Novel Clinical Trial Designs In Oncology: An Histology-Independent Approach

Cancer’s heterogeneity is a leading reason why product development in this therapeutic area is so risky and costly – and why the pharma industry has not yet addressed all needs in the market. Because low-incidence tumors present recruitment challenges and have lower commercial potential than high-incidence ones, they receive less attention from drug developers; meanwhile, new therapies for high-incidence tumors must demonstrate higher efficacy than therapeutic alternatives in order to differentiate themselves.

• Advances in our understanding of the molecular basis of cancer will likely fragment treatment of larger tumors into many smaller patient subpopulations, creating challenges for drug developers to design studies using histology – or tumor site – as the basis for recruiting patients.
• There is growing interest in innovative clinical study designs in which patients across multiple tumor types are pooled into a single study based on the molecular signature of their tumors (also known as histology-independent, histology-agnostic, basket studies) allow sponsors to enter the market with multiple indications simultaneously, drastically reducing development time lines, saving resources, creating marketing synergies, and supporting rapid penetration.
• Yet, the pros and cons of using histology-independent studies for registration are still largely unexplored, and companies need clear decision criteria as they create their development strategies. Histology-independent trial designs, which are potentially applicable in a range of situations, must be evaluated based on many clinical, regulatory, and commercial factors.
• The most important step is to create alignment with regulatory authorities, starting with a broader dialogue between stakeholders.

The Conventional Approach: Histology-Based Study Designs

For years, drug developers have used histology as the fundamental analytical unit by which they study new agents for their effectiveness. Homogenous patient cohorts (based on their histological distinctions) produce more predictable outcomes. This rationalization is consistent with the mechanisms of action for many cytotoxic agents and some classes of targeted therapies. For example, the cytotoxic effects of taxanes in breast, ovarian, prostate, and lung tissues are likely attributable to inherent metabolic properties of the tumor, such as drug absorption/uptake, DNA replication/cell division, and mechanisms of drug resistance. Likewise, the effects of angiogenesis inhibitors (e.g., bevacizumab) in certain solid tumors are related to the neovascularization properties of specific tumors and the molecular targets that are involved.

Despite the technical merits – and proven success – of this path for drug development, histology-based approaches can be inefficient and costly as they require sponsors to conduct separate Phase II and Phase III studies in each tumor type – often in a staggered or sequential manner.

Although this approach does reduce up-front investment risk, it has significant drawbacks. Launch delays can result in the loss of substantial commercial value, not to mention prevent cancer patients from accessing effective therapies. Developers have reduced incentives to invest in rare tumor types and low-frequency mutations in moderate-to-large tumor types because the patient base is small, disease awareness is low, and recruitment difficult.

Science Now Supports Histology-Independent Approaches

With advances in our understanding of the molecular basis of cancer, we can pinpoint the molecular targets responsible for disease onset and progression – an ability that should, in time, extend to most cancer types. In such cases, patient cohorts can be distinguished based on their genetic/molecular signature, rather than on their tumor histology. (Tumor histology is, after all, nothing more than a proxy for molecular segmentation.) This alternative way of achieving homogeneity in the study sample affords developers the opportunity to study multiple cancers having a common molecular target/pathway in a single clinical trial.

Pooling patients across multiple rare tumors into a single study can provide a sample of sufficient size to assess the agent's overall efficacy using common endpoints. While manufacturers do conduct multiple clinical studies in parallel across multiple cancers in exploratory (Phase I) studies, there are only a few examples of this approach being used in registrational studies, with Novartis AG’s Gleevec (imatinib) being the best known. (See Exhibit 1.)

