|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology
The Multifunctional, Multi-Institutional, and Sometimes Even Global Phase I Study: A Better Life for Phase I Evaluations or Just "Living Large"?Institute for Drug Development, Cancer Therapy and Research Center, University of Texas Health Science Center at San Antonio, San Antonio, TX IN THIS ISSUE OF the Journal of Clinical Oncology, Baselga et al1 report on the results of a phase I, pharmacokinetic, and pharmacodynamic study of the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor ZD1839, which was one of the first selective inhibitors of the EGFR tyrosine kinase to enter clinical evaluation. The principal objectives of the study were to evaluate the feasibility, safety, pharmacokinetics, pharmacodynamics, and preliminary antitumor activity of ZD1839 administered on a continuous once-daily dosing schedule. The study represents an important contribution to the arena of developmental therapeutics targeting EGFR. The authors should be congratulated for providing a body of seminal information that has served as a foundation for broad disease-directed clinical evaluations including two randomized, multinational phase III trials, in which previously untreated patients with advanced non–small-cell lung carcinoma were randomized to treatment with either a conventional chemotherapy regimen alone or ZD1839. However, several features of the phase I study design reported here, including the participation of unusually large numbers of actively accruing institutions and patients at each dose level, as well as more than just a "preliminary" slant toward efficacy end points, are incongruent to the stated phase I–related aims of the study and are concerning from a safety standpoint. Nevertheless, this study is not the first to introduce these multi-institutional and multifunctional design elements and will likely not be the last.2 Instead, it is representative of a growing trend in which large, multifunctional phase I evaluations, championed by the pharmaceutical and biotechnology industries, are progressively being performed in lieu of smaller, more intimate trials that have been traditionally conducted at single institutions and strictly focused on dose finding and characterizing the toxicological and pharmacological profiles of new anticancer therapeutics. This trend undoubtedly stems from mounting competitive pressures in these industries, resulting in a quest for maximal efficiency in patient resource utilization and a strict adherence to often-unrealistic management-driven time lines. These competitive corporate pressures have been progressively shifted over the years to the clinic and are now being observed downstream at the earliest phase of therapeutic evaluations, which were once largely immune from most study design imperfections that would even slightly increase the risk for patients, inasmuch as the overriding theme in the design and conduct of the phase I studies has always been related to minimizing risk with the principal dictum being to cut no corners. The dose escalation scheme and the multi-institutional aspects of the present study of ZD1839 run counter to the safeguards implemented and accepted in new phase I study methodologies such as accelerated dose titration designs and continuous reassessment methods, which represent investigator efforts to maximize efficiency and solve ethical dilemmas in this critical stage of therapeutic development.3,4 These approaches were principally developed to reduce the number of patients treated at potentially subtherapeutic dose levels and, more importantly, to reduce the numbers of patients treated at doses that exceed the maximum-tolerated dose. In retrospect, the design elements of the present study are congruent with the first aim, as recent results of clinical trials indicate that ZD1839 doses of 250 and 500 mg/d may be equally efficacious, and the starting dose of 150 mg/d was therefore unlikely to have been subtherapeutic. However, the study design was not conducive to maximal risk reduction as it potentially exposed more patients than necessary to doses exceeding the maximum-tolerated dose. For example, there was an unacceptably high incidence of intolerable toxicity by any standards in patients treated at the 1,000-mg/d dose level of ZD1839 in the present study; six (50%) of 12 patients experienced severe (grade 3) nonhematological dose-limiting toxicity, including five patients in the first course of treatment rather than the maximally accepted norm of two patients. These events highlight one of the major vulnerabilities inherent in the multi-institutional approach to phase I studies: Rapid accrual may actually obscure the recognition of toxicities and hinder appropriate decisions to impede patient accrual to reduce risk to other patients. In multi-institutional