|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology
The Phase III Trial in the Era of Targeted Therapy: Unraveling the "Go or No Go" Decision
From the Division of Hematology/Oncology, Massachusetts General Hospital, Harvard Medical School, Boston; and the Program on the Pharmaceutical Industry, Massachusetts Institute of Technology, Cambridge, MA. Address reprint requests to Thomas G. Roberts Jr, MD, Massachusetts General Hospital, Cox 640, 100 Blossom St, Boston, MA 02114; e-mail: troberts{at}partners.org.
Purpose: To review characteristics of contemporary phase III oncology trials and create an explicit framework to help clinical researchers prioritize novel therapies for phase III testing. Methods: We searched the MEDLINE and EMBASE databases for all reviews of phase III trials; cataloged all phase III trials in two national clinical trial databases; and reviewed approval criteria of recently approved oncology drugs from public data provided by the US Food and Drug Administration. Industry data not available elsewhere in the medical literature were obtained from a sourcebook published by a large contract research organization. Results: Phase III oncology trials are the most expensive and time-consuming aspect of the drug development process. The results of these trials continue to exert the greatest influence on the treatment decision of oncologists and remain pivotal to the granting of drug approval. Making optimal decisions about which agents to advance to phase III testing may decrease the overall cost of cancer drug development and limit the number of patients exposed to ineffective drugs. A conceptual decision model for prioritizing novel therapies for phase III testing is presented. Conclusion: Cancer drug development has become more complex and expensive, whereas overall clinical progress remains slow. The transition from phase II to phase III requires a strategic decision that is based on new considerations. A greater investment in phase I and II drug trials may be required to provide the information necessary for phase III planning.
IT HAS been almost 50 years since the National Cancer Institute (NCI) initiated, in 1954, the first randomized trial in medical oncology.1,2 It would be difficult to overstate the contribution that randomized trials have made to advance cancer care since then. There have been dramatic instances when novel agents transformed the treatment of a disease before phase III testing (eg, imatinib mesylate [Gleevec; Novartis, East Hanover, NJ]), but these have been rare. Most advances have instead been incremental; sufficiently powered phase III trials have been required to establish clinical benefit.3,4 These trials are now particularly critical to the testing of molecularly targeted therapies (MTTs) that may exhibit little evidence of activity in preceding developmental trials (phases I and II). Although it is obvious that randomized testing of all potentially useful novel therapies is not feasible, cost effective, or ethical,5 there is no emerging consensus on how to prioritize novel targeted therapies for phase III testing.6–8 This has become especially true for the subset of MTTs that are neither cytotoxic nor proapoptotic and therefore produce modest changes in tumor size.9–11 The lack of consensus on the testing of these agents coincides with increasing criticism of the national cancer drug development system.6,12 Despite the expenditure of billions of dollars by government and industry on cancer drug discovery and development, only two to five oncology-related drugs are approved each year, a number that is neither increasing nor sufficient to meet the needs of the majority of cancer patients with advanced disease. Within this context, we sought to provide a systematic overview of the characteristics of current phase III cancer treatment trials and offer an explicit framework to help prioritize novel agents for phase III testing.
Characteristics of Current Phase III Clinical Trials To obtain characteristics of contemporary phase III oncology trials, we cataloged selected information about all ongoing phase III oncology trials listed, as of May 2003, in two national databases: the Physician Data Query (PDQ) database, maintained by the NCI,13 and the ClinicalTrials.gov database, maintained by the National Library of Medicine.14 The PDQ database, with approximately 1,800 abstracts of cancer clinical trials, contains all NCI-sponsored phase II and III trials submitted to the US Food and Drug Administration (FDA) under Investigational New Drug regulations, as well as some trials supported by European cooperative groups and industry. Because the PDQ database does not capture all industry-sponsored trials, we expanded our analysis to include all industry-sponsored phase III trials listed in the ClincalTrials.gov database. This second database contains approximately 7,700 clinical trials; under section 113 of the FDA Modernization Act, all industry sponsors are required to submit clinical trial information to this database if they are testing the effectiveness of a drug to treat a serious or life-threatening disease. For those trials contained in both databases, we abstracted the data primarily from the PDQ database. We obtained data for primary sponsor, treatment modality, disease setting, line of treatment, accrual goal, and planned accrual time for all identified phase III trials. For drug or biologic treatment trials, we obtained additional information on route of administration, FDA approval status at time of trial initiation, and molecular target, if indicated. Agents with an FDA approval for any indications were considered FDA approved even if the trial was conducted in another treatment setting. Differences in trial characteristics for continuous variables were evaluated using the Student’s t test and analysis of variance.
Searching for Relevant Studies in the Medical Literature
Regulatory Data
Phase III Trial Characteristics We identified 409 active phase III trials; these included 362 trials contained in the PDQ database and an additional 47 industry-sponsored trials contained in the ClinicalTrials.gov database that were not found in the PDQ database. Patient accrual goals and planned accrual periods were available for approximately 90% and 70% of the identified trials, respectively. Summary statistics for these trials are presented in Table 1  . The median accrual goal for the identified trials was 481 patients, with a median planned accrual period of 3.8 years. Because of a small number of trials with large accrual goals (eg, a 20,000-patient accrual goal in a trial comparing prolonged v shorter treatments with tamoxifen in early breast cancer), mean accrual goals generally exceeded the medians.
