Five Steps for Structural Reform in Clinical Cancer Research

Abstract

Despite advances in the prevention and early detection of cancer and the treatment of some malignancies, clinical research has not yet delivered treatment benefits of the magnitude anticipated after the launch of imatinib, which established highly effective new treatment standards. The primary impediments to progress are scientific, but the efficiency of research is also affected by structural deficiencies relating to where and by whom it is conducted, as well as how it is organized and regulated.

To optimize the research environment and maximize the benefits of improved funding, adjustments in the roles of government, industry, the academic community, national research bodies, and regulatory authorities are needed. A patchwork of reforms that are enabling in character and build on existing expertise can deliver substantial progress without the need for radical intervention.

THE DESIRE OF THE NEW US political administration to find “a cure for cancer in our time”¹ will provide a welcome impetus to research. Significant progress has been made during the past decade with respect to screening, prevention, basic biology, and the treatment of some cancers^2–4; however, notwithstanding these successes, there have been only modest advances in the treatment of most malignancies, and the claim that gradual progress is being made and cures will emerge at some future point may appear to skeptics to have become a jaded mantra. The targeted therapy paradigm has yet to deliver the benefits many anticipated after the launch, now 9 years ago, of imatinib, which led to changes in treatment standards for chronic myeloid leukemia and gastrointestinal stromal tumors.⁵ Other agents such as bevacizumab, sunitinib, bortezemib, cetuximab, and erlotinib have limited benefits for most patients, generally extending survival in advanced cancer by only a few months.^6–10

The new political agenda provides a timely opportunity to consider why interventional cancer research is not progressing at a satisfactory rate. The question requires consideration at a number of interlocking levels, including not only the scientific, conceptual, and technological but also the political and economic and, finally, the structural and cultural (Figure 1). Here I focus primarily on these latter aspects of the research environment and, in particular, on clinical research and its structural conditions, including where and by whom research is done, how it is organized and regulated, and how the varied constituencies within the overall research enterprise interrelate. Optimizing these aspects of research will be vital to securing the full benefits of increased federal funding.

FIGURE 1.

Selected domains influencing progress and stagnation in cancer research.

Note. The output of cancer research is conditioned and constrained by a diversity of factors, including those shown here, each of which influences and is influenced by the others. The term “facilitative science” refers to research auxiliary to basic and interventional science, such as studies of clinical trial designs, biomarkers, and endpoints. “Structural factors” are those relating primarily to the organization and infrastructure of cancer research.

SCIENTIFIC LIMITATIONS IN RESEARCH

Our understanding of cancer is advancing rapidly and systematically, as part of the ongoing revolution in molecular and cell biology. Nonetheless, the scientific challenges confronting basic and interventional research remain immense, which has structural repercussions. Molecular studies of cancer cell states have identified many potential therapeutic targets,¹¹ but our knowledge of most signaling pathways and the principles of cell cybernetics remains tenuous. Consequently, our understanding of how to manipulate cell states is poor, and it is not surprising that most emerging therapeutics have only modest benefits.

Knowledge of host immunological responses, the tumor microenvironment, and tumor–host biology is also limited, impeding the development of immunotherapies and agents such as angiogenesis inhibitors.¹² Molecular subtyping of malignancies remains in its infancy.¹³ The strategy of targeting single signaling pathways may yet prove to have only limited potential in the treatment of many common cancers, and more sophisticated approaches involving multiple interventions may prove necessary, building on principles that have informed treatment in other fields such as HIV.¹⁴ Furthermore, there has been only modest commercialization to date of biological approaches outside the targeted therapy paradigm, such as those directly using host immune cells.^15,16

The limited effectiveness of many emerging treatments has adverse implications for clinical research. Large clinical trials are required to demonstrate small benefits, early identification of promising agents is challenging, it is becoming increasingly difficult to demonstrate effects on overall survival as the number of weakly active second- and third-line treatments increases, and there are difficulties in validating surrogate endpoints.^17–19 Until more effective therapeutics are developed, or more efficient and sensitive modes of clinical assessment are devised, the pace of clinical research will not improve and the cost of developing therapeutics will remain disproportionately high, as will the retail price of these agents.