phase I studies, investigators from different sites are, in effect, competing with one another for treatment slots. For a multitude of reasons, including no shortage of eligible patients with advanced malignancies and access to novel and potentially beneficial therapeutics, investigators are highly motivated to quickly accrue their patients. However, the reported absence of toxicity in a single patient after a 28-day observation period provides little statistical confidence to ensure the safety of 14 additional patients. Furthermore, the simultaneous entry of multiple patients provides little opportunity to apply the "brakes" that should be built into the study to prevent or minimize toxicity in other patients once severe toxicity is recognized. The present study illustrates many of the inordinate complexities of study design and logistics that are being progressively incorporated into phase I evaluations in this brave world of "buzz" words that include "global," "multitasking," and "supersize-it." Once the first patient treated at each ZD1839 dose level had safely cleared the 28-day observation period, 14 patients from as many 10 institutions, seven nations, two continents, and two hemispheres, spanning countless time zones and even one international dateline, could be treated simultaneously at the newly declared "safe" dose level and a single subject was able to be concomitantly treated at the next higher dose. Certainly, the resources allocated to coordinate the dynamic logistical aspects innate to the design of this phase I study, particularly the real-time monitoring and assessment of toxicities, not to mention the need to evaluate adverse events and communicate assessments to all sites in multiple languages, must have been an enormous undertaking. Unfortunately, the potential safety risks inherent in such efforts, as well as the loss of other intangibles incurred by the dilution of observations among investigators and institutions, may have also been enormous, but for what possible gain? Complex logistical and multifunctional design elements have been incorporated into early evaluations in other therapeutics areas for many years. However, phase I studies of anticancer agents, which have much narrower therapeutic margins and higher risk-benefit ratios, have traditionally been performed by a small number of experienced investigators at a maximum of one or two highly specialized sites. Similar to good laboratory practices, which mandate the use of the fewest well-calibrated instruments as possible to minimize experimental variability due to instrumentation, this seemingly minor, albeit powerful, logistical feature has traditionally facilitated the acquisition of expertise by both investigators and research staff because it enables them to make detailed observations over the entire range of dose levels. The resultant intimacy enhances the ability of investigators to detect subtle, albeit potentially consequential, adverse effects, and to readily compare toxicities from dose level to dose level, patient to patient, and schedule to schedule. In addition, the geographic concentration of adverse events has undoubtedly accelerated the derivation of measures to minimize adverse effects, which could have otherwise led the development of many important therapeutics astray for many years or possibly forever. Recent examples include premedication to minimize severe hypersensitivity reactions due to paclitaxel, third-space fluid accumulation related to docetaxel, and even the recognition that treatment with therapeutics targeting EGFR could be safely continued despite the fulminant development of dermatological toxicity.5-8 In essence, this rather low-tech approach to phase I evaluations has resulted in the accurate, safe, expedient, and successful characterization of the toxicological and pharmacological profiles of a multitude of anticancer agents over the last several decades. Common sense would argue that the participation of many investigators in a vast array of institutions, countries, and continents may impair the ability of these clinicians to recognize diverse adverse effects of therapeutics. Although the hackneyed expression "nothing beats experience" is heard as part of almost everyone’s medical training, the educational literature suggests that the dilution of experience may hinder clinical diagnosis.9-13 Toxicity and safety data for phase I studies are principally derived through the process of focused medical interrogation. The investigator must, in effect, identify and diagnose drug-induced adverse effects among myriad signs and symptoms due to the underlying malignancy. Pivotal to the interpretation of these events and the attribution of relatedness (ie, separating "signal from noise") is pattern recognition.10-13 Isolated symptoms may be recognized as