Adopting a histology-independent approach offers many benefits, including:

• The potential to receive marketing approval for multiple indications from a single successful trial. The FDA, for example, approved Gleevec, a BCR-ABL and c-Kit inhibitor, in 2009 for four rare indications.
• Shorter clinical time lines than conventional approaches. The conventional, sequential approach to pivotal studies, when successful, results in approvals for multiple indications separated by years. In contrast, a single pivotal trial spanning multiple cancers can compress the development time line for a compound dramatically.
• A feasible way to conduct well-powered studies in rare cancers. With a histology-based approach, it can be very challenging to recruit enough patients for clinical studies in rare cancers (such as cholangiocarcinoma and hairy cell leukemia) and in rare mutations of more common cancers. Enrolling a sufficient number of patients can often take two or more years, and such trials still run a high risk of failing to demonstrate a statistically significant benefit due to an inadequate sample size. In contrast, a study investigating multiple cancers simultaneously targets a larger patient population, providing sufficient statistical power to assess efficacy and safety.

Exhibit 1

Examples Of Histology-Independent Approaches

Agent/Sponsor	Patient Population	Clinical Trial Design	Outcome	Reference
Gleevec (imatinab)/Novartis	Evidence of KIT, PDGFRA or PDGFRB overexpression	Phase II open-label single-arm study. 186 patients with 40 malignancies enrolled.	Confirmed response in 8.9% of solid tumors and 27.5% of hematologic malignancies. Notable activity observed in four tumor types (DFSP, HES, MDS, ASM).	Heinrich et al. Clin Cancer Res 2008;14:2717–2725
Tykerb (lapatinib)/GSK	HER2 amplified gastroesophageal, bladder, ovarian, or uterine tumors	Phase II double-blinded discontinuation study. 141 patients screened, 32 patients enrolled.	At week 12, 1 (3%) patient had a complete response, 9 (28%) had stable disease, 20 (63%) had progressive disease, and 2 (6%) were unknown. Low response rate, slow screening, and limited enrollment led to study closure.	Galsky et al. Invest New Drugs 2012;30:695–701
Tafinlar (dabrafenib) and Mekinist (trametinib)/GSK	Patients harboring BRAF V600E mutation	Phase II open-label non-randomized study	Study ongoing	Clinicaltrials.gov (NCT02034110)

Abbreviations: myelodysplatic/myeloproliferative diseases (MDS/MPD), aggressive systemic mastocytosis (ASM), hypereosinophilic syndrome (HES) and/or chronic eosinophilic leukemia (CEL), dermatofibrosarcoma protuberans (DFSP)

IMS

What’s Holding Sponsors Back?

A small number of companies are actively exploring how to design and conduct histology-independent pivotal clinical trials, but few have thus far used the approach in pivotal studies to support registration. One recent example is GlaxoSmithKline PLC’s Phase II trial of the combination therapy of Tafinlar (dabrafenib) and Mekinist (trametinib) in patients with rare cancers that bear the targeted BRAF mutation. These agents are already approved for metastatic melanoma, but GSK is investigating the effectiveness of this combination therapy in other rare cancers, including anaplastic thyroid cancer, biliary tract cancer, gastrointestinal stromal tumor, non-seminomatous germ cell tumor/non-geminomatous germ cell tumor, hairy cell leukemia, multiple myeloma, and adenocarcinoma of the small intestine. The study is designed to determine the overall response rate (ORR) of dabrafenib and trametinib in subjects with rare BRAF V600E mutated cancers.

Of course, the effectiveness of a targeted therapy may indeed be histology specific, in which case using the conventional approach is appropriate. Phase I or II studies should be used to clarify this, paving the way for conducting histology-independent pivotal trials when appropriate. However, even after this basic question has been answered, the decision to pool multiple tumor types involves weighing the inherent clinical, regulatory, and commercial challenges of the approach against its benefits.

Clinical Issues

The basic challenge is to identify metrics that are robust enough to measure the variance in responses across cancers. Even if all cancers included in a trial are responsive to the targeted therapy, the level of response is likely to vary by tumor type due to histology-specific factors. Overall survival (OS) and progression-free survival (PFS) rates – which vary by organ site – will likely be very different across histologies. Thus, response rate may be the only common endpoint.