Breast cancer and non–small-cell lung cancer were the two most well-represented diseases in the set of phase III trials, constituting 55 (13%) and 33 (8%) of the 481 trials, respectively. The disease chosen for testing strongly influenced the patient accrual goals and planned accrual periods: breast, prostate, and colorectal cancer trials had the three highest accrual goals, whereas prostate cancer and non-Hodgkin’s lymphoma trials had the longest planned accrual periods. There was no statistical difference between mean accrual goals for phase III trials sponsored by industry, the NCI, or the European Organisation for Research and Treatment of Cancer. However, trial sponsorship strongly influenced planned accrual periods; for example, the average planned accrual time of the industry-sponsored trials was 2.2 years, compared with 4.2 years for the NCI-sponsored trials (P < .001 between categories). The primary treatment modality under investigation (ie, chemotherapy, radiotherapy, or surgery) was not associated with differences in patient accrual goals or planned accrual periods. Of the 409 active phase III trials, we identified 316 (77%) phase III trials that primarily evaluated drugs or biologics. A minority (24%) of these trials focused on biologic agents; these trials that focused on biologics had significantly smaller patient accrual goals and accrual periods, compared with drug trials. Seventy-seven (29%) of the treatment trials focused on the evaluation of an agent not approved by the FDA at the time of trial initiation; these 77 trials had significantly lower mean accrual goals and shorter planned accrual periods, compared with the 235 trials evaluating already-approved agents. Interestingly, the 34 trials that evaluated classic cytotoxic agents did not differ significantly in patient accrual goals or planned accrual periods from the 282 trials that evaluated agents with identified molecular targets. We also analyzed trial characteristics that were based on the intended treatment line (ie, first line, second line, or adjuvant). Adjuvant trials had the largest mean accrual goals (1,895 patients), whereas second-line or greater trials had the smallest mean accrual goals (478 patients).
Factors That Are Stimulating Demand for Phase III Trials
In addition to the scientific need to compare clinical benefit of these novel treatments, the FDA will continue to require randomized trials of novel agents in their granting of marketing approval. Historically, randomized controlled trials have provided the basis for approval of approximately two-thirds of the more than 85 currently approved oncology drugs in the United States. In fact, full marketing approval has typically required two positive phase III trials.21 Since 1992, there has been a mechanism for accelerated marketing approval of cancer drugs on the basis of surrogate end points, such as response rates in phase II trials, provided that the surrogate is deemed likely, usually in the judgment of the Oncology Drug Advisory Committee, to predict improved survival or symptom palliation.22,23 Although 12 of the 29 drugs approved from 1995 through 2002 have used the accelerated approval mechanism,19,24 there are three reasons the FDA will continue to require phase III trials. First, accelerated approvals require that the sponsors complete postmarketing controlled trials to upgrade to full approval.16,21 Second, the accelerated approval mechanism is only available for an indication where the new therapy offers promise over the existing standard. As new drugs fill the gaps in second- and third-line indications for the major cancers, there will be less opportunity for accelerated approval. Third, because of the concern that accelerated approval creates incentive to perform large, single-arm studies in progressively refractory patient populations (as was the case with gefitinib (Iressa; Astra Zeneca, Wilmington, DE) in third-line non–small-cell lung cancer), some officials of the FDA have indicated a preference for interim analyses of phase III data as the basis for accelerated approval.24,25 The recent approvals of anastrazole (Arimidex; AstraZeneca, Wilmington, DE) for the adjuvant treatment of postmenopausal women with hormone receptor–positive breast cancer and oxaliplatin (Eloxatin; Sanofi-Synthelabo, Paris, France) for the treatment of second-line metastatic colon cancer were based on the data from such interim analyses (Table 2 ).
A third factor driving the need for phase III trials of novel agents is the recognition that many of these drugs will have modest activity in phase II. To save time and effort, it is an increasingly common strategy to abbreviate phase II testing and initiate early phase III evaluation. Two strategies in particular are dominating: testing an approved regimen with or without an MTT in the first- or second-line setting (strategy A); and testing MTTs alone versus observation in minimal residual disease states (strategy B). Of the 71 active phase III trials evaluating unapproved agents, we found that approximately 30% were employing strategy A and another 20% were employing strategy B. It is not clear how successful strategy B will be; on the basis of our review of available data, as of May 2003, the FDA has yet to approve a cancer drug (excluding supportive agents) that has not shown a clear, single-agent response rate in at least one phase II trial.
Factors Limiting the Ability to Perform Phase III Trials
Extrapolating from available data, the direct cost of each industry-sponsored phase III trial for a cancer drug can exceed $10 to $20 million. Phase III oncology trials sponsored by the cooperative groups of the NCI tend to be less expensive, but they still exceed, on average, $1 million per trial. The high cost is driven by the requirements for a large number of patients, coordination over many centers, and detailed and long-term data collection (Table 3
Not only are phase III trials costly, they also extend the time to marketing approval. For industry sponsors, delays in time to approval mean a loss of patent exclusivity time and diminished revenue potential. Using data supplied by the pharmaceutical industry and maintained by the Tufts Center for the Study of Drug Development, the mean length of phase III trials in oncology from 1990 to 1997 was estimated at 4.5 years.15 This time period was slightly longer than the mean planned accrual period of 3.9 years that we calculated for the 409 active trials identified and analyzed in our study. In an effort to speed accrual, sponsors seem to be increasing the number of open study sites, adding to the cost and complexity of the trial. In our analysis of industry-sponsored phase III trials of unapproved agents, the average number of open sites per trial was 27. Although the FDA review phase is often cited for unduly prolonging development time, recent data suggest that for cancer drugs, the clinical trials and not the review phase are responsible for increases in development time. Whereas the average time of clinical development (from Investigational New Drug filing to FDA marketing approval) increased from 5 years for the seven cancer drugs approved in the 1960s to 10.8 years for the 20 drugs approved from 1990 to 1996, the FDA approval phase decreased from 2 years in the 1960s to approximately 1.6 years in the 1990s.22 This translates into a clinical phase development time that has more than doubled when compared with development time in the 1960s, whereas FDA review times have actually decreased by 20%. The most recent data available indicate that phase III trials in oncology are now the longest among all therapeutic categories of drugs—20% longer than phase III cardiovascular trials and almost double the mean length of phase III trials for anti-infectives.15 For some sponsors, this concern about clinical development delays exerts greater influence on the go or no go decision than the trial cost itself.29 A third factor that limits oncology drug development is the relatively high probability that a new cancer drug will fail to obtain marketing approval. Of the 280 antineoplastic agents brought into clinical trials from 1975 to 1994, only 29 (10%) of them ultimately received FDA marketing approval.30 This yield translated into the second lowest approval rate of any therapeutic drug class (with drugs targeting the CNS having the lowest yield).31 It is important to note that all drug failures are not alike: the later they occur, the greater their financial impact. The most costly drug failures are those that make it to an FDA review but are then rejected. Rejections of oncology-related new drug applications are relatively uncommon, occurring approximately 20% of the time21; but the financial impact for the sponsor can be devastating, especially for single-product biotechnology companies. Almost as costly is the failure to show benefit in phase III trials. On the basis of a recent published pharmacoeconomic model, a 4% decrease in the rejection rate for novel agents could alone reduce the total capitalized cost per approved drug by $100 million.32,33 These data emphasize how late drug failures in particular drive the overall cost of drug development and ultimately lead to higher prices for those drugs that do gain approval.