being related to treatment only when experience indicates that a pattern has emerged, and then awareness is raised for the same toxicity or pattern in subsequent patients. In large multi-institutional, not to mention multinational phase I studies, individual investigators may gain no more than just anecdotal experience with a compound. If anticancer agents were associated with high therapeutic indices or broad safety margins, this issue would be no more than a theoretical concern, perhaps to be pondered by oncological philosophers. However, our cumulative experience with both nonspecific cytotoxic agents and more recently with rationally designed, targeted therapeutics indicates that therapeutic indices and safety margins remain low and narrow, and the economic and health impact of unrecognized or underappreciated toxicities in cancer patients can be enormous.14-16 It is ironic that large multicenter studies, particularly those evaluating phase I end points such as the present study, limit the ascertainment of the highest level of expertise attainable by any single individual or group of investigators, especially when such expertise is the principal criterion used to select investigators, along with their research staff and institutions, to participate in the study in the first place. In response to economically driven competitive pressures to maximize efficiency in early evaluations of anticancer therapeutics, phase I trials are now incorporating inordinately unrealistic goals and time lines and are being designed to kill as many birds as possible with a single stone. Phase I evaluations are now seeking to simultaneously address both toxicity and efficacy questions and concurrently satisfy the requirements of multiple regulatory agencies. The multi-institutional phase I study represents a somewhat misguided attempt to simultaneously accomplish multiple objectives and beat the clock set by corporate management based on the notion that a large number of institutions will likely increase the rate of patient accrual. However, the rate of patient accrual in phase I trials is tightly regulated by fixed observation periods, which are implemented a priori as safeguards to minimize the numbers of patients treated with potentially intolerable doses. Furthermore, multiple dose levels, including those that may not even be biologically relevant, are often sized a priori above and beyond that required to detect a sufficiently low incidence of intolerable effects. Although this is usually performed under the guise of better characterizing the pharmacokinetics and pharmacodynamics of new agents, the main reason is to simultaneously address efficacy questions. Nevertheless, delayed tumor growth, which is the predominant antitumor effect of rationally designed, target-based therapeutics, is not likely to be appreciated in nonrandomized clinical trials, and a dose cohort size of 14 patients is insufficient to even roughly assess the relationship between dose and antitumor activity. And then there is the progressive dilution of the role and responsibilities of the principal investigator in large multi-institutional studies, particularly those with phase I objectives. Even when an industry-sponsored multi-institutional phase I study bears the name of a single principal investigator on the front-sheet of the protocol, which still does occasionally occur from time to time, the main concern is whether any single investigator can actually provide careful, real-time oversight of all toxicities and safety issues, particularly those that span many countries, continents, and time zones. In most cases, this role must be functionally relinquished to the corporate sponsor and its clinical research team, who have access to the data at all participating institutions and therefore can compile the results for assessment. In essence, the individual or entity who obtains access to the emerging toxicity data ascends to the role of principal investigator, making judgments about safety issues and dose escalation without either the benefit of direct clinical experience with the agent under testing or financial independence from the sponsor. Irrespective of increasing public and regulatory scrutiny of the processes by which novel therapeutics are developed, the functional delegation of this responsibility to a third party cannot be permitted in phase I studies in which safety issues dominate, particularly if the third party is potentially conflicted. Instead, the study must be designed so that its logistical elements are conducive for a principal investigator to truly serve in this capacity. In multi-institutional phase I studies, such as that presented by Baselga et al, the use of an external data safety monitoring committee should be considered to provide additional real-time checks, balances, and