Phase I trial design: are new methodologies being put into practice?

The 3 + 3 design in dose-finding studies with small sample sizes: Pitfalls and possible remedies

In the last few years, numerous novel designs have been proposed to improve the efficiency and accuracy of phase I trials to identify the maximum-tolerated dose (MTD) or the optimal biological dose (OBD) for noncytotoxic agents. However, the conventional 3+3 approach, known for its and poor performance, continues to be an attractive choice for many trials despite these alternative suggestions. The article seeks to underscore the importance of moving beyond the 3+3 design by highlighting a different key element in trial design: the estimation of sample size and its crucial role in predicting toxicity and determining the MTD. We use simulation studies to compare the performance of the most used phase I approaches: 3+3, Continual Reassessment Method (CRM), Keyboard and Bayesian Optimal Interval (BOIN) designs regarding three key operating characteristics: the percentage of correct selection of the true MTD, the average number of patients allocated per dose level, and the average total sample size. The simulation results consistently show that the 3+3 algorithm underperforms in comparison to model-based and model-assisted designs across all scenarios and metrics. The 3+3 method yields significantly lower (up to three times) probabilities in identifying the correct MTD, often selecting doses one or even two levels below the actual MTD. The 3+3 design allocates significantly fewer patients at the true MTD, assigns higher numbers to lower dose levels, and rarely explores doses above the target dose-limiting toxicity (DLT) rate. The overall performance of the 3+3 method is suboptimal, with a high level of unexplained uncertainty and significant implications for accurately determining the MTD. While the primary focus of the article is to demonstrate the limitations of the 3+3 algorithm, the question remains about the preferred alternative approach. The intention is not to definitively recommend one model-based or model-assisted method over others, as their performance can vary based on parameters and model specifications. However, the presented results indicate that the CRM, Keyboard, and BOIN designs consistently outperform the 3+3 and offer improved efficiency and precision in determining the MTD, which is crucial in early-phase clinical trials.

Comparison of efficacy discrepancy between early-phase clinical trials and phase III trials of PD-1/PD-L1 inhibitors

BACKGROUND

Phase III clinical trials are pivotal for evaluating therapeutics, yet a concerning failure rate has been documented, particularly impacting oncology where accelerated approvals of immunotherapies are common. These failures are predominantly attributed to a lack of therapeutic efficacy, indicating overestimation of results from phase II studies. Our research aims to systematically assess overestimation in early-phase trials involving programmed cell death-1 (PD-1)/programmed cell death-ligand 1(PD-L1) inhibitors compared with phase III trials and identify contributing factors.

METHODS

We matched 51 pairs of early-phase and phase III clinical trials from a pool of over 9,600 PD-1/PD-L1 inhibitor trials. The matching criteria included identical treatment regimens, cancer types, treatment lines, and biomarker enrichment strategies. To assess overestimation, we compared the overall response rates (ORR) between early-phase and phase III trials. We established independent variables related to eligibility criteria, and trial design features of participants to analyze the factors influencing the observed discrepancy in efficacy between the two phases through univariable and multivariable logistic analyses.

RESULT

Early-phase trial outcomes systematically overestimated the subsequent phase III results, yielding an odds ratio (OR) comparing ORR in early-phase versus phase III: 1.66 (95% CI: 1.43 to 1.92, p<0.05). This trend of inflated ORR was consistent across trials testing PD-1/PD-L1 monotherapies and combination therapies involving PD-1/PD-L1. Among the examined factors, the exclusion of patients with autoimmune diseases was significantly associated with the disparity in efficacy between early-phase trials and phase III trials (p=0.023). We calculated a Ward statistic of 2.27 to validate the effectiveness of the model.

CONCLUSION

These findings underscore the tendency of overestimation of efficacy in early-phase trials involving immunotherapies. The observed differences could be attributed to variations in the inclusion of patients with autoimmune disorders in early-phase trials. These insights have the potential to inform stakeholders in the future development of cancer immunotherapies.

Beyond the dose-limiting toxicity period: Dermatologic adverse events of patients on phase 1 trials of the Cancer Therapeutics Evaluation Program

BACKGROUND

Dermatologic adverse events (AEs) can be key determinants of overall drug tolerability and of the maximum tolerated and recommended phase 2 doses in phase 1 trials. The authors present the largest dedicated analysis of dermatologic AEs on phase 1 trials to date.

METHODS

Data from a prospectively maintained database of patients with solid tumors who were enrolled onto Cancer Therapeutics Evaluation Program (CTEP)-sponsored phase 1 trials of cytotoxic or molecularly targeted agents (MTAs) from 2000 to 2010 were analyzed. Cumulative incidence, site, and type of drug-related dermatologic AEs were described and compared. The timing of worst drug-related dermatologic AEs was summarized.