It is therefore difficult to select the appropriate success criteria that can be applied to all histologies. The endpoints chosen will be of interest to: 1) regulatory agencies as they evaluate safety and efficacy and 2) payors and health technology assessment (HTA) agencies as they conduct cost-benefit analyses. Regulators may, in fact, be quite accepting of nontraditional endpoints, whereas it remains to be seen if payors will understand and appreciate the study findings as they look at treatment cost and comparative effectiveness.

One practical approach is to establish success criteria for each specific histology; however, an aggregate, multi-histology response rate will facilitate evaluation of trial results. Thus, the selection of endpoints will be a key consideration in the choice of the cancers to be included in a histology-independent study.

It's also important to design the assay system to screen patients efficiently and consistently; an inability to do so may impact patient accrual and quality of clinical data. In many instances, a biomarker test may not be commercially available at the time of study initiation. Regulators have been willing, however, to work with lab-developed tests and accept validation of commercial tests later.

Regulatory Issues

1. Breaking new ground in how clinical research is conducted comes with a caveat: regulatory risk is high. At this writing, the FDA has not issued specific guidance on histology-independent trials; however, the agency clearly is not against the concept. In fact, at the November 2011 Conference on Cancer Clinical Research hosted by the Friends of Cancer Research and the Brookings Institution, Janet Woodcock, MD, director of the Center for Drug Evaluation and Research at the FDA, commented on the acceptability of multi-histology trials as long as a clear targetable population can be identified (e.g., through a biomarker).

1. Moreover, the agency’s top drug official, Richard Pazdur, MD, has assured sponsors that the existing regulatory framework could allow for use of pivotal trials that move outside the bounds of histology. “We do not want to look [like] the regulatory pathways are inhibiting the progress of molecular pathways or molecularly directed therapies. One does not need to have approval for gastric cancer or colon cancer or a specific subtype of colon cancer,” the Office of Hematology and Oncology Products director said on November 4, 2013 at the Institute of Medicine’s National Cancer Policy Summit. “There is nothing in the legislation that would prohibit us from approving a drug for inhibiting pathway X or inhibiting pathway Y without any reference to a known established disease here. So that is open, we do have that degree of flexibility.”

1. Still, it might be a hard sell to use a histology-independent approach in applying for approval across more common cancers, recognizing that tumors can respond differently based upon their histology. In the near term, the FDA may be lukewarm to the idea of multi-histology studies in large tumor types, so the approach is best reserved for cancers with high unmet need and few effective therapy alternatives.

Commercial Issues

A number of commercial issues will also have a bearing on a company's histology-independent development strategy, extending from market development requirements and geographical considerations to pricing and reimbursement decisions.

A thorough understanding of the current and future evolution of individual tumors is critical in determining the value proposition of clinical data coming from a histology-independent study. This market assessment must consider competitor development strategies and the impact of their data on one's own study design. Also – especially when multiple rare cancers are indicated – companies may need to adopt different models for engaging stakeholders (e.g., physician education, patient advocacy) to drive appropriate utilization. Identifying and engaging target patients may be as challenging as with orphan diseases. And, because the prevalence of specific mutations or rare cancers may be concentrated geographically, marketing and educational programs will need to be directed accordingly.

With simultaneous approval for multiple indications, a company may be faced with some difficult pricing trade-offs. Is it better to base the product’s price on its use in a rare tumor with high unmet medical need (relatively high price per patient) or in treating a larger set of target patients suffering from more common forms of cancer (relatively low price per patient)? Another challenge may be in convincing payors to accept "less robust" endpoints, such as ORR.

All of the above challenges have been met in prior specific situations, so none are insuperable. One key to success is to work closely with regulatory agencies from the beginning to understand their views and align the approach. Although some initial groundwork has been laid, a broader dialogue between biopharmaceutical companies, the FDA, and the European Medicines Agency (EMA) is needed to clarify expectations and reduce risks.

When Are Histology-Independent Studies Worth Exploring?

