|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 12 (April 20), 2008: pp. 1926-1931 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.3793 Multi-Institutional Phase I Trials of Anticancer Agents
From the Division of Hematology/Oncology and the Department of Epidemiology and Biostatistics, Case Western Reserve University; University Hospitals Ireland Cancer Center; and the Case Comprehensive Cancer Center, Cleveland, OH Corresponding author: Afshin Dowlati, MD, Division of Hematology/Oncology, University Hospitals Case Medical Center, 11100 Euclid Ave, Cleveland, OH 44106; e-mail: afshin.dowlati{at}case.edu
Purpose Physicians involved in the conduct of phase I studies of novel anticancer agents have raised concerns about the emergence of multi-institutional phase I trials and about using the optimal biologic dose (OBD) as an alternative to the maximum-tolerated dose (MTD) as the primary end point in early drug development. We sought to determine the factors associated with multi-institutional phase I studies and OBD determination. Patients and Methods We reviewed all published phase I trials between January 1998 and June 2006 from two major clinical cancer journals. The following components from each trial were determined: number of participating sites, sponsor, nation where study was conducted, MTD or OBD established, number of patients accrued, mechanism of action of the studied agent, accrual time, and tumor type. Results We identified 463 trials. Fifty-six percent were performed in single institutions. Only 30% reported accrual time. The number of patients enrolled on single institution studies was significantly lower than on multi-institutional studies (P < .05), but there was no difference in accrual time. There was no association between the number of institutions and the sponsor or the mechanism of drug action. National Institutes of Health–sponsored trials enrolled fewer patients per trial than pharmaceutical-sponsored trials (P < .05). Although 99% of trials with cytotoxic agents determined an MTD, only 64% of trials with targeted agents did. Conclusion Multi-institutional phase I studies do not decrease the time to study completion and result in an increase in number of patients per trial. One third of trials with targeted agents failed to determine an MTD.
The explosion of our knowledge about the biology of cancer has resulted in the appearance of new targets for anticancer drug development. In addition, high throughput evaluation of small molecules has further increased the number of agents that require clinical development. First-time human administration of candidate anticancer drugs or of novel combinations is performed in the context of phase I trials. Traditional phase I trials involve a dose escalation in cohorts of three to six patients with the primary end points of dose-limiting toxicity (DLT) determination; maximum-tolerated dose (MTD; ie, the dose less than the DLT at which one or fewer of the three to six patients in the cohort develops a DLT); and, finally, a recommended phase II dose.1 Newer designs include pharmacologically guided dose escalation, statistically based methods of dose escalation, and accelerated titration designs. However, none of these newer trial methodologies deviates from the primary end points, and, in all cases, a minimum number of patients are enrolled.1 Given the regulatory and logistical complexities, the conduct of phase I trials has been limited to institutions with experience in this field. Many physicians involved with the design and conduct of these studies have noticed a change in the overall landscape of phase I drug development. Among the changes include a deviation from the end points of determining the DLT and MTD. Many studies now fail to demonstrate a DLT or an MTD and are focusing on determining the optimal biologic dose (OBD), which can be elusive.2,3 The OBD lacks a standard definition, although many consider it a recommended phase II dose that is determined by biologic end points rather than by toxicity. In an attempt to find an OBD, investigators perform intensive pharmacokinetic and pharmacodynamic sampling, which has added a further layer of complexity to early-phase drug development. In some cases, sequential tumor biopsies have been performed for these pharmacodynamic evaluations.4 Another perceived change includes the emergence of so-called multi-institutional phase I trials. In this context, multiple institutions participate in a single phase I trial, and enrollment is generally done by allocating patient slots to each participating site. In the absence of enrollment completion in the given time frame, the slot is open to all sites on a first-come, first-served basis. The responsibility for allocating patient slots generally lies with the sponsor and requires significant coordination among the participating sites. Multi-institutional phase I trials, however, have come under criticism recently.5 In an attempt to better understand this changing landscape in phase I drug development, we conducted a review of the available phase I studies published during an 8-year period.