RESULTS

In total, 3517 patients with solid tumors and 6165 unique, drug-related dermatologic AEs were analyzed, including 1545 patients on MTA-only trials, 671 on cytotoxic-only trials, and 1392 on combination MTA and cytotoxic trials. Of 1270 patients who had drug-related dermatologic events, the timing of the worst AE was as follows: 743 (cycle 1), 303 (cycle 2), and 224 (cycle 3 or later). Although the cumulative incidence of grade ≥3 drug-related AEs increased to 2.4% by cycle 6, it was only 1.6% at the end of cycle 1. The cumulative incidence of drug-related AEs was highest in patients who received MTA-only therapy (P < .001) and differed by dose level (P < .001). In patients who received MTA-only therapy, drug-related AEs were most common for combination kinase inhibitor-containing therapy (P < .001).

CONCLUSIONS

A substantial proportion of drug-related dermatologic AEs occur after the traditional dose-limiting toxicity monitoring period of phase 1 clinical trials. Future designs should account for late toxicities.

Evolution of clinical trial design in early drug development: systematic review of expansion cohort use in single-agent phase I cancer trials

PURPOSE

To evaluate the use and objectives of expansion cohorts in phase I cancer trials and to explore trial characteristics associated with their use.

METHODS

We performed a systematic review of MEDLINE and EMBASE, limiting studies to single-agent phase I trials recruiting adults and published after 2006. Eligibility assessment and data extraction were performed by two reviewers. Data were assessed descriptively, and associations were tested by univariable and multivariable logistic regression.

RESULTS

We identified 611 unique phase I cancer trials, of which 149 (24%) included an expansion cohort. The trials were significantly more likely to use an expansion cohort if they were published more recently, were multicenter, or evaluated a noncytotoxic agent. Objectives of the expansion cohort were reported in 74% of trials. In these trials, safety (80%), efficacy (45%), pharmacokinetics (28%), pharmacodynamics (23%), and patient enrichment (14%) were cited as objectives. Among expansion cohorts with safety objectives, the recommended phase II dose was modified in 13% and new toxicities were described in 54% of trials. Among trials aimed at assessing efficacy, only 11% demonstrated antitumor activity assessed by response criteria that was not previously observed during dose escalation.

CONCLUSION

The utilization of expansion cohorts has increased with time. Safety and efficacy are common objectives, but 26% fail to report explicit aims. Expansion cohorts may provide useful supplementary data for phase I trials, particularly with regard to toxicity and definition of recommended dose for phase II studies.

Phase I clinical trial of vinorelbine in tumor-bearing cats

BACKGROUND

Vinorelbine (VRL) has been investigated in dogs, but its use in cats has not been studied.

HYPOTHESIS/OBJECTIVES

To determine the maximal tolerated dose (MTD) and dose-limiting toxicity (DLT) of VRL in tumor-bearing cats.

ANIMALS

Cats were included in this prospective phase I trial if they had confirmed malignancy, received ≥1 VRL treatment, and had adequate follow-up. Previous treatment was acceptable, but concurrent chemotherapy or radiotherapy was not permitted.

METHODS

Using a modified phase I design, cats were enrolled in cohorts of 3 at a starting dosage of 9 mg/m(2) . Cats tolerating the first treatment well were eligible to receive additional VRL treatments at escalating dosages; escalations beyond the perceived MTD were permitted based on individual tolerance. Intended treatment interval was 7 days. Patient histories, physical examinations, and complete blood counts were performed weekly.

RESULTS

Nineteen cats were included. Sixty-one VRL treatments were administered. Median number of treatments was 2 (range, 1-9). Starting dosages were 9-12 mg/m(2) . Maximal dosage administered was 15.5 mg/m(2) . The MTD was 11.5 mg/m(2) . Acute DLTs were neutropenia, vomiting, and nephrotoxicity. Other notable toxicities were weight loss and anemia.

CONCLUSIONS AND CLINICAL IMPORTANCE

Vinorelbine is tolerated in cats at a weekly interval. Recommended starting dosage is 11.5 mg/m(2) . Neutropenia was transient, lasting <7 days; vomiting was self-limiting in most cases. Although VRL-associated nephrotoxicity has not been reported, potential attribution of this toxicity to VRL must not be discounted. Further investigation of the efficacy of VRL in feline malignancies is warranted.