Critical Factors for Entering a Phase III Trial: A Conceptual Decision Model
Specific criteria for choosing agents for testing in phase III trials are not well described in the literature. We identified only two studies that quantitatively address a method to choose anticancer agents for phase III testing.39,40 Both studies offer formal power calculations derived from the results of phase II trials as the basis for choosing agents for phase III testing, but both are specific to lung cancer, and neither has been validated prospectively. We identified an additional eight articles that address the problem more generally; but none of the analyses offered an overall conceptual model of the go or no go decision.2,11,23,41–45 Using data collected from our review, we therefore created such a model to frame the decision on areas of uncertainty and risk. We grouped decision critical factors into five key considerations unique to each agent in question (Table 4
Factor 1: The target validity and preclinical activity. Our review of 316 contemporary drug or biologic treatment trials found that 282 (89%) are evaluating an agent with at least one identifiable molecular target. Three forces underlie the target focus. First, there is a growing consensus that there is little chance that an important breakthrough will come from new cytotoxic chemotherapy.46 Second, the enormous gains in the understanding of cancer biology during the last 20 years have identified pathways of carcinogenesis, invasion, angiogenesis, and metastases that provide a new mechanistic basis for drug discovery.47 Third, clinical and commercial successes of targeted agents, including monoclonal antibodies and imatinib mesylate, have begun to justify the enormous investment in this approach (Table 5 ).48
A conceptual basis from prioritizing which agents to test in phase III trials on the basis of perceived importance of the molecular target has not been carefully described. Few targets have been adequately validated in human systems. The presence of redundant and compensatory cellular pathways, innate and acquired drug resistance, and genetic instability of tumors can all decrease the clinical value of a particular drug target.49,50 For the purposes of this framework, a target is considered valid to the extent that it is both uniquely or markedly overexpressed in the specific type of cancer (ie, tumor-specific; criterion A), and essential for the survival of the cancer in question (ie, rate limiting; criterion B). For each tumor, there may be only a handful of targets that meet these strict criteria. The risk of initiating phase III trials of targeted agents where the target has not been adequately validated is becoming increasingly clear. The early experience with matrix metalloproteinase inhibitors (MMPIs) is illustrative. At least five MMPIs advanced to late stages of development before the precise role of their targets was known. For these agents, the target (matrix metalloproteinase) was overexpressed and enzyme activity was demonstrable, but not unique to tumors (partially meets criterion A); but its inhibition did not lead to cell death in already metastatic disease (does not meet criterion B). Therefore, only one of the two criteria was fulfilled, and partially at best. Of 15 phase III trials undertaken with MMPIs, none has shown a significant survival benefit, and two have shown poorer survival as compared with placebo.51
The MMP-I experience can be contrasted with the development of imatinib mesylate for chronic myelogenous leukemia (CML). The role of Bcr-Abl kinase, imatinib mesylate’s target in CML, was rigorously characterized before the drug’s clinical development. The Bcr-Abl kinase is an activated oncogene product, which can be detected in virtually all patients with the CML. Its normal counterpart, the Abl kinase, is found in nonmalignant tissues, but it is not critical to normal physiology (meets criterion A). It is sufficient to cause the disease in animal models and absolutely required to sustain the malignant phenotype in CML cells (meets criterion B). When this target is inhibited in cell culture experiments, the malignant cells die, but normal bone marrow cells do not. Thus the target fulfills the above criteria and is deemed valid. The importance of the target coupled with imatinib mesylate’s own pharmacologic merits contributed to its speedy approval on the basis of phase I and II data: it was approved in just more than 3 years from its initial testing in humans and required less than 3 months under FDA review (Table 3 The validity of human epidermal growth factor-2 (HER-2) as a target has been established in successful trials of trastuzumab (Herceptin; Genentech Inc, South San Francisco, CA) in combination with chemotherapy. Preliminary positive results of ongoing phase III trials of bevacizumab (Avastin, Genentech) and cetuximab (Erbitux, C225; Bristol-Myers Squibb, New York, NY) suggest that vascular endothelial growth factor receptor-2 kinase domain receptor and epidermal growth factor receptor will also be valid targets, although the benefit is less dramatic than with imatinib mesylate and requires the presence of cytotoxic drugs. This set of trials is of particular importance in that it demonstrates that targeted drugs that have minimal single-agent activity in phase II trials may exert clearly beneficial effects in phase III combination trials. An interesting target may override the negative effects of modest phase II results. Factor 2: Pharmacokinetic properties in phase I and II trials. PK properties provide a second criterion that can be used to rank novel MTTs for preferential phase III testing. PKs act as a bridge integrating data from preclinical testing through early clinical trials.52 Because PK properties (eg, bioavailability, peak and trough concentrations, area