While a histology-independent clinical trial is not recommended when seeking approval to treat a high-incidence tumor in a drug’s first indication, there are several potential applications that do make sense. We've identified some below:

Potential Application #1: In a collection of rare tumors that share a common gene target

Studying multiple tumors simultaneously – with some rarer tumor types in the mix – is one way to test efficacy in a range of indications without investing substantial time and resources. Novartis adopted this approach with imatinib. (See sidebar, “Gleevec: A Groundbreaking Example.”) Conducting an entire trial in a rare malignancy such as dermatofibrosarcoma protuberans (DFSP) would have been time-consuming, expensive, and high risk.

The emergence of new molecular targets will open other opportunities for this type of study design. (See Exhibit 3.) Some possible promising future applications include phosphoinositide 3-kinases (PI3Ks), a family of proteins involved in the regulation of cell survival, growth, metabolism, and glucose homeostasis. They play an active role in a number of tumors, some of which have a high incidence (e.g., multiple myeloma, diffuse large B-cell lymphoma, chronic lymphocytic leukemia, follicular lymphoma) and which may warrant their own clinical trials. However, a histology-independent approach could be used for less common tumors such as marginal zone lymphoma, mantle cell lymphoma, Waldenström's macroglobulinemia, Burkitt’s lymphoma, and primary CNS lymphoma where it may not be feasible to run independent trials.

Exhibit 3

Potential Applications Of Histology-Independent Approaches Across Rare Tumors

IMS

Potential Application #2: In tumors without an effective treatment solution, but with a solid gene target This application is promising because there is a lower regulatory bar for agents targeting diseases without treatment alternatives. For example, imatinib’s approval in dermatofibrosarcoma protuberans was based on data from only 18 patients (12 from a Phase II histology-independent trial and six from independent case reports). Although this is an extreme case, it suggests that the FDA may be more lenient when evaluating a new agent for a disease with high unmet need and limited effective alternatives.

Having an identified solid gene target prior to patient enrollment is critical. Because a genetic marker is the primary criterion for selecting patients for a histology-independent trial, assays must be well developed and tested at the time of study initiation. Moreover, it may be beneficial to partner with a central reference lab to ensure every part of the testing value chain (tissue collection, preparation, and storage; testing; and results interpretation) is managed carefully. Any breakdown in this chain may impact the quality of the final clinical results.

Potential Application #3: With shelved agents that may be effective against rarer tumors

A third application is to re-examine agents that have already failed in development. Some, for instance, may have failed due to limited scientific understanding at the time, but may actually present applications in other tumor types. This strategy is especially useful in addressing rarer tumors that would otherwise not be commercially attractive.

Several companies (GlaxoSmithKline being one) are sharing their clinical data with outside investigators – a step that could lead to the discovery of subpopulations that could benefit from shelved drugs. The initiative Discovering New Therapeutic Uses for Existing Molecules, led by the National Center for Advancing Translational Sciences (NCATS), is an example of the type of program that could use histology-independent studies to identify new therapies for rare cancers (assuming expansion of its therapeutic focus to include oncology).

Potential Application #4: In low-frequency mutations within small and large tumors One such example is Pfizer’s Xalkori (crizotinib), which was approved based on strong efficacy data in EML4-ALK-positive non-small cell lung cancer (NSCLC) patients (representing roughly 3% to 5% of all NSCLC patients). Newly emerging data suggest that the drug has potential benefits in other tumor types, such as neuroblastoma and anaplastic large-cell lymphoma. In retrospect, it may have been possible to pursue a single, histology-independent study to gain approval in all three indications.

Determining When To Use A Histology-Independent Design

Despite its many benefits, the histology-independent approach is not appropriate for pivotal studies in all tumor types and indications. So, before moving forward with a histology-independent approach, sponsors should carefully consider a number of scientific, regulatory, and commercial factors.

Which Tumor Types Are Appropriate?