How First-Time-in-Human Studies Are Being Performed: A Survey of Phase I Dose-Escalation Trials in Healthy Volunteers Published Between 1995 and 2004

First-time-in-human studies are small, time-lagged dose-escalation studies including volunteer subjects evaluating safety and tolerability. There is little consensus in the design of a first-time-in-human study, and it is difficult to get an overview of studies performed. One hundred five studies comprising 3323 healthy volunteers published in the 5 major clinical pharmacology journals since 1995 were analyzed. The average trial was placebo controlled, double blind including 32 subjects at 5 dose levels but with great variation in cohort size and dose-escalation method. The parallel single-dose design was the most common design, with the crossover designs being more frequent in the early publications. Despite discussions on the optimization of phase I trials, little seems to be happening. The development of study designs and evaluation methods for cancer trials is extensive, but formal statistically based methods and more scientific study designs are unusual in phase I dose-escalation trials in healthy volunteers.

What does a modified-Fibonacci dose-escalation actually correspond to?

BACKGROUND

In most phase I oncology trials, it is often stated that the dose increments follow a "modified-Fibonacci sequence". This term, however, is vague.

METHODS

To better characterize this sequence, we reviewed 81 phase I trials based on this concept.

RESULTS

Out of 198 phase I oncology trials, 81 (41%) are based on modified-Fibonacci series. Actual incremental ratios varied in a large range from 0.80 to 2.08. The median of actual increments was about 2.00, 1.50, 1.33, 1.33, 1.33, 1.33, 1.30, 1.35…. The "modified Fibonacci-sequence" gathers heterogeneous variation of the genuine sequence, which does not tend to a constant number at higher dose-levels.

CONCLUSION

This confusing term should be avoided.

Life-expectancy of patients enrolled in phase 1 clinical trials: a systematic review of published prognostic models

Life-expectancy superior to 3 months is a key-eligibility criterion for contemporary oncology phase 1 trials. Nevertheless, there is no reliable and consensual guidance for estimating this criterion. We have conducted a systematic review of published studies investigating the risk factor for 90-day mortality and the inherent generated scores. Nine studies have been published on this topic. Only two of these prognostic models have been validated on an independent dataset. Most of the models are based on a very subjective and investigator-dependent parameter: the performance status. The predictive performance of these prognostic models is poorly evaluated.

Plasma levels of vascular endothelial growth factor during and after radiotherapy in combination with celecoxib in patients with advanced head and neck cancer

Celebrex and radiotherapy in advanced head and neck cancer. This phase I dose-escalation study seeks to determine the phase II recommended dose of cyclooxygenase type 2 (COX-2) inhibitor in patients with locally advanced squamous cell head and neck (H&N) cancer, treated with accelerated radiotherapy. Anti-vasculogenic effect of this treatment on serum vascular endothelial growth factor (VEGF) is examined. Patients were irradiated with curative intent (72Gy in 6weeks). Celecoxib was administered throughout the radiotherapy course. Serum VEGF level were tested during radiotherapy and in follow-up. Tumor specimens were stained to quantify the COX-2 expression. Thirty-two patients completed the treatment. The dose of celecoxib was escalated (200, 400 and 800mg bid, then de-escalated to 600mg bid). The acute toxicity related to the treatment in the first and second cohort did not reach grade III; in the third cohort three patients had grade III radiation toxicity and one had celecoxib-related toxicity. In the last fourth cohort the toxicity was acceptable. Significant VEGF level drop (p=0.011) was found between radiation day 1 and post-treatment visit. Significant decrease (p=0.022) of the VEGF level was shown in patients with high COX-2 expression in the tumor. Phase II recommended dose of celecoxib combined with accelerated radiotherapy in advanced H&N cancer was 600mg bid. A significant decrease of the post-treatment serum VEGF level compared to the initial level was noticed only in patients with high COX-2 expression in tumors.

Toxicity burden score: a novel approach to summarize multiple toxic effects

BACKGROUND

Toxicity data from cancer trials are summarized into a single outcome, dose-limiting toxicity (DLT), which does not account for multiple lower grade toxic effects nor differentiates between toxicity types and gradations within DLT.

METHODS

Toxicity data were summarized into a toxicity burden score (TBS) using a weighted sum. The severity weights were estimated via regression using historical data. We demonstrated the method using historical data from a bortezomib trial and illustrated the advantages of defining DLT based on TBS in a simulated dose-finding trial.

RESULTS

The estimated weights were 0.17, 0.40 and 0.85 for grade 1/2, grade 3 and grade 4 platelets, respectively; 0.19, 0.64, 1.03 and 2.53 for grade 1, 2, 3 and 4 neuropathy, respectively and 0.17 for each grade 3 or higher nonhematologic toxic effects unrelated to treatment. In the simulated trial, the probability of selecting doses above the maximum tolerated dose decreased when using the DLT defined based on TBS.

CONCLUSIONS

TBS is a feasible approach to summarize toxicity. It includes information from the grades and types of multiple toxic effects and can be applied in all phases of drug development. Further efforts should focus on validating the method in a large prospective study before applying it in practice.