Data Acquisition We reviewed all published phase I trials between January 1998 and June 2006 from two clinical cancer journals—Journal of Clinical Oncology and Clinical Cancer Research—that routinely publish phase I studies. Trials that had a combined phase I/II design were excluded from the analysis. After a thorough review of each manuscript, we identified the following components: number of participating sites, sponsor (National Institutes of Health [NIH] or NIH foreign equivalent v pharmaceutical v institutional), nation(s) where study was conducted (North America v Europe/Australia), establishment of an MTD or an OBD, number of patients accrued, mechanism of action of the studied agent (eight different categories; Table 1), accrual time, and tumor type (general solid tumor v hematologic malignancies v disease-specific solid tumor) in which the trial was conducted. Agents with a targeted mechanism of drug action were defined as agents that have determined molecular targets that do not fall into the traditional class of chemotherapeutic agents (eg, alkylating agents, antitumor antibiotics, platinum agents, toposiomerase inhibitors, antimetabolites).
Statistical Evaluation The association between two factors was examined by using the  2 or Mantel-Haenszel 2 test (between two ordinal variables). The difference between accrual and number of patients among groups was examined by the Kruskal-Wallis test, a nonparametric test that is based on Wilcoxon scores followed by the Benferroni multiple-comparison procedure (Wilcoxon rank sum test if there only two groups). All statistical analyses were done with SAS software (SAS Institute, Cary, NC), and a P value less than .05 was considered statistically significant.
General We identified 463 phase I trials. The majority of trials evaluated patients with solid tumors; only 23 (5%) were performed on patients with hematologic malignancies (leukemia, lymphoma, and myeloma). The number of phase I studies that were conducted specifically in hematologic malignancies increased with time from 4.3% in 1998 to 19.2% in 2005 (P < .01). This increase was caused by an increase in the number of targeted agents evaluated. The majority of studies were done in North America (70.1%), followed by Europe/Australia (23.7%), then jointly between North America and Europe (3.2%). Only 2.1% of studies were conducted outside of North America and Europe. Pharmaceutical companies were the major sponsors (58.8%), followed by the NIH or foreign equivalent (35.2%), and the principal investigators’ institutions (4.3%). Cytotoxic chemotherapies were the most investigated agents (43.6%) of all studies, followed by targeted agents (30.6%). Of the 463 published studies, 138 (30%) reported the accrual time.
Multi-Institutional Phase I Studies
Pharmaceutical Versus NIH Sponsorship There was marginally significant difference in the number of patients per study among sponsors (P = .07; Fig 2). NIH-sponsored studies enrolled a median of 28 patients (range, 12 to 76), versus pharmaceutical-sponsored trials that enrolled a median of 31 patients (range, 9 to 99), which was significant by pairwise comparison (P < .05).
Disease-Specific Phase I Trials (Single-Disease Solid Tumors) Of the 463 trials evaluated, 63 (13%) only recruited patients with a single, solid-tumor, disease type (eg, breast, lung, renal). These disease-specific, solid-tumor, phase I trials were highly associated with targeted agents (Table 3). In addition, disease-specific phase I studies enrolled considerably fewer patients per trial than the studies in either general solid tumors or hematologic malignancies (Table 4; P < .01). Accrual time, however, was not different among disease-specific trials or trials that involved general solid tumors but was longer in hematologic malignancy–specific phase I trials (Table 5; P = .04). Disease-specific phase I studies also are more likely to be multi-institutional in nature compared with studies in general solid tumors. Seventeen percent of disease-specific studies had at least four institutions involved, as opposed to only 6% of studies in general solid tumors.
OBD Versus MTD The number of studies using the OBD end point as a percentage of all phase I studies did not increase during the study period (P = .77). OBD studies were determined mainly for targeted agents. Targeted agents defined an MTD in 64% and an OBD in 36% of studies compared with cytotoxic drugs that defined an MTD in 99% of studies. We did not find any association between the end point of the trial (OBD v MTD) and the number of institutions involved in the phase I trial (Table 6; P = .32). However, there was a significant difference in the number of patients per study between MTD and OBD studies (P < .01), and MTD studies accrued significantly more patients than OBD studies. MTD studies accrued a median of 31 patients (range, 12 to 99) as opposed to OBD studies that accrued a median of 22 patients (range, 9 to 68). There was also a significant difference in the accrual time per study between the MTD (median, 21 months; range, 6 to 82 months) and the OBD studies (median, 19 months; range, 3 to 30 months; P = .03).