under the curve [AUC], clearance, and half-life) determine the appropriate route of administration, treatment schedule, and dose to be used in later trials,53 they should represent an important component of any phase III decision. Favorable PK properties include: high percentage and consistent bioavailability, if orally administered; a long half-life (ie, greater than 8 hours, which allows for once-per-day dosing of oral agents); and drug levels at a trough that exceeds the threshold dose needed to inhibit the target in preclinical experiments. A new drug should at a minimum achieve an AUC in humans that equals the effective AUC in animals; if it does not, there is little point in advancing to phase III. These properties are necessary but not sufficient in deciding to test an agent in a phase III trial. All things being equal, however, agents with favorable PKs in phase I and II should be ranked preferentially in the decision to proceed with phase III testing.54 Factor 3: Activity and toxicity in phase I and II trials. The third criterion that can be used to rank novel MTTs for phase III testing is the activity and toxicity seen in phase I and II trials. Even in the absence of antitumor effects or demonstrable effects on its target pathway, most agents that have an acceptable toxicity and favorable PK profile in phase I trials undergo additional efficacy testing in phase II trials. For the MTTs, simultaneous phase I and II trials in combination with selected cytotoxic agents are often initiated in parallel with single-agent phase II trials to take advantage of synergy observed in preclinical studies. Among their several advantages, phase II trials allow verification of tolerability of dose and schedule before advancing to a definitive phase III effort; they also provide, it is hoped, concrete evidence of target inhibition and some evidence of antitumor activity. The results from phase II trials, even among the MTTs, still exert great influence in determining whether to advance to phase III. When outstanding activity and a superior safety profile are noted during a phase I or phase II trial in an otherwise untreatable disease (eg, imatinib mesylate in gastrointestinal stromal tumor [GIST]), phase III trials are unnecessary for initial approval. For other agents and in most disease settings, the decision to advance to a phase III trial rests at least in part on whether evidence of an agent’s biologic activity (alone or in combination) in phase I and II trials exceeds a threshold. There are several measurements of activity that can be employed for this purpose. Response rates of novel agents in developmental trials have traditionally played a major role in the decision about whether to proceed with phase III testing. (They also have contributed to many FDA approvals.) From January 1, 1990, to November 1, 2002, tumor response, both in phase II and phase III, provided a major basis for FDA approval in 26 of 57 (47%) anticancer drug claims.24 Currently, one of the major challenges facing the go or no go decision-maker is the lack of impressive response rates for many of the novel agents. Response rates may not be helpful at all in evaluating targeted agents that have growth inhibition as their primary effect (ie, cytostatic agents).8,11 Accordingly, several novel phase II designs have been offered. Rosner et al,45 for example, proposed a randomized discontinuation design for the phase II evaluation of cytostatic agents using time to progression as the major end point. In this design, only those patients whose disease is stable on therapy are randomly assigned to either continuing therapy or the administration of a placebo. When applied to the testing of a novel agent in the treatment of renal cell carcinoma, this approach proved feasible and allowed for smaller sample sizes because it minimized patient heterogeneity by randomly assigning only those patients with stable disease. Among the novel agents that do effect tumor regression, evaluating antineoplastic drugs by response rates as a surrogate to clinical benefit relies on the prediction that tumor shrinkage will lead to either lengthened survival or improved quality of life.11 Although this prediction seems to hold true for hematologic malignancies, the correlation of response rates in phase II trials to improved survival in phase III in solid tumors is uncertain until proven for any given agent and tumor.55 If the duration of response is brief or if the percentage of response is low, there may be no increase in median survival in a phase III trial. Even within phase III trials, there may be little correlation between response rates and survival benefit.56 Because of patient selection and other biases, response rates should be interpreted cautiously when deciding whether to test an agent in phase III.21,55,57 Because of these considerations, the FDA not only assesses overall response rates but also the number of complete responses; the duration of responses; the reproducibility of response rates in multiple trials; and the history of using the response rate surrogate for the disease and class of drug in question.55 In selected cases, such as drug-refractory patients and rare diseases, bona fide responses in a small number of patients in phase II trials may be sufficient for approval. For example, 13 complete responses provided the major basis for approval of ifosfamide for third-line testis cancer.19,55 The question of whether to advance an agent into phase III testing on the basis of response rates in phase II trials is therefore complex. The answer depends on the disease and the clinical setting in question, whether a response threshold has been eclipsed relative to the standard of care, and, most important, whether the agent has demonstrated a favorable therapeutic ratio (activity to toxicity ratio).