Dose-levels and first signs of efficacy in contemporary oncology phase 1 clinical trials

PURPOSE

Phase 1 trials play a crucial role in oncology by translating laboratory science into efficient therapies. Molecular targeted agents (MTA) differ from traditional cytotoxics in terms of both efficacy and toxicity profiles. Recent reports suggest that higher doses are not essential to produce the optimal anti-tumor effect. This study aimed to assess if MTA could achieve clinical benefit at much lower dose than traditional cytotoxics in dose seeking phase 1 trials.

PATIENTS AND METHODS

We reviewed 317 recent phase 1 oncology trials reported in the literature between January 1997 and January 2009. First sign of efficacy, maximum tolerated dose (MTD) and their associated dose level were recorded in each trial.

RESULTS

Trials investigating conventional cytotoxics alone, MTA alone and combination of both represented respectively 63.0% (201/317), 23.3% (74/317) and 13.7% (42/317) of all trials. The MTD was reached in 65.9% (209/317) of all trials and was mostly observed at the fifth dose level. First sign of efficacy was less frequently observed at the first three dose-levels for MTA as compared to conventional cytotoxics or combinations regimens (48.3% versus 63.2% and 61.3%). Sign of efficacy was observed in the same proportion whatever the treatment type (73-82%). MTD was less frequently established in trials investigating MTA alone (51.3%) or combinations (42.8%) as compared to conventional cytotoxic agents (75.6%).

CONCLUSION

First sign of efficacy was less frequently reported at the early dose-levels and MTD was less frequently reached in trials investigating molecular targeted therapy alone. Similar proportion of trials reported clinical benefit.

A systematic methodology review of phase I radiation dose escalation trials

BACKGROUND AND PURPOSE

The purpose of this review is to evaluate the methodology used in published phase I radiotherapy (RT) dose escalation trials. A specific emphasis was placed on the frequency of reporting late complications as endpoint.

MATERIALS AND METHODS

We performed a systematic literature review using a predefined search strategy to identify all phase I trials reporting on external radiotherapy dose escalation in cancer patients.

RESULTS

Fifty-three trials (phase I: n = 36, phase I-II: n = 17) fulfilled the inclusion criteria. Of these, 20 used a modified Fibonacci design for the RT dose escalation, but 32 did not specify a design. Late toxicity was variously defined as > 3 months (n = 43) or > 6 months (n = 3) after RT, or not defined (n = 7). In only nine studies the maximum tolerated dose (MTD) was related to late toxicity, while only half the studies reported the minimum follow-up period for dose escalation (n = 26).

CONCLUSION

In phase I RT trials, late complications are often not taken into account and there is currently no consensus on the methodology used for radiation dose escalation studies. We therefore propose a decision-tree algorithm which depends on the endpoint selected and whether a validated early surrogate endpoint is available, in order to choose the most appropriate study design.

Justification of the starting dose as the main determinant of accrual time in dose-seeking oncology phase 1 trials

INTRODUCTION

New drug development is a time- and resource-consuming process. Phase 1 trials constitute a major key-step of this development. Shortening the accrual time is of major importance.

METHODS

292 published phase-1-trials were retrospectively reviewed to establish the determinants of accrual time using Log-rank test and then Cox Model.

RESULTS

Out of 292 trials (1997-2008), only 107 reports (36%) described the accrual time (median: 20 months, 5-72). Phase-2-recommended dose was established in 87 studies (81%). Most studies investigated regimens including cytotoxic drugs (77%) or molecular targeted therapies (29%). Under univariate analysis, two parameters shortened the accrual time: studies conducted in USA vs. other places (19 vs. 21 months p = 0.03) and regimen with more than 2 dose-escalated drugs (13 vs. 21 months, p = 0.003). One parameter was significantly associated with longer accrual time: starting dose justified by animal toxicology data vs. previous clinical trials (22 vs. 19 months, p = 0.03). Most of parameters did not significantly affect the accrual time: nature of investigated drugs, duration of treatment cycle, phase 1 dedicated to specific tumoral subtypes, number of centers, method of drug escalation (classical 3+3 vs. accelerated titration design), type of increment (modified Fibonacci method vs. others) and presence of expansion of cohort at the phase-II-recommended dose. Cox model analysis retained one determinant: starting dose justified by animal toxicology data: HR = 2.00 [1.45-5.20], p = 0.047.

CONCLUSION

Few parameters influence the accrual time of dose-escalation phase-1 trials. Real first-in-man phase 1 studies based on starting dose estimated from animal toxicological data require longer accrual time.