STRUCTURAL AND CULTURAL IMPEDIMENTS

A variety of structural and, to a lesser extent, cultural aspects of clinical cancer research compound these scientific difficulties (Table 1). One key group of structural challenges involves the pharmaceutical industry (hereafter pharma). Pharma is dedicated to improving patient outcomes, but the burden of developing novel therapeutics has become excessive. Recent estimates of the attrition rate for compounds in oncology range from 77% to 92%, and the average cost of successfully developing a novel therapeutic, when failures are taken into account, is more than $800 million.^20,21 The burden on pharma is due in part to its own failure to develop more effective, or patient-subtyped, treatments whose utility could be demonstrated in smaller and shorter trials; however, it is also a function of regulatory requirements, as discussed subsequently. Pharma's predicament is exacerbated by declining periods of market exclusivity as the duration of premarket development increases.²¹

TABLE 1.

Selected Structural Issues Within Contemporary Clinical Cancer Research

Sector or Constituency	Current Structural Issues
Government	Need to reinvigorate co-coordinating functions after period of stagnation
Pharmaceutical industry	Large, slow, expensive clinical trials
	Onerous burden of developing novel therapeutics
	Focus on individual products rather than treatments
	Focus on targeted rather than biological therapies
	Need for better identification of responsive patient/tumor subtypes
	Acceptance of mediocre but marketable interventions
	Need for more original research alongside established paradigms
	Need for greater cooperation within industry and academia
Noncommercial clinical research (independent, charitable, and NCI)	Shortage of funds
	Shortage of truly effective therapies entering clinical trials
	Large, slow, expensive clinical trials
	Inability to conduct comprehensive treatment optimization research program owing to the cost of therapeutics, duration of trials, and volume of emerging products
Regulators (FDA)	Shortage of funds
	Excessive administrative load
	Standard approval procedures/philosophy problematic with respect to cancer research
University spin-out/biotechnology sectors	Inefficient translation of promising university science into viable therapeutic models
	Financial viability of companies/innovations in the absence of revenue generation
	Achieving viable commercialization/business models
Industry-funded nonprofit sector	Limited development to date; contribution of industry inadequate as yet with respect to scale of patient need

Note. FDA = Food and Drug Administration; NCI = National Cancer Institute.

These circumstances suggest a need for reform; however, pharma's contribution to clinical oncology is also influenced by its internal traditions and culture, and here too reform may be required. Large corporations are generally geared to the development of individual drugs rather than more complex, multi-intervention treatments, and they may give insufficient weight to multitherapeutic thinking during basic research or clinical development. Pharma corporations may also be risk averse with respect to the balance between conservative development programs that yield modestly active but marketable agents and more original research with less immediate commercial potential.

In addition, pharma exhibits greater interest in targeted therapies than individualized biological approaches because the targeted therapy concept is both better aligned with traditional pharmaceutical therapeutics and more commensurate with existing manufacturing and sales models. The industry has also performed poorly in identifying narrow but more responsive patient groups and developing personalized treatment algorithms. Finally, pharmaceutical corporations may adopt an unduly proprietary approach to clinical data and biospecimens and devote insufficient interest to commercial or academic collaboration. The industry is highly committed to cancer research, but, notwithstanding the constraints imposed by regulatory requirements and market realities, its approach and output have been suboptimal with respect to the scale and gravity of patient need.

A second group of structural issues concerns noncommercial clinical research, including government-run, government-funded, charitable, and independent academic projects. In the United States, the National Cancer Institute (NCI) and subsidiary bodies have organized a well-funded program of in-house and independent research.^22–24 Whereas industry's traditional role has been to develop individual therapeutics, the NCI focuses its clinical programs on treatment optimization and has conducted important trials of combined targeted therapies.²³

Such research is, however, constrained by the same problems of trial size, duration, design, and endpoints faced by industry trials. Moreover, the cost of purchasing cancer therapeutics is high, and pharma companies are unlikely to provide support for investigations that might constitute a commercial risk (e.g., head-to-head and combination studies). The implicit settlement in which the industry develops drugs and academics optimize their use is under strain in clinical oncology, because the resources of noncommercial science are inadequate to process the burgeoning number of highly expensive, slow-to-evaluate products efficiently.