For all cancer treatments under development, activity in the context of toxicity—the therapeutic ratio—must be assessed. In fact, the FDA’s major criterion for efficacy is evidence of patient benefit with a favorable risk-benefit profile.21 To visualize this trade-off, a developmental matrix is offered to demonstrate how the balance between activity and toxicity informs the decisions the test an agent in phase III (Fig 3
By way of orientation, investigational treatments can be placed on the matrix on the basis of their biologic activity in developmental trials on the x-axis and relative toxicity on the y-axis. Those treatments with strong evidence of activity (including response rates and biologic effects on the target) and modest toxicity compared with the standard of care are located toward the lower left and are termed superstars. Agents with less evidence of activity (response rate < 20%) but still with relatively modest toxicity are located toward the bottom of the matrix and are referred to as incrementalists. Those potential treatments with strong evidence of activity but with significant toxicity compared with the standard of care are termed trade-offs and are located in the upper left. Last, the treatments that are unfortunate enough to have both little evidence of biologic activity and significant toxicity are located toward the upper right and are termed castaways, because they are unlikely to enter further clinical testing. The matrix is useful because the designations given to treatments help determine the need for phase III testing. For different reasons, the superstar and castaway treatments typically do not enter phase III trials initially. The superstar classification refers to novel treatments with vastly improved activity (with limited toxicity) relative to the standard of care for the disease under investigation. Because of their impressive characteristics, these treatments may not require phase III testing for initial FDA approval because they become approved (or become the standard of care) on their phase I or phase II clinical results. Examples of superstar treatments among the cytotoxics include paclitaxel for platinum-resistant ovarian cancer (approval was based on five uncontrolled phase II studies) and cladribine for the treatment of hairy-cell leukemia (approval was based on two phase II studies). The paradigmatic and most recent superstar among the MTTs is imatinib mesylate for CML and GIST. Another example of a superstar is the differentiating agent all-trans-retinoic acid in the treatment of acute promyelocytic leukemia (approval based on two phase II studies). Because of their favorable activity and toxicity profiles compared with available treatments, none of these examples required a phase III trial for FDA approval or widespread acceptance. The only problem with superstars in the drug development process is their rarity. In contrast to the superstars, the castaway treatments show little evidence of improved biologic activity for the diseases they are intended to treat or have poor toxicity profiles. An example of a cytotoxic castaway treatment is mitoguazone, which has a unique mechanism of action, as an inhibitor of S-adenosylmethionine decarboxylase, and modest activity against acute myelogenous leukemia and lymphomas.58 Because its activity was not an improvement over the standard treatments at the time of development, and because unacceptable mucositis was associated with its use, development was terminated.30 Although this agent was revived for the treatment of AIDS-related non-Hodgkin’s lymphoma, its relatively poor ratio of activity to toxicity continued to limit developmental progress. Several of the solid tumor vaccines provide examples of castaways that displayed relative modest toxicity but with little if any biologic activity, thus falling to the lower right of the matrix. Drugs that fall in the castaway category, with or without poor toxicity profiles, generally do not warrant later-stage testing. The other two treatment classifications, the incrementalists and trade-offs, typically require phase III testing to evaluate clinical benefit. As much as the superstars are rarified, our review showed that incrementalist treatments represent the typical characterization of a promising novel agent at the completion of phase II testing. These treatments show their promise in phase II by both demonstrating some biologic activity and acceptable toxicity; their activity, however, is neither curative nor dramatic, with response rates less than 20%. Because it is usually impossible to determine the contribution of patient selection bias to the outcomes in phase II, the incrementalists may require phase III testing if FDA approval is to be sought or if clinical benefit is to be confirmed. Examples of cytotoxic incrementalists include single-agent vinorelbine (Navelebine; Glaxo Smithkline, Research Triangle Park, NC) in the treatment of non–small-cell lung cancer and paclitaxel in the adjuvant treatment of node-positive breast cancer. Both of these agents required phase III testing for FDA approval for their respective indication. For the MTTs, an example of an incrementalist is trastuzumab for HER-2/neu-overexpressing breast cancers. For its first-line indication in breast cancer, trastuzumab required testing in a large, phase III trial to show that the combination with chemotherapy (anthracycline plus cyclophosphamide or paclitaxel) had greater efficacy when compared with chemotherapy alone. If no approved treatment exists for a particular indication, an incrementalist, such as Iressa for third-line non–small-cell lung cancer, may achieve approval after phase II, but the outcome is never certain. The trade-off designation is the last of the four. These treatments have evidence of biologic activity but at the price of significant toxicity. An example of such a trade-off treatment is ecteinascidin-743 (ET-743). This agent is a natural product that may work by binding in the minor groove in guanine-cytosine–rich sequences of DNA and alkylating an amino group of guanine.59 The two US phase II trials with ET-743 in soft tissue sarcoma produced response rates of 14% and 8% for chemotherapy-naive and pretreated patients, respectively.60 These responses were accompanied by significant toxicity, including grade 3 to 4 transaminitis and grade 2 to 3 fatigue in 32% of the patients.61 Without the toxicity, these responses may have been enough to establish clear clinical benefit in patients with this drug-resistant disease; in the presence of significant toxicity, they are not, and phase III testing will likely be required in the United States for drug approval. Before proceeding to phase III trials with a trade-off agent, it is prudent to attempt to minimize toxicity by modifying dose or schedule or by using supportive medications.30 In the case of ET-743, phase I trials of weekly 1-hour infusions were initiated and may show better patient tolerance compared with the every-third-week regimen employed in the aforementioned trials.62 If toxicity cannot be circumvented in a novel trade-off agent with promising activity, the agent will usually require phase III testing against another active agent for demonstration of overall clinical benefit.