Dose escalation methods in phase I cancer clinical trials

Phase I clinical trials are an essential step in the development of anticancer drugs. The main goal of these studies is to establish the recommended dose and/or schedule of new drugs or drug combinations for phase II trials. The guiding principle for dose escalation in phase I trials is to avoid exposing too many patients to subtherapeutic doses while preserving safety and maintaining rapid accrual. Here we review dose escalation methods for phase I trials, including the rule-based and model-based dose escalation methods that have been developed to evaluate new anticancer agents. Toxicity has traditionally been the primary endpoint for phase I trials involving cytotoxic agents. However, with the emergence of molecularly targeted anticancer agents, potential alternative endpoints to delineate optimal biological activity, such as plasma drug concentration and target inhibition in tumor or surrogate tissues, have been proposed along with new trial designs. We also describe specific methods for drug combinations as well as methods that use a time-to-event endpoint or both toxicity and efficacy as endpoints. Finally, we present the advantages and drawbacks of the various dose escalation methods and discuss specific applications of the methods in developmental oncotherapeutics.

Two-stage model-based clinical trial design to optimize phase I development of novel anticancer agents

Background The phase I program of anticancer agents usually consists of multiple dose escalation studies to select a safe dose for various administration schedules. We hypothesized that pharmacokinetic and pharmacodynamic (PK–PD) modeling of an initial phase I study (stage 1) can be used for selection of an optimal starting dose for subsequent studies (stage 2) and that a post-hoc PK–PD analysis enhances the selection of a recommended dose for phase II evaluation. The aim of this analysis was to demonstrate that this two-stage model-based design, which does not interfere in the conduct of trials, is safe, efficient and effective. Methods PK and PD data of dose escalation studies were simulated for nine compounds and for five administration regimens (stage 1) for drugs with neutropenia as dose-limiting toxicity. PK–PD models were developed for each simulated study and were used to determine a starting dose for additional phase I studies (stage 2). The model-based design was compared to a conventional study design regarding safety (number of dose-limiting toxicities (DLTs)), efficiency (number of patients treated with a dose below the recommended dose) and effectiveness (precision of dose selection). Retrospective data of the investigational anticancer drug indisulam were used to show the applicability of the model-based design. Results The model-based design was as safe as the conventional design (median number of DLTs = 3) and resulted in a reduction of the number of patients who were treated with a dose below the recommended dose (−27%, power 89%). A post-hoc model-based determination of the recommended dose for future phase II studies was more precise than the conventional selection of the recommended dose (root mean squared error 8.3% versus 30%). Conclusions A two-stage model-based phase I design is safe for anticancer agents with dose-limiting myelosuppression and may enhance the efficiency of dose escalation studies by reducing the number of patients treated with a dose below the recommended dose and by increasing the precision of dose selection for phase II evaluation.

Lenalidomide, adriamycin, and dexamethasone (RAD) in patients with relapsed and refractory multiple myeloma: a report from the German Myeloma Study Group DSMM (Deutsche Studiengruppe Multiples Myelom)

We conducted a phase 1/2 trial combining lenalidomide (R) with adriamycin (A) and dexamethasone (D) for relapsed and relapsed-refractory myeloma to determine tolerability and efficacy of this novel regimen, RAD, delivered for six 28-day cycles. A total of 69 intensively pretreated patients with a median age of 65 years (range, 46-77 years) were enrolled. Using pegfilgrastim (G), the maximum tolerated dose (MTD) was formally not reached at the highest dose level (R, 25 mg on days 1-21; A, 9 mg/m(2) intravenously on days 1-4; and D, 40 mg on days 1-4 and 17-20; dose level 5+G), which was then used to determine efficacy. Grades 3/4 neutropenia and thrombocytopenia were seen in 48% and 38% of patients, respectively. Thromboembolic events occurred in 4.5% and severe infections in 10.5% of patients. On an intent-to treat analysis, overall response rate (ORR) was 73% for the whole study and 77% including 74% complete response (CR) plus very good partial response (VGPR) for dose level 5+G. Response rates and progression-free survival did not differ between relapsed and relapsed-refractory patients. Deletion of chromosome 17p and elevated beta(2)-microglobulin were associated with significantly inferior response and shortened time to progression. In conclusion, RAD induces substantial and durable remission with an acceptable toxicity profile in patients with relapsed and relapsed-refractory myeloma. This trial was registered at www.ClinicalTrials.gov as no. NCT00306813.

Translation of Innovative Designs Into Phase I Trials

Purpose

Phase I clinical trials of new anticancer therapies determine suitable doses for further testing. Optimization of their design is vital in that they enroll cancer patients whose well-being is distinctly at risk. This study examines the effectiveness of knowledge transfer about more effective statistical designs to clinical practice.

Methods

We examined abstract records of cancer phase I trials from the Science Citation Index database between 1991 and 2006 and classified them into clinical (dose-finding trials) and statistical trials (methodologic studies of dose-escalation designs). We then mapped these two sets by tracking which trials adopted new statistical designs.