Science should lead the way in determining which tumors (both large and small) will be most relevant. The first step is to identify and characterize the relationship between the populations that express the biomarker and their response to drug therapy, as biomarker positivity itself does not necessarily correlate with response. The second step is to define, across tumor opportunities, the incidence/prevalence threshold above which regulators are unlikely to accept a histology-independent approach. The FDA likely accepted Gleevec's trial design because the agent targeted low-incidence tumors in the refractory setting for which very few treatment options existed. When this is not the case and the addressable population for a tumor type is large, regulators will probably want a single histology study, stratified by biomarker status and with compelling endpoints such as OS and PFS.

Given the universe of potential histologies that can be explored in a histology-independent approach, sponsors can opt for hybrid designs that reduce the size and risk of the study. For example, companies can choose a limited set of histologies upfront or conduct early proof-of-concept studies to select a sub-set tumors to investigate downstream.

What Is The Commercial Opportunity?

A thorough understanding of the market size, level of unmet need, competitive dynamics, and pricing/access implications within each tumor is a prerequisite to developing a business rationale for a histology-independent approach. Sponsors need to define the patient populations that should be targeted, choosing between relapse/refractory settings or earlier lines of therapy. They must also understand how the emerging competitive situation will influence the future treatment landscape and the value proposition of the drug in development. And, lastly, they must gain a sense of how payors will respond to the efficacy data from these studies (e.g., ORR).

What Are The Study Design Parameters?

Given the complexities of designing and executing histology-independent studies, it is critical to work closely with regulatory authorities to minimize delays and risks of failure.

Choosing the most appropriate comparator drug based on the current and evolving standards of care influences both regulatory acceptance and commercial relevance. Historically, multi-histology studies have targeted the relapse/refractory settings where regulatory requirements are less stringent (single-arm, open label studies).

Another issue is that trading the accepted endpoints of OS and PFS in single-histology studies for the blended outcomes (e.g., ORR) of multi-histology studies may pose difficulties in gaining approval and access. These challenges can be mitigated by prioritizing the histologies that generate the most significant therapeutic effect and by setting the bar high for entry into the study.

Also, especially in rare tumors, it may be necessary to work with multiple sites in order to recruit sufficient numbers of patients. Before including a specific cohort into a study, sponsors should evaluate recruitment hurdles carefully to minimize delays. Additionally, studies involving multiple histologies must allow for adaptable patient selection and enrichment of patient populations as needed. This necessitates having a pre-defined process for evaluating incoming data in real time as well as having a decision framework for prioritizing which histologies to proceed with (and possibly enrich).

Is Our Organization Prepared?

Companies planning to adopt a histology-independent approach will need to think and work differently throughout the development and commercialization process. Key considerations include:

• Creating a data analysis process and decision framework to evaluate and prioritize histologies in real time, as the data become available.
• Ensuring that a validated biomarker will be ready to support the clinical research and then, ultimately, the product’s commercialization.
• Integrating the consideration and development of pharmacodiagnostics into the drug development process.
• Developing commercial programs to find and gain access to patients – especially in rare/orphan disease areas.

Conclusions

There are opportunities to employ a single, histology-independent study to accelerate development and gain approval across multiple tumor types. Doing so yields significant benefits across the compound's life cycle – beginning with compressed development time lines, reduced investment, and lowered risk and leading to marketing synergies and rapid penetration. Most important, it provides a feasible way for companies to conduct well-powered studies in rare cancers and deliver beneficial treatments for cancer patients.

While the FDA is not against the concept, neither has it provided any guidance on how companies should proceed. The fact is, histology-independent studies are innovative and new, so there is no great body of experience on which to draw. Therefore, there’s a need for greater dialogue, additional learning, and continued evolution before this approach can be used to its full potential. The thinking here should serve as a basis to begin that process.

Andy Wong, PhD ([email protected]) is a Principal in IMS Consulting Group’s Strategy & Portfolio Analysis Group and is based in New York, NY. William Leaf-Herrmann, PhD, ([email protected]) is a Senior Principal in IMS Consulting Group’s Strategy & Portfolio Analysis Group and is based in Cambridge, MA. Lindsay Wich Leahy, MPH ([email protected]) is an Associate Consultant with IMS Consulting Group and is based in Chicago, IL.