A third set of structural problems pertains to regulators, notably the Food and Drug Administration (FDA) in the United States.²⁵ The increasingly detailed information required during the regulatory process has gradually increased the burden on industry and regulators alike.^21,26 In addition, several basic aspects of regulatory procedure are poorly adapted to the needs of contemporary oncology.

The traditional approach to evaluating cancer products in phase II and phase III programs, which begins with proof of single-agent activity in second- or third-line treatment before proceeding to first-line use, may become increasingly problematic if the orchestrated use of several agents is required to elicit a clinically significant response and as molecularly defined disease subtypes grow in importance. Although FDA approvals are contextualized—for instance, oxaliplatin was approved for advanced colorectal cancer in the setting of a specific fluoropyrimidine regimen²⁷—and provision can be made for combinations of agents when individual products are inactive, the focus on single products remains dominant within the approval process, contributing to the shortage of combined-intervention research, particularly at the preclinical and premarketing stages.

Regulatory procedure is also problematic with respect to the threshold for marketing approval. Historically, this has involved a single, critical hiatus that, once achieved, provides insufficient incentive for pharma to provide support to subsequent optimization studies, particularly if there is a risk the results may be commercially disadvantageous.

A closely related issue is the question of how active a product should be to secure approval. The 1962 Kefauver–Harris Amendment required evidence of a substantial benefit,²⁶ but recent practice in oncology has been to require only modest therapeutic effectiveness; for instance, erlotinib was approved for the treatment of advanced pancreatic cancer despite prolonging median overall survival by only 2 weeks.²⁸ This practice is in part a reflection of the reality that more effective treatments are unavailable, but it also reflects the need to sustain industry investment and has the benefit of making available agents that might contribute to more effective multiagent strategies or be customized to smaller but more responsive patient groups.

Conversely, however, a low threshold for approval could be viewed as providing pharma with an incentive to develop weak agents, to avoid incisive research such as patient enrichment studies that might erode sales potential, and to continue to conduct large, slow trials. It is the difficult role of the regulator to reconcile these considerations, but the height of the bar for full approval may now need to be raised with respect to efficacy.²⁹

CULTURAL CHANGE AND THE “WAR ON CANCER”

How can these structural and cultural difficulties be addressed? Given the scale and heterogeneity of clinical cancer research within the market economy, a patchwork of reforms targeting specific challenges and opportunities is preferable to a grand plan. Actions should be parsimonious and enabling rather than crudely interventionist, and they should be based as far as possible on existing practices and expertise.

The NCI and FDA have key roles in addressing the structural challenges of clinical cancer research. Both have already fostered beneficial structural and cultural evolution, as well as driving scientific innovation. The NCI has launched numerous initiatives, including the Translational Research Working Group, the National Institutes of Health “road map” initiatives, and basic science programs ranging from the Cancer Genome Atlas to think tanks in cancer biology.^30–33 The FDA's Critical Path Initiative, which promotes scientific, technological, and manufacturing innovations across all areas of medicine,^34–36 is of particular importance to clinical oncology. In addition to FDA and NCI programs, numerous other initiatives have been launched in cancer research. For instance, the C-Change program has for more than a decade sought to coordinate the diverse constituencies of cancer research and striven for progress in cancer prevention, detection, and treatment.³⁶

These programs have had a beneficial impact on the organization of commercial and academic research, in addition to encouraging the growth of industry-funded nonprofit initiatives as well as institutions such as the Critical Path Institute and the Foundation for the National Institutes of Health Biomarkers Consortium.^37,38 Despite such developments, however, the pace and output of cancer research remain inadequate, and more substantial measures should be considered.