Results

One thousand two hundred thirty-five clinical and 90 statistical studies were identified. Only 1.6% of the phase I cancer trials (20 of 1,235 trials) followed a design proposed in one of the statistical studies. These 20 clinical studies showed extensive lags between publication of the statistical paper and its translation into a clinical paper. These 20 clinical trials followed Bayesian adaptive designs. The remainder used variations of the standard up-and-down method.

Conclusion

A consequence of using less effective designs is that more patients are treated with doses outside the therapeutic window. Simulation studies have shown that up-and-down designs treated only 35% of patients at optimal dose levels versus 55% for Bayesian adaptive designs. This implies needless loss of treatment efficacy and, possibly, lives. We suggest that regulatory agencies (eg, US Food and Drug Administration) should proactively encourage the adoption of statistical designs that would allow more patients to be treated at near-optimal doses while controlling for excessive toxicity.

Risks and benefits associated with novel phase 1 oncology trial designs

BACKGROUND

Although aggressive dose escalation strategies were designed to improve the risk-benefit profile of phase 1 oncology trials, they have not been adequately studied. The prevalence of several novel trial designs and their association with a variety of clinical endpoints was evaluated.

METHODS

A review of the literature was performed to identify phase 1 oncology studies of cytotoxic agents published from 2002 through 2004.

RESULTS

Of 955 phase 1 oncology articles initially identified, 149 studies, comprising 4532 patients, were analyzed. Only 34% of studies utilized aggressive dose escalation schemes, 22% permitted intrapatient dose escalation, and only 28% enrolled fewer than 3 patients to any dose level. Studies that allowed intrapatient dose escalation or used fewer than 3 patients per dose were not associated with different rates of response or toxicity, nor did they increase the number of patients who received the recommended phase 2 dose. However, aggressive dose escalations were associated with increased rates of both hematologic (17% vs 10%) and nonhematologic (17% vs 13%) toxicity with similar response rates. Only studies that used conservative dose escalation designs and those that studied U.S. Food and Drug Administration (FDA)-approved agents required fewer patients to complete a trial. Trials of FDA-approved agents were also associated with higher response rates than trials of non-FDA-approved agents (10% vs 2%), without an increased risk of toxicity.

CONCLUSIONS

Several novel aggressive design strategies intended to improve the risk-benefit profile of phase 1 oncology trials are not associated with improved clinical outcome, and may be harmful in certain instances.

Phase I Studies of Chemotherapeutic Agents in Cancer Patients: A Review of the Designs

I review the designs available for Phase I dose-finding studies of chemotherapeutic agents in cancer patients. The designs are based on the assumption that both efficacy and toxicity increase with dose, and thus attempt to minimize the number of patients treated at low doses, and also to minimize the chance that patients will be treated at excessively toxic or lethal doses. The designs fall into two classes: rule-based and model-guided. Rule-based designs can always determine a reasonable maximum tolerable dose based on observed toxicity, but when model assumptions are not satisfied, many model-guided designs will not.

Using the continual reassessment method: lessons learned from an EORTC phase I dose finding study

Many clinicians often do not feel comfortable with the Continual Reassessment Method (CRM). This article reviews its implementation, showing the characteristics, advantages and limitations of this method in Phase I studies as an alternative to the classical 'Fibonacci' escalation schema. A two center, dose escalation phase I study of rViscumin was carried out. Thirty-seven patients were included at 14 different dose-levels (10 to 6400 ng/kg). The complete clinical results are presented elsewhere. A 2-step CRM design enables one to speed-up the study and most importantly to obtain an accurate estimate of the maximum tolerated dose (MTD). Different management issues related to a multicenter study are illustrated and we show how the method can go wrong when severe toxicity, or dose limiting toxicity (DLT), is not considered by the clinician as being sufficient to limit dose escalation (here a grade 3 asthenia related to the drug). This would have affected any dose finding methods. We believe that CRM is a good alternative to the standard method from both a statistical and a practical point of view but further methodological research is necessary to address the issues related to the composite nature of the endpoint.

Phase I clinical trial of CEP-2563 dihydrochloride, a receptor tyrosine kinase inhibitor, in patients with refractory solid tumors

