing and developing biomarkers for integration into clinical trials is a core competency for genomics-based drug development. A 2015 study in Nature Genetics estimated that using the growing wealth of human genetic data to select the most promising clinical targets and indications could improve the current (low) success rate in drug development by a factor of two.
Deciding whether to implement a pharmacogenomics or biomarker companion test in drug development is challenging. Uncertainty around the validity of many genomic tests (due to the complexity of gene expression) makes biomarker expertise even more crucial. PAREXEL internal data suggest that genomics is becoming more prevalent in all phases of clinical development and as many as 83% of developers are outsourcing at least some portion of their genomic medicine activities. Companies also are partnering with healthcare service providers, especially in the areas of quality control, genomic data management/stewardship, results interpretation, and the ability to leverage genetic evidence to support regulatory submissions.
Using genomic markers to stratify patients allows companies to run smaller, more focused trials, but they still must ensure that sample sizes are large enough to prove a drug’s efficacy, as well as the efficacy of any associated genetic test. In addition, sample collection must be traced to precise patient populations to demonstrate a test’s utility. Companies must also navigate market timing challenges, assessing exactly when a test needs to be available in order to be used in a trial.
The industry’s growing focus on niche markets and pharmacogenomic drug development has raised evidentiary questions about smaller clinical trials and patient enrichment strategies based on biomarkers, per one interview-based study of stakeholders. Some companies have faced backlash from patients and advocacy groups when studies of promising experimental agents enrolled fewer subjects than they once did, due to strict genetic criteria.
New guidelines governing testing and genomics abound. The U.S.FDA, European Medicines Agency (EMA), and the Japanese Pharmaceutical and Medical Devices Agency (PMDA) have all published regulatory guidelines on pharmacogenomics, but it’s still up to developers to collect compliant genomic data. Genomic information should directly inform decision-making, and should allow accurate assessments of the benefit/risk ratio and the medical value of a drug in its intended patient population.
Finally, improving the success rate of clinical trials by enrolling only patients who can benefit is pointless unless the resulting products are approved, reimbursed, and cost-effective. Commercial access issues and patient-related outcomes need to remain central to development.
Companies practicing genomic-based development need to keep the following in mind:
- Outsourcing all or a portion of the pharmacogenomics, data analytics, bioinformatics, or target validation work may be a developer’s most efficient option, allowing access to specialized skills and deep experience at potentially lower cost
- Smaller trials are good; trials too small to draw conclusions from are not
- Pharmacogenomics data should articulate the benefit/risk ratio and medical value of a drug
- All drugs need to be cost effective to succeed in today’s market
Maintain high ethical and privacy standards
The possible mishandling of genomic samples and information, and the potential for discrimination based on genomic information remains a huge and to-date unresolved risk of personalized medicine.
The possible harmful impacts on patients (for instance, medically actionable results may prompt some patients to undergo invasive procedures), and the affordability of genetic tests are all ethical dilemmas of genomic testing. Companies must navigate these issues to bring products successfully to market. Education – of providers, patients, investigators, and other stakeholders – will be critical to ensuring the uptake and proper use of personalized medicine products.
For many firms, the complexities and expense of genomics-based patient enrollment necessitate finding an external partner with the proper expertise, technology, and global footprint to execute clinical trials. The diversity of global regulations governing biobanking (biospecimen collection and management), the need to retain samples during development and post-marketing, and the quality of collection and annotation required to ensure valid results can be daunting. Trials may involve extensive training of investigators regarding sample collection and processing requirements, not to mention full documentation of the chain of custody of samples and data.
Reinventing drug development
Are healthcare stakeholders – sponsors, physicians, patients, payers, and regulatory authorities – fully prepared for the era of personalized medicine? Probably not. And it remains unclear whether current pharmaceutical development economic models work for genomic medicine.
However, at PAREXEL, we believe adapting to personalized medicine and the host of changes it will bring is a necessity and an opportunity. Developing the right skill sets for collaboration, for incorporating new technologies and processes, and for handling genetic data securely and ethically will demand building expertise in-house, working with external partners, and finding optimal combinations of both.