As a preliminary step toward more specific interventions, a cultural and attitudinal sea change is required in the way clinical cancer research is professionally comprehended and publicly represented. It should be acknowledged that research has not met expectations and that the death of many millions worldwide each year constitutes an ongoing public health calamity. An overt redeclaration of the “war on cancer” may not be desirable, given that the phrase recalls the unrealistic expectations associated with the 1971 initiative of President Richard Nixon³⁹ and is also suggestive of the recent “war on terror.” However, a sense of urgency and collective purpose is desirable and should inform policy with respect to executive mobilization and the respective support accorded to narrow vested interests and broad public health priorities.

Furthermore, although systematic, incremental studies should remain central to research, greater innovation must also be promoted, particularly with respect to models of intervention and their transition to the clinic. Finally, prevention, screening, and early detection programs will remain of great importance alongside the basic and interventional research issues discussed here.

FIVE PROPOSALS FOR STRUCTURAL ADJUSTMENT

To improve the structural conditions of interventional cancer research, adjustments should be considered with respect to the roles of government, industry, the academic sector, and regulators (Table 2). Several of the measures described in this section have been formulated as specific proposals, but they are also intended to illustrate types of interventions capable of fostering structural reform.

TABLE 2.

Examples of Measures to Facilitate Structural Reforms in Clinical Cancer Research

Example Measure	Objectives	Features
Cancer research review body	Advise government on cancer research funding/strategy	Established by and reporting to government
		Permanent secretariat
	Create combined platform for policy contributions by NCI, FDA, etc.	Faculty from NCI, FDA, professional societies, leading journals, patient groups
	Assess progress	Recommendations on research policy/specific initiatives
	Provide information for health care specialists, lawmakers, and public debate	Develop/curate inventory of current cancer research
		Periodic report on state of research, unmet needs, “blue-sky” thinking
	Foster unity and common purpose
Rapid action unit	Expedite measures to support effective research (e.g., integrated trial databases, biospecimen repositories)	Established by and reporting to government
		Expedites specific projects recommended by review body and agreed to by government
		Identifies and manages steps necessary to realize specific projects
Support for basic and early interventional research	Improve quality/number of emerging therapeutics/treatments	Optimized mix of incremental/innovative research
		Reviews therapeutic potential of current academic research
		Increased support for university spin-out/biotechnology sectors
		Increased support for NCI/FDA translational/facilitative research initiatives
Modification of FDA approval criteria	Support early/niche market access while encouraging development of more effective products/treatments	Revisits basic principles of phase III trials/single-agent activity
		Increased emphasis on treatments alongside individual drugs
	Support multiagent treatment optimization	Raised bar for full approval
		Multistage, graduated approval with increasing marketing rights (“star” system to clarify degree of benefit for patients/purchasers)
		Public research program: assignment of drug to NCI/ODAC for use in optimization studies to run in parallel with FDA clinical assessment
Support for industry	Incentives for pharmaceutical industry	Simplify approval procedure for obtaining initial, limited marketing rights
		Indemnify liability during public research program assessment
		Prospect of expansion to new indications
		Extend market exclusivity for highly effective treatments
	New business models for biologics and personalized medicine	Encourage pharmaceutical industry development of new business models
		Support university spin-out/biotechnology sectors
		Consider approaches to supporting/creating novel for-profit enterprises

Note. FDA = Food and Drug Administration; NCI = National Cancer Institute; ODAC = Oncologic Drugs Advisory Committee.

Proposal 1: Establish a Cancer Research Review Body

The federal government should consider establishing a review body with responsibility for evaluating the overall state of the campaign against cancer and making funding and policy recommendations. The body should have a modest permanent secretariat, but its faculty should be formed of representatives of government, the NCI and FDA, professional societies, industry, leading journals, and patient groups. The administrative load should be carried by the secretariat, with no additional burdens placed on the contributing organizations. The C-Change organization should also be consulted regarding the formation and membership of the review body.