CEP-2563 dihydrochloride (CEP-2563) is a soluble lysinyl-beta-alanyl ester of CEP-751, a potent inhibitor of the trk family of receptor tyrosine kinases and the platelet-derived growth factor (PDGF) receptor tyrosine kinase. CEP-2563 was developed because of the limited aqueous solubility of CEP-751. Preclinical models have demonstrated that both CEP-751 and CEP-2563 have antitumor activity in a variety of tumors. A Phase I clinical trial involving 18 patients was conducted to determine the toxicity profile, maximum tolerated dose (MTD), toxicity profile, and pharmacokinetics of CEP-2563 in patients with advanced solid tumors refractory to standard therapy. CEP-2563 was administered over 1 hour via a central venous catheter once daily for five consecutive days every three weeks. A rapid dose titration strategy with initial single patient cohorts and 100% dose escalations was used. With the appearance of drug-related toxicity, escalations were decreased to 50% or 25% and cohorts were expanded to 3 or 6 patients until establishment of the MTD. Dose escalation rapidly proceeded to 320 mg/m(2)/d. The dose limiting toxicities (DLTs) observed were grade 3 hypotension and grade 2 allergic reaction. Other toxicities included anemia, thrombocytopenia, anorexia, asthenia, diarrhea, fatigue, headache, nausea, vomiting, and rash. Pharmacokinetic analysis showed that CEP-2563 is reliably converted to CEP-751. This study demonstrated that single agent CEP-2563 therapy is feasible with acceptable toxicities. The recommended phase II dose is 256 mg/m(2)/d. Rapid dose escalation with single patient cohorts was a safe and efficient method of conducting this phase I trial.

New Paradigm in Dose-Finding Trials: Patient-Specific Dosing and Beyond Phase I

We propose a new paradigm for the clinical evaluation of new cancer therapies. It entails adjusting the search for the optimal dose on the basis of measurable patient characteristics that may be predictive of adverse responses to treatment, and extending this search beyond phase I and into phases II and III. We provide examples of (a) how the fine-tuning of dose may involve utilization of patient-specific attributes to obtain a personalized treatment regimen, and (b) how novel methods for phase I design can be used to update the working dose for the conduct of phase II and III cancer clinical trials. These examples should be interpreted as an enticement for the development of new methods to implement the proposed new paradigm.

Risks and benefits of phase 1 oncology trials, 1991 through 2002

BACKGROUND

Previous reviews of phase 1 oncology trials reported a rate of response to treatment of 4 to 6 percent and a toxicity-related death rate of 0.5 percent. These results may not reflect the rates in current phase 1 oncology trials.

METHODS

We reviewed all nonpediatric phase 1 oncology trials sponsored by the Cancer Therapy Evaluation Program at the National Cancer Institute between 1991 and 2002. We report the rates of response to treatment, of stable disease, of grade 4 toxic events, and of treatment-related deaths.

RESULTS

We analyzed 460 trials involving 11,935 participants, all of whom were assessed for toxicity and 10,402 of whom were assessed for a response to therapy. The overall response rate (i.e., for both complete and partial responses) was 10.6 percent, with considerable variation among trials. "Classic" phase 1 trials of single investigational chemotherapeutic agents represented only 20 percent of the trials and had a response rate of 4.4 percent. Studies that included at least one anticancer agent approved by the Food and Drug Administration constituted 46.3 percent of the trials and had a response rate of 17.8. An additional 34.1 percent of participants had stable disease or a less-than-partial response. The overall rate of death due to toxic events was 0.49 percent. Of 3465 participants for whom data on patient-specific grade 4 toxic events were available, 14.3 percent had had at least one episode of grade 4 toxic events.

CONCLUSIONS

Overall response rates among phase 1 oncology trials are higher than previously reported, although they have not changed for classic phase 1 trials, and toxicity-related death rates have remained stable. Rates of response and toxicity vary, however, among the various types of phase 1 oncology trials.

Phase I-II trial of twice-weekly gemcitabine and concomitant irradiation in patients undergoing pancreaticoduodenectomy with extended lymphadenectomy for locally advanced pancreatic cancer

PURPOSE

Define the maximum tolerated dose (MTD), tolerability, and efficacy of gemcitabine given concomitantly with radiotherapy in patients with locally advanced pancreatic cancer.

METHODS AND MATERIALS

Patients were required to have locally advanced T1-T3 resectable pancreatic cancer. Gemcitabine, given twice weekly before irradiation as a 30-min infusion, was tested at 3 dose levels: 20, 50, and 100 mg/m(2). The radiation dose was 50.4 Gy (ICRU) in 28 fractions. The targeted irradiation volume included the tumor, edema, and a 1-cm margin.

RESULTS

Twenty-eight of 34 patients was eligible for analysis of the treatment. The median age was 67 years (range 38-82). Six patients had T1, 9 had T2, and 19 had T3 diseases (AJCC). Dose-limiting toxicities were Grade 4, fatigue and nausea; Grade 3, thrombocytopenia, diarrhea, and infection. The MTD established was at the 50-mg/m(2) gemcitabine dose. A total of 21 of 28 patients underwent surgery: 18 had pancreaticoduodenectomy, 2 had total pancreatectomy, and 1 for palliative surgery. At the time of analysis, 13 of 28 (46%) were disease-free. The estimated median survival was 25 months and overall survival rate at 2 years (Kaplan-Meier) was 55%.