By placing selective unapproved agents that were in phase III trials as of November 2002 onto the proposed classification matrix (Fig 4
Factor 4: Potential role either alone or in combination in treating a disease. The fourth issue must be addressed before pursuing an expensive registration trial: For what disease and for which line of treatment will marketing approval be sought? This question is critical because it can affect the likelihood and speed of gaining marketing approval as well as the ultimate market size and treatment niche of the agent if approved. In general, a first-line approval strategy is riskier, because the agent is compared with the most active treatment; the potential return, however, is greater because of a larger market size. Depending on the activity of the agent in early efficacy trials, the risk may be justified by the potential for sales after approval. It is generally true that most agents approved for second- and third-line indications are approved as single agents (eg, paclitaxel for second-line ovarian cancer and ifosfamide for third-line testicular cancer). Conversely, most agents with first-line indications are approved for use in combination (eg, irinotecan plus fluorouracil and leucovorin for first-line colon cancer). One important advance over the last several years deserving of discussion is the demonstration of distinct patterns of gene expression in tumors that seem to be histologically identical.63,64 The unrecognized molecular heterogeneity may be associated with differences in prognosis and variability in patient response to chemotherapy. If not controlled for, molecular heterogeneity can result in clinical trials that are underpowered and may fail to provide evidence of efficacy of active therapies.65 If, however, one can identify by genetic or imaging approaches a prognostic profile that distinguishes responders from nonresponders, one might exclude the majority of nonresponders and enrich the response rate to a more acceptable level. This approach is already routinely taken in measuring hormonal status and HER-2 expression in breast cancer and KIT expression in GIST; proponents argue that similar efforts should be undertaken before phase III testing for all new targeted therapies.66 Factor 5: Potential for market niche and strategic role for the sponsor. Regardless of how promising an agent seems with respect to the other four factors, a drug is less likely to advance to phase III testing if the market size is insufficient to support a positive return on investment or if the agent does not play a strategic role in the sponsor’s portfolio. In most cases, the decision to continue development of a drug at any stage depends on whether the drug’s estimated present value of future revenues exceeds the anticipated total development, distribution, and marketing costs. Because a phase III trial represents the single largest clinical development cost, sponsors evaluate the likely demand for the drug before investing millions of dollars on a phase III effort. These demand estimates must take into account off-label usage,67 additional potential indications, changes in reimbursement schedules, and alterations in the standard of care for the disease in question during the several years of phase III development. This demand must then be balanced against the likelihood of achieving marketing approval and a market niche if a phase III registration trial is successful. It is estimated that more than 30% of drugs entering clinical trials are abandoned primarily on the basis of these economic considerations.31 If a compound’s efficacy seems promising, but its predicted return on investment seems unfavorable, it can be offered to the NCI or the cooperative groups to develop under their sponsorship and at their expense.
For traditional cytotoxic drugs, it has generally been difficult to estimate market size, and in many cases, such as paclitaxel, initial market calculations significantly underestimated the ultimate expansion into other indications. Significant postmarketing approval investment in clinical trials that explore new combinations of drugs and new disease settings usually leads to expansion of sales. Even imatinib mesylate (approved for a small market) will likely approach $1 billion in annual sales on the basis of its continuous use in patients who will live for an extended period of time (Table 5
This study provides a systematic overview of characteristics of contemporary phase III trials and formalizes major criteria that sponsors use to make the go or no go decision. The findings demonstrate that despite their utility, phase III trials are the most time-consuming, scrutinized, and expensive aspect of the drug development process.31,68 Because of the resources they each consume, these trials are risky from a sponsor’s perspective. A positive result will likely lead to marketing approval and make the effort worthwhile; a negative result can easily kill a drug or at least produce a major setback for the drug and its sponsor. The recent failure of high-profile drugs such as Iressa in phase III trials highlights the risk of subjecting modestly active drugs to large, randomized trials and the need to examine clinical research strategies. At the same time, the agents themselves present greater challenges for development, given that many are noncytotoxic and therefore more difficult to assess than traditional cytotoxics. The transition from phase II to phase III requires a strategic decision based on several new considerations. No longer are clinical response rates in phase II the only driving factor in supporting the decision to embark on phase III trials. Uniqueness of target and its preclinical validation, evidence of biologic effect on the target, surrogates for response, pharmacokinetic characteristics, and potential for enhancement of cytotoxic therapy all play an increasingly important role in supporting the decision to proceed to phase III. Each of these new considerations relies on preclinical data and early clinical information not ordinarily of critical importance in classic cytotoxic drug development, in which single-agent response rates in phase I and II have been paramount. Thus, paradoxically, a greater investment in the phase I and II drug trials may be required to provide the information necessary for phase III planning and commitment. New approaches, incorporating molecular imaging, genomics, and pharmacogenetics, as well as invasive tissue sampling, will be required to provide this background information. To this point, despite the visible transition from cytotoxics to molecular targeted drug discovery, the ability to assess biologic end points in the clinic has not kept pace. This analysis revealed that many of the criteria that sponsors use to make go or no go decisions are beyond the control of clinical researchers (eg, resources available to the sponsor, FDA policies, market size, and competition). One area, however, that demands greater participation from clinical researchers is the identification of enriched patient populations for phase II and III testing. Although the vast majority of phase III trials that we examined were testing agents with identifiable molecular targets, we found that few sponsors identified subgroup populations likely to respond to these agents by the end of phase II. The failure to identify subgroups of responders contributed to the categorization of most therapies at the completion of phase II as incrementalists. Because the accelerated approval mechanism allows such agents to achieve approval after phase II in many instances, as illustrated by the recent approval of Iressa, sponsors may have less incentive to identify subgroups of responders because of the potential effect this approach can have on reducing market size. It therefore becomes especially important for the clinical research community to encourage industry and government sponsors to support appropriate biologic correlative studies in phase I and II to design phase III studies that contain target-enriched patient populations. This approach may ultimately lead to more superstars after phase II and the need for smaller, faster, and less expensive confirmatory phase III trials. If we ignore this approach, we may be doomed to make expensive mistakes and to end a rational preclinical undertaking with an ill-conceived phase III trial.
The following authors or their immediate family members have indicated a financial interest as follows: No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Owns stock (not including shares held through a public mutual fund): Bruce A. Chabner, Pharmamar, Kosan, Cell Genesis. Acted as a consultant within the past 2 years: Bruce A. Chabner, Baxter, Eli Lilly, Cephalon, Pharmamar, Cell Genesis; Thomas J. Lynch Jr, Aventis. Served as an officer or member of the board of a company: Bruce A. Chabner, Kosan Bioscience. Received more than $2,000 a year from a company for either of the last 2 years: Thomas J. Lynch Jr, AstraZeneca, Lilly, Aventis.
We thank Peg Hewitt of the Tufts Center for the Study of Drug Development for providing librarian support and an anonymous referee for providing helpful comments.
T.G.R. is supported in part by National Cancer Institute grant R25T CA 92203.