The review body should provide a shared conduit through which its component constituencies can contribute to policy. An additional function should be to ensure that the status of the campaign against cancer is publicly clarified. Accordingly, the secretariat should establish an inventory of current research, including basic, translational, and clinical studies. The body should also periodically publish a short, nontechnical, strategic overview of the campaign with policy recommendations. This report should complement recent NCI reports^40,41 by appearing in a scheduled series and evaluating research strictly against the goal of curing cancer rather than in terms of advances over previous standards. Such a report would help maintain the public profile of cancer research, support a sense of shared purpose and identity throughout the many constituencies engaged in cancer research, and provide a reference point for critique and debate.

Proposal 2: Create a Rapid Action Unit to Drive Specific Initiatives

When priority actions to support cancer research are identified (whether through the suggested review body or alternative mechanisms), they should be expedited rapidly. One approach to achieving this would be the establishment by the government of a rapid action unit capable of swiftly identifying and managing the steps necessary to realize specific objectives.

For instance, it would greatly benefit cancer research if integrated multitrial databases and biospecimen repositories were established for all major cancers, pooling data from phase II and phase III clinical trials, including those conducted by industry, with due consideration for commercial and patient confidentiality. If such a step were recommended by the review body and approved by the government, it would be the task of the rapid action unit to identify the requisite legal, logistical, and funding steps, as well as appropriate incentives, in liaison with all stakeholders and partners and to oversee their timely implementation. Such a unit, which would require a mix of legal, scientific, and organizational skills, might constitute a useful model for expediting and co-coordinating urgent projects across other areas of government.

Proposal 3: Provide Support for Basic and Applied Research

Clinical oncology is in urgent need of better therapeutics. This requires an improved understanding of cancer biology and the generation of clinically testable models of intervention more rapidly and in greater numbers than has been achieved to date. Detailed discussion of these challenges is beyond this article's scope, but organizations such as the NCI should debate publicly whether they are achieving the optimal mix of conservative and more adventurous research, both in pure and applied settings.

Academic institutions should also review their research and consider whether the possibilities for interventional science are being fully realized. In addition, there is a need for the government to review whether the university spin-out (i.e., commercialized applied academic research) and biotechnology sectors are functioning adequately, from the standpoints both of medical–scientific opportunity and finance. Finally, there is a need to review how the NCI and FDA translational and facilitative research initiatives, such as the Critical Path Initiative,⁴² can be optimally supported.

Proposal 4: Adjust Regulatory Procedures to Support More Incisive Research

Regulatory approval procedures have a significant impact on clinical and preclinical research. They have the potential to drive the generation of more effective treatments and to increase the vibrancy of academic research, but the burdens they place on industry and on regulatory bodies themselves should be minimized. Current FDA practices and approval criteria should be evaluated with respect to their degree of reliance on large phase III trials and the demonstration of single-agent activity in diverse patient populations. It should also be considered whether the provision of trial data and biospecimens to academic repositories should, with due consideration for commercial rights and patient privacy, become mandatory for new drug applications. I propose 3 additional adjustments (which would require legislative support) for discussion.

First, the bar for full approval should be raised, such that only highly effective products (or multiproduct treatments) comparable in efficacy to imatinib are granted full marketing approval. Second, rather than denying marketing rights to treatments that fail to meet this ambitious standard, the all-or-none approval philosophy should be replaced by a graduated system in which less effective products can be approved, but with more limited marketing rights. For instance, direct-to-patient and television marketing might be available only for treatments achieving full approval. Allocation of a certain number of “stars” could readily identify to purchasers and patients which treatments were considered highly effective by regulators. Clear identification of truly effective treatments would act as a strong incentive to industry to move beyond the development of modestly effective mass-distribution therapeutics while preserving market viability.

Finally, 1 approach to achieving more effective multiagent treatment optimization and driving the development of more active interventions would be a requirement that all therapeutics other than those achieving full approval commit to “public research program” assessment, running in parallel with advanced phase II through phase IV clinical development (i.e., all randomized clinical trials designed to evaluate efficacy and safety) and managed by the NCI in consultation with the FDA and its Oncologic Drugs Advisory Committee. Within the proposed public research program, companies would be required to contribute a supply of their product to the program at discounted rates or free of charge. The drug would be used both for in-house NCI research and in studies by bidding academic groups. Companies should have no veto over the studies carried out, which might include phase II combination therapy studies with competitor products, head-to-head comparisons, and more extensive patient subtyping analyses.