CONCLUSION

Gemcitabine 50 mg/m(2) given twice weekly with concomitant irradiation induces acceptable and manageable toxicity and might prolong survival.

Multi-species toxicology approaches for oncology drugs

The Toxicology and Pharmacology Branch (T&PB) of the National Cancer Institute (NCI) performs pharmacological and toxicological evaluations of new oncology agents according to an agent-directed paradigm in which all studies are tailored to each agent. The United States Food and Drug Administration (US FDA) requires that preclinical toxicology studies be conducted in two species, a rodent and a non-rodent for all small molecules, and T&PB has successfully used this formula. While pharmacokinetic (PK) studies are considered optional, T&PB routinely develops new methods for plasma/tissue drug analysis and employs this methodology throughout development to determine kinetics in various species and toxicokinetics in the toxicity studies. In the current era of molecular target-based development, the T&PB also develops or employs methodology to evaluate effects of the new chemical entity on appropriate biomarkers in tumour and normal tissues. In this comprehensive programme, T&PB is able to correlate safety and toxicity with both plasma drug levels and biomarker modulation in two species for a seamless entry into Phase I.

Phase I trial and pharmacokinetic study of BMS-247550, an epothilone B analog, administered intravenously on a daily schedule for five days

PURPOSE

The epothilones are a novel class of nontaxane microtubule-stabilizing agents. BMS-247550 is a semisynthetic analog of the natural product epothilone B. We conducted a phase I study administering BMS-247550 as a 1-hour intravenous infusion daily for 5 consecutive days every 21 days.

PATIENTS AND METHODS

Twenty-one patients received BMS-247550 without filgrastim in the first cycle. An additional six patients were enrolled at a starting dose of 8 mg/m2/d with filgrastim support. Twenty-one of the 27 patients had received prior paclitaxel, docetaxel, or both.

RESULTS

One hundred seven cycles were administered to 27 patients. The maximum-tolerated dose was 6 mg/m2 of BMS-247550 administered as a 1-hour intravenous infusion daily for 5 consecutive days every 21 days. Dose-limiting toxicity at a dose of 8 mg/m2/d was neutropenia with or without filgrastim support. Nonhematologic grade 3 toxicities included fatigue (seven cycles), stomatitis (two cycles), and anorexia (one cycle). The mean terminal half-life of BMS-247550 was 16.8 +/- 6.0 hours, the volume of distribution at steady-state was 798 +/- 375 L, and the clearance was 712 +/- 247 mL/min. Objective responses were observed in patients with breast, cervical, and basal cell cancer. Reductions in CA-125 levels were noted in patients with ovarian cancer.

CONCLUSION

The recommended phase II dose of BMS-247550 on the daily schedule for 5 days is 6 mg/m2/d. Neutropenia was dose limiting, but higher doses were tolerated by a large fraction of patients with filgrastim support. Peripheral neuropathy was mild, even after multiple cycles of therapy, and was not dose limiting.

Adaptive dose finding for phase I clinical trials of drugs used for chemotherapy of cancer

Phase I clinical trials of cancer chemotherapy drugs are intended to determine the maximum tolerable dose (MTD). Thestandard method employed is a rule-based dose-escalation scheme in which escalation depends on the number of patients at a dose level that have dose-limiting toxicity (DLT). The MTD is thus defined in terms of the rules and a series of dose levels selected for sampling. For some trials it is desirable to have a more precise definition of the MTD, and to determine the MTD more accurately than possible with the rule-based schemes. Continuous reassessment methods (CRMs) define the MTD to be the dose at which a fixed fraction of patients experience DLT, and thus appear suited to these trials. It is shown, however, that these methods can have failure modes that in fact make them unattractive. An alternative data-driven dose-finding method is described that combines the robustness of the rule-based methods and with features of CRMs. The method has two stages. In the first stage, doses are escalated by a factor of 1.5. In the second stage, which begins at the first instance of DLT, a two-parameter logistic dose-response model estimates the MTD from the DLT experience of all patients. The model is initialized by setting the dose (d10) at which 10 per cent of patients would experience DLT to half the dose at which the first DLT was observed, and the dose (d90) at which 90 per cent would experience DLT to ten times d10. Weights are assigned such that the information at d10 and d90 is equivalent to that of one patient at each of the two doses. Cohorts of three patients are treated in both stages, and the dose for a new cohort in the second stage is the estimated MTD. The only prior information required to specify the design completely is the dose which will be given to the first cohort. Two stopping rules are investigated; among the requirements for these are that at least three (or four) DLTs be observed and at least nine patients be treated in the second stage. Simulations show that a coefficient of variation of approximately 0.4 on a target DLT probability of 0.3 is obtained over a wide variation in dose-response characteristics of the study drug. The performance of