oversight. "Living large" with multifunctional, multipurpose, and multi-institutional phase I studies will not accelerate the development of rationally designed, target-based anticancer therapeutics, and it runs counter to the well-entrenched dictum of doing no harm by designing phase I clinical trials to ensure maximal patient safety and oversight. Furthermore, as in all high-quality research, early clinical investigations must be designed to address only a limited number of specific questions. Too often in the competitive arena of drug development, attempts are made to address multiple unrelated, and often conflicting, issues in a single phase I study. Although discerning antitumor benefit, developing and validating biologic assays, and assessing the effects of therapeutics on the tumor target or surrogate tissues are certainly laudable goals in the development of targeted therapeutics, they must not adversely impact on the principal objectives of the study, which are directed at safety and dose-delineation with an eye toward directly applying this knowledge in subsequent disease-directed evaluations and translational studies. Efforts to design phase I studies to include only patients with limited specific histologic subtypes to preliminarily assess antitumor activity, or to restrict eligibility only to patients amenable to serial tumor sampling, particularly in the absence of validated biologic assays, do not only blur the distinctions between phase I and phase II/III clinical trials, but also risk missing bona fide antitumor activity in other unexpected indications or may obscure the appreciation of a relevant magnitude of activity in selected indications due to low statistical power. Nevertheless, although traditional phase I study designs have proven serviceable over the last several decades, they should not be adhered to unquestionably, and improvements are warranted, particularly in the case of those involving rationally designed, target-based therapeutics that may portend superior therapeutic indices if optimal biologic doses, rather than maximum-tolerated doses, are delineated and used. However, the essential characteristics of past phase I study designs should not be sacrificed in the design of new phase I study methodologies. REFERENCES
1. Baselga J, Rischin D, Ranson M, et al: Phase I safety, pharmacokinetic, and pharmacodynamic trial of ZD1839, a selective oral epidermal growth factor receptor tyrosine kinase inhibitor, in patients with five selected solid tumor types. J Clin Oncol 20: 4292-4302, 2002
2. Ranson M, Hammond LA, Ferry D, et al: ZD1839, a selective oral epidermal growth factor receptor-tyrosine kinase inhibitor, is well tolerated and active in patients with solid, malignant tumors: Results of a phase I trial. J Clin Oncol 20: 2240-2250, 2002
3. Simon R, Freidlin B, Rubinstein L, et al: Accelerated titration designs for phase I clinical trials in oncology. J Natl Cancer Inst 89: 1138-1147, 1997 4. O’Quigley J, Pepe M, Fisher L: Continual reassessment method: A practical design for phase 1 clinical trials in cancer. Biometrics 46: 33-48, 1990[CrossRef][Medline] 5. Weiss RB, Donehower RC, Wiernik PH, et al: Hypersensitivity reactions from Taxol. J Clin Oncol 8: 1263-1268, 1990[Abstract]
6. Peereboom DM, Donehower RC, Eisenhauer EA, et al: Successful re-treatment with Taxol after major hypersensitivity reactions. J Clin Oncol 11: 885-890, 1993
7. Hidalgo M, Siu LL, Nemunaitis J, et al: Phase I and pharmacologic study of OSI-774, an epidermal growth factor receptor tyrosine kinase inhibitor, in patients with advanced solid malignancies. J Clin Oncol 19: 3267-3279, 2001 8. Cortes JE, Pazdur R: Docetaxel. J Clin Oncol 13: 2643-2655, 1995[Abstract] 9. Grant J, Marsden P: Primary knowledge, medical education and consultant expertise. Med Educ 22: 173-179, 1988[Medline] 10. Wood JK: Diagnosis by pattern recognition. Lancet 1: 801, 1971[Medline] 11. Ridderikhoff J: Medical problem-solving: An exploration of strategies. Med Educ 25: 196-207, 1991[Medline] 12. Dudley HA: Pay-off, heuristics, and pattern recognition in the diagnostic process. Lancet 2: 723-726, 1968[CrossRef][Medline] 13. Dunn MM, Woolliscroft JO: Assessment of a pattern-recognition examination in a clinical clerkship. Acad Med 69: 683-684, 1994[Medline]
14. Rothenberg ML, Meropol NJ, Poplin EA, et al: Mortality associated with irinotecan plus bolus fluorouracil/leucovorin: Summary findings of an independent panel. J Clin Oncol 19: 3801-3807, 2001
15. Friedrich MJ: Cardiotoxicity concerns prompt data review in breast cancer trial. J Natl Cancer Inst 94: 650-651, 2002
16. Cohen MH, Williams G, Johnson JR, et al: Approval summary for imatinib mesylate capsules in the treatment of chronic myelogenous leukemia. Clin Cancer Res 8: 935-942, 2002 This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