1. Frei E III, Holland JF, Schneiderman MA, et al: A comparative study of two regimens of combination chemotherapy in acute leukemia. Blood 13:1126–1147, 1958 2. Sylvester R, Van Glabbeke M, Collette L, et al: Statistical methodology of phase III cancer clinical trials: Advances and future perspectives. Eur J Cancer 38:S162–S168, 2002 (suppl 4) 3. Simon R: Randomized clinical trials in oncology: Principles and obstacles. Cancer 74:2614–2619, 1994[Medline] 4. Byar DP, Simon RM, Friedewald WT, et al: Randomized clinical trials: Perspectives on some recent ideas. N Engl J Med 295:74–80, 1976[Abstract] 5. Nottage M, Siu LL: Principles of clinical trial design. J Clin Oncol 20:42S–46S, 2002 (suppl)
6. Schein PS: The case for a new national program for the development of cancer therapeutics. J Clin Oncol 19:3142–3153, 2001
7. Schellens JH, Ratain MJ: Endostatin: Are the 2 years up yet? J Clin Oncol 20:3758–3760, 2002 8. Stadler WM, Ratain MJ: Development of target-based antineoplastic agents. Invest New Drugs 18:7–16, 2000[CrossRef][Medline]
9. Gelmon KA, Eisenhauer EA, Harris AL, et al: Anticancer agents targeting signaling molecules and cancer cell environment: Challenges for drug development? J Natl Cancer Inst 91:1281–1287, 1999
10. Ratain MJ, Stadler WM: Clinical trial designs for cytostatic agents. J Clin Oncol 19:3154–3155, 2001
11. Korn EL, Arbuck SG, Pluda JM, et al: Clinical trial designs for cytostatic agents: Are new approaches needed? J Clin Oncol 19:265–272, 2001
12. Bailar JC III, Gornik HL: Cancer undefeated. N Engl J Med 336:1569–1574, 1997 13. National Cancer Institute Physicians Query Database. http://www.cancer.gov/cancerinfo/pdq/cancerdatabase. Accessed May 13, 2003 14. ClinicalTrials.gov: A Service of the National Institutes of Health, Developed by the National Library of Medicine. http://www.clinicaltrials.gov/. Accessed May 13, 2003 15. Mathien MP: Parexel’s Pharmaceutical R&D Statistical Sourcebook 2002/2003. Boston, MA, Parexel International Corporation, 2002 16. US 21 Code of Federal Regulations, Parts 314.500–314.560 17. US Food and Drug Administration, Guidance for Industry: FDA Approval of New Cancer Treatment Uses for Marketed Drugs and Biological Products (March 13, 1997). Rockville, MD, FDA guidance doc 18. Food and Drug Administration Modernization Act of 1997, Pub. L. No. 105–115, 111 Stat. 2296 (Nov. 21, 1997) 19. US Food and Drug Administration Center for Drug Evaluation and Research Oncology Tools web site.http://www.fda.gov/cder/cancer/. Accessed March 13, 2003 20. New medicines in development for cancer. Washington, DC, Pharmaceuticals Research and Manufacturers Association of America, 2001, pp 1–56 21. Hirschfeld S, Pazdur R: Oncology drug development: United States Food and Drug Administration perspective. Crit Rev Oncol Hematol 42:137–143, 2002[Medline] 22. Shulman SR, Wood-Armany MJ: Accelerating access to cancer drugs. J Biolaw Bus 2:38–44, 1999
23. Schilsky RL: End points in cancer clinical trials and the drug approval process. Clin Cancer Res 8:935–938, 2002
24. Johnson JR, Williams G, Pazdur R: End points and United States Food and Drug Administration approval of oncology drugs. J Clin Oncol 21:1404–1411, 2003 25. Interim analysis of phase III data preferable for accelerated approval, FDA’s Pazdur says. Cancer Lett 28:1–4, 2002 26. DiMasi JA, Hansen RW, Grabowski HG: The price of innovation: New estimates of drug development costs. J Health Econ 22:151–185, 2003[CrossRef][Medline] 27. DiMasi JA, Hansen RW, Grabowski HG, et al: Cost of innovation in the pharmaceutical industry. J Health Econ 10:107–142, 1991[CrossRef][Medline]
28. Roche K, Paul N, Smuck B, et al: Factors affecting workload of cancer clinical trials: Results of a multicenter study of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 20:545–556, 2002
29. Carter SK: Clinical strategy for the development of angiogenesis inhibitors. Oncologist 5:51–54, 2000 (suppl 1) 30. Von Hoff DD: There are no bad anticancer agents, only bad clinical trial designs: Twenty-first Richard and Hinda Rosenthal Foundation Award Lecture. Clin Cancer Res 4:1079–1086, 1998[Abstract] 31. Dimasi JA: Risks in new drug development: Approval success rates for investigational drugs. Clin Pharmacol Ther 69:297–307, 2001[CrossRef][Medline] 32. Tufts Center for the Study of Drug Development quantifies savings from boosting new drug R&D efficiency. Boston, MA, Tufts Center for the Study of Drug Development, September 9, 2002. http://csdd.tufts.edu/NewsEvents/RecentNews.asp?newsid=20. Accessed November 18, 2002 33. DiMasi JA: The value of improving the productivity of the drug development process: Faster times and better decisions. Pharmacoeconomics 20:1–10, 2002 (suppl 3)[Medline]
34. Wittes RE, Friedman MA: Accrual to clinical trials. J Natl Cancer Inst 80:884–885, 1988 35. Winn RJ: Obstacles to the accrual of patients to clinical trials in the community setting. Semin Oncol 21:112–117, 1994[Medline] 36. Wright JR, Crooks D, Ellis PM, et al: Factors that influence the recruitment of patients to phase III studies in oncology: The perspective of the clinical research associate. Cancer 95:1584–1591, 2002[CrossRef][Medline]