However, products should be marketable (subject to provisional regulatory approval) during this period, provided use was monitored and clinical data provided to the FDA and NCI. “Public research” status might indeed be a useful marketing tool. Within these proposed arrangements, approval for initial, limited market access should be simplified and the entry bar kept low for products with therapeutic potential in combination regimens or subgroups.

If regulatory adjustments such as those just described were adopted, noncommercial clinical research would benefit greatly. Highly effective drugs and multidrug treatments would pass rapidly through graduated approval to unrestricted marketing, and the economics of developing niche products would become more favorable. Of note, moreover, most of these recommended adjustments are extensions of existing principles developed by the FDA. Graduated approval is an extension of the established concept of accelerated but provisional approval,⁴³ whereas the increased emphasis on treatments alongside therapeutics extends the practice of specifying the management framework in which a novel therapeutic can be marketed.^27,44

The requirement for companies to provide a supply of their product for use in optimization studies approved by the NCI and the FDA's Oncologic Drugs Advisory Committee is a more substantial step, but one that is necessary to help ensure that effective multi-intervention and patient-subtyped treatments are identified more efficiently. To prevent such arrangements from discouraging investment in oncology, similar measures should be considered in other areas of medicine and required for overseas companies seeking access to the US market.

Proposal 5: Support Industry Through Regulatory Adjustments and New Business Models

Alongside these measures, pharma requires greater support. The regulatory process should be simplified to reduce development times and costs. Wardell et al.²¹ have proposed how regulatory procedures could be simplified at the investigational new drug and new drug application stages. Public research program assessments should offer the prospect of rapid expansion into new clinical settings or disease states as a result of investigation by independent groups; other incentives might include indemnified liability for public research program studies and prolonged market exclusivity for highly effective products and treatments achieving full approval, a policy building on measures already used in the context of pediatric indications and orphan drugs.⁴⁵ Extensions to market exclusivity could also be used to reward companies for committing promising agents to the approval process at an early stage.

Pharma should also be encouraged in the development of new business models. Currently, some corporations are seeking to develop innovative biologics businesses or adopt more flexible and personalized approaches to medicine with increased reliance on external innovators.^46,47 Although such trends are primarily a matter for the market, government should consult with pharma to identify how its efforts to develop more innovative approaches to cancer research can best be supported. With respect to the university spin-out and biotech sectors, the government should consider how, in the absence of sustained private finance or pharma support, it can best foster the establishment of viable for-profit organizations dedicated to the commercialization of biological and targeted strategies.

POLITICAL WILL, PRACTICAL ACTION

The case for improved federal funding of clinical cancer research has been eloquently made by Winer et al.⁴⁸ However, improved funding will have the greatest benefit if it is accompanied by a renewed urgency, a culture of innovation, and, above all, structural adaptation. Recent events in the global financial system illustrate the need for appropriate regulations to ensure that industries operate efficiently, sustainably, and for the public good. Nonetheless, in the past decade the regulation of the pharmaceutical industry has become burdensome with respect to product development and marketing approvals while simultaneously failing to encourage the development of highly effective treatments. During the same period, independent academic research has become underfunded and increasingly marginal to the clinical development process. The NCI and FDA have succeeded in driving innovation in a challenging era, but government must provide increased support and make greater use of their expertise and leadership.

If these varied issues are to be addressed, political will must be translated into practical action. A patchwork approach is preferable to a grand plan, and the examples offered here illustrate the types of interventions with the potential to have a beneficial impact on the spirit, culture, organization, regulation, and commercialization of research. Cancer is not business as usual but a deadly and immediate public health crisis; as such, it requires sustained governmental attention and provides a critical test of the commitment to “restore science to its rightful place, and wield technology's wonders to raise health care's quality and lower its cost.”⁴⁹