To gain insight into the prevalence and factors associated with multi-institutional phase I studies, we reviewed phase I trials published in two major clinical oncology journals. We recognize that a potential bias in our study may come from the analysis of two journals based in the United States only. However, it is unlikely that trial designs differ extensively elsewhere in the world. Our analysis showed that 56% of phase I studies were performed at a single institution. We were unable to demonstrate an increase in the percentage of multi-institutional phase I studies during the studied period. We did not, however, investigate the period before 1998, which is a limitation of our study. We then attempted to analyze factors that are associated with the decision to place the trial at multiple sites rather than one single institution. We found no association between the number of institutions, the sponsor, or the mechanism of action of the studied agent. Thus multi-institutional phase I studies are not unique to the pharmaceutical sponsored–studies but occur with the same frequency as with NIH sponsorship. Furthermore, phase I studies are not unique to so-called targeted agents but occur just as frequently in trials of cytotoxic agents. There are several potential reasons why multi-institutional phase I studies are being performed. The first theoretical argument relates to shortening the accrual time. Because sponsors follow strict drug development timelines, rapid accrual and determination of a recommended phase II are crucial. This concept has been borrowed from our knowledge of phase II and III studies, in which many centers routinely are involved to accelerate accrual. Our analysis clearly demonstrates that accrual time is not shortened by performing a multi-institutional phase I study. We propose that this lack of improvement in accrual rate relates to several factors. One may be that, given the proliferation of new agents that require clinical development, each site is opening several competing studies, which would hamper accrual. The second possibility may relate to the investigators’ lack of enthusiasm and interest in a trial when multiple principal investigators are involved. The second potential reason for performing multi-institutional phase I trials is to gain early efficacy data, thus deviating from the traditional end points of DLT, MTD, and recommended phase II dose determination. These so-called multifunctional phase I studies5 attempt to obtain early evidence of clinical efficacy (not a traditional primary end point in phase I studies) and to allow the sponsor to take less risks in drug development or even to bypass the need for phase II studies. Indeed, the concept of disease-specific phase I studies (ie, performance of a phase I study in a single disease type) may stem from the wish to obtain such early efficacy data. Our data show that 13% of studies in solid tumor were disease-specific phase I trials, and these studies were more likely to be multi-institutional in nature. A third possible reason for performing multi-institutional studies may be that additional sites are recruited by sponsors as insurance against the risk that a single site may have unexpected delays in opening the study. The retrospective analysis of this study cannot fully address this hypothesis. Nevertheless, the absence of a difference in the accrual time between single- and multiple-site studies argues against the benefit of this approach for sponsors. Multi-institutional phase I trials may pose potential problems. Does investigator experience with any given agent have importance in terms of recognition of toxicity? If so, does dilution of this experience by multiplying the number of investigators on a given trial have an effect on this experience?5 In the setting of multi-institutional studies, toxicity assessments that use the National Cancer Institute Common Terminology Criteria for Adverse Events are transmitted to the sponsor. Grade 1 to 2 toxicities are generally not reported in real time and may go unnoticed for a period of time at institutions in which the initial toxicity did not occur. As grade 1 to 2 toxicities can be a prelude to higher grade toxicities, the absence of direct experience by an investigator may affect the investigator's ability to recognize higher grade toxicities in a timely fashion. Another practical problem with multi-institutional phase I studies relates to activation of the trial at different times by the participating institutions. This too will prevent certain investigators from acquiring experience with lower doses of the investigational agent, and they may then put patients only on the highest dose levels, in which toxicities are more likely to occur. Our study also demonstrated an increased number of patients enrolled on multi-institutional