37. Viability of cancer clinical research: Patient accrual, coverage, and reimbursement—American Medical Association Council on Scientific Affairs. J Natl Cancer Inst 83:254–259, 1991 38. Simon R, Thall PF, Ellenberg SS: New designs for the selection of treatments to be tested in randomized clinical trials. Stat Med 13:417–429, 1994[Medline]
39. Chen TT, Chute JP, Feigal E, et al: A model to select chemotherapy regimens for phase III trials for extensive-stage small-cell lung cancer. J Natl Cancer Inst 92:1601–1607, 2000
40. Freidlin B, Breathnach OS, Johnson BE: A model to select regimens for phase III trials for patients with advanced-stage non-small cell lung cancer. Clin Cancer Res 9:917–922, 2003 41. Fazzari M, Heller G, Scher HI: The phase II/III transition: Toward the proof of efficacy in cancer clinical trials. Control Clin Trials 21:360–368, 2000[CrossRef][Medline] 42. Yao Q, Wei LJ: Play the winner for phase II/III clinical trials. Stat Med 15:2413–2423, 2455–2458, 1996[CrossRef][Medline]
43. Scher HI, Heller G: Picking the winners in a sea of plenty. Clin Cancer Res 8:400–404, 2002 44. De Gruttola VG, Clax P, DeMets DL, et al: Considerations in the evaluation of surrogate endpoints in clinical trials: Summary of a National Institutes of Health workshop. Control Clin Trials 22:485–502, 2001[CrossRef][Medline]
45. Rosner GL, Stadler W, Ratain MJ: Randomized discontinuation design: Application to cytostatic antineoplastic agents. J Clin Oncol 20:4478–4484, 2002
46. Carney DN: Lung cancer: Time to move on from chemotherapy. N Engl J Med 346:126–128, 2002 47. Hanahan D, Weinberg RA: The hallmarks of cancer. Cell 100:57–70, 2000[CrossRef][Medline]
48. Chabner BA: The oncologic four-minute mile. Oncologist 6:230–232, 2001 49. Reddy A, Kaelin WG: Using cancer genetics to guide the selection of anticancer drug targets. Curr Opin Pharmacol 2:366–373, 2002[CrossRef][Medline]
50. Chabner BA, Boral AL, Multani P: Translational research: Walking the bridge between idea and cure—Seventeenth Bruce F. Cain Memorial Award lecture. Cancer Res 58:4211–4216, 1998
51. Coussens LM, Fingleton B, Matrisian LM: Matrix metalloproteinase inhibitors and cancer: Trials and tribulations. Science 295:2387–2392, 2002
52. Collins JM: Pharmacology and drug development. J Natl Cancer Inst 80:790–792, 1988 53. Johnson J, Monks A, Hollingshead M, et al: Preclinical aspects of cancer drug discovery and development, in Chabner BA, Longo D (eds): Cancer Chemotherapy and Biotherapy. Philadelphia, PA, Lippincott Williams & Wilkins, 2001, pp 17–36 54. Newell DR: Pharmacologically based phase I trials in cancer chemotherapy. Hematol Oncol Clin North Am 8:257–275, 1994[Medline]
55. Pazdur R: Response rates, survival, and chemotherapy trials. J Natl Cancer Inst 92:1552–1553, 2000 56. Buyse M, Thirion P, Carlson RW, et al: Relation between tumour response to first-line chemotherapy and survival in advanced colorectal cancer: A meta-analysis—Meta-Analysis Group in Cancer. Lancet 356:373–378, 2000[CrossRef][Medline]
57. Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors: European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 92:205–216, 2000 58. Warrell RP Jr, Burchenal JH: Methylglyoxal-bis(guanylhydrazone) (methyl-GAG): Current status and future prospects. J Clin Oncol 1:52–65, 1983[Abstract] 59. Pommier Y, Kohlhagen G, Bailly C, et al: DNA sequence- and structure-selective alkylation of guanine N2 in the DNA minor groove by ecteinascidin 743, a potent antitumor compound from the Caribbean tunicate Ecteinascidia turbinata. Biochemistry 35:13303–13309, 1996[CrossRef][Medline] 60. Demetri GD: ET-743: The US experience in sarcomas of soft tissues. Anticancer Drugs 13:S7–S9, 2002 (suppl 1) 61. Demetri GD, Manola J, Harmon D, et al: Ecteinascidin-743 (ET-743) Induces durable responses and promising 1-year survival rates in soft tissue sarcomas (STS): Final results of phase II and pharmakoknetic studies in the U.S.A. Proc Am Soc Clin Oncol 20:352a, 2001 (abstr 1406)
62. Chabner BA: Cytotoxic agents in the era of molecular targets and genomics. Oncologist 7:34–41, 2002 (suppl 2) 63. Perou CM, Sorlie T, Eisen MB, et al: Molecular portraits of human breast tumours. Nature 406:747–752, 2000[CrossRef][Medline] 64. Alizadeh AA, Eisen MB, Davis RE, et al: Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511, 2000[CrossRef][Medline]
65. Betensky RA, Louis DN, Cairncross JG: Influence of unrecognized molecular heterogeneity on randomized clinical trials. J Clin Oncol 20:2495–2499, 2002
66. Castro M: The simpleton’s error in drug development. J Clin Oncol 20:4606–4607, 2002
67. Laetz T, Silberman G: Reimbursement policies constrain the practice of oncology. JAMA 266:2996–2999, 1991 68. Lasagna L, Frei E III: The impact of regulations, tradition, and experimental design on clinical cancer trials: Report and recommendations resulting from Washington Cancer Trials Conference. Am J Clin Oncol 19:325–329, 1996[CrossRef][Medline] Submitted January 27, 2003; accepted July 8, 2003.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