phase I studies. Most, but not all, phase I designs attempt to expose the least amount of patients to subtherapeutic doses, as most responses occur within 80% to 120% of the recommended phase II dose.1 When an increased number of patients are enrolled per trial, there is a risk that more patients will be exposed to subtherapeutic doses, although some designs (eg, the continual reassessment method) attempt to maximize the posterior probability that a patient is treated at a dose that is associated with the target DLT rate. Our finding that pharmaceutical-sponsored trials enroll more patients than NIH-sponsored phase I trials suggests a role for the sponsor in the increase in patient numbers. This may be a result of the logistics of slot allocation to each site. Our study also shows that many phase I studies do not determine a final MTD. This is strongly associated with so-called targeted therapies (36% of all such studies). As mentioned earlier, however, it is not associated with the number of institutions involved. Also, to our surprise, studies that used an OBD end point enrolled fewer patients than studies that used an MTD end point. In addition, OBD studies had a shorter accrual time; however, this may be explained by the fewer number of patients enrolled. One could hypothesize that, given less significant myelotoxicity with many targeted agents, investigators would be evaluating more dose levels with these agents; therefore, studies that used an OBD would be enrolling an excess number of patients. The fewer number of patients enrolled on OBD studies may reflect the early conclusion of these phase I studies once a target plasma concentration is achieved. Another alternative maybe that newer, accelerated titration designs are more commonly used with targeted agents or that the starting dose is higher than prior traditional starting doses for phase I studies. Although we saw no change with time in the percentage of OBD studies, we believe this may reflect the time period we analyzed. As most studies have been completed a year or more before publication, it is conceivable that analysis of more recent studies would show a trend toward an increase in the OBD determination. In conclusion, multi-institutional phase I studies fail to shorten accrual time and are associated with an increased number of patients per trial. The utility of placing phase I trials at multiple sites is thus unclear. Sponsors of such trials should consider the potential increased cost of involvement of several institutions without any apparent benefit of such an approach. The reason for a greater number of patients on pharmaceutical-sponsored trials also needs to be evaluated further.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a "U" are those for which no compensation was received; those relationships marked with a "C" were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: Afshin Dowlati, Eli Lilly & Co, Genentech, Aventis Research Funding: Afshin Dowlati, GlaxoSmithKline, Celgene, OSI Pharmaceuticals Expert Testimony: None Other Remuneration: None
Conception and design: Afshin Dowlati Administrative support: Afshin Dowlati, Lauren Patrick, Pingfu Fu Provision of study materials or patients: Afshin Dowlati Collection and assembly of data: Afshin Dowlati, Sudhir Manda, Joseph Gibbons, Lauren Patrick, Pingfu Fu Data analysis and interpretation: Afshin Dowlati, Sudhir Manda, Joseph Gibbons, Scot C. Remick, Pingfu Fu Manuscript writing: Afshin Dowlati, Sudhir Manda, Joseph Gibbons, Scot C. Remick, Lauren Patrick, Pingfu Fu Final approval of manuscript: Afshin Dowlati, Sudhir Manda, Joseph Gibbons, Scot C. Remick, Lauren Patrick, Pingfu Fu
Supported by Grants No. 5K23 CA109348-01 and U01 CA62502 from the National Institutes of Health. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Eisenhauer EA, O'Dwyer PJ, Christian M, et al: Phase I clinical trial design in cancer drug development. J Clin Oncol 18:684-692, 2000 2. Adjei AA: What is the right dose? The elusive optimal biologic dose in phase I clinical trials. J Clin Oncol 24:4054-4055, 2006 3. Dowlati A, Robertson K, Radovotyicivh T, et al: Novel phase I dose de-escalation design trial to determine the biological modulatory dose of the anti-angiogenic agent SU5416. Clin Cancer Res 11:7938-7944, 2005 4. Dowlati A, Haaga J, Remick SC, et al: Sequential tumor biopsies in early phase clinical trials of anticancer agents for pharmacodynamic evaluation. Clin Cancer Res 7:2971-2976, 2001 5. Tolcher AW, Takimoto CH, Rowinsky EK: The multifunctional, multi-institutional, and sometimes even global phase I study: A better life for phase I evaluations or just "living large"? J Clin Oncol 20:4276-4278, 2002 Submitted July 5, 2007; accepted December 11, 2007.
Related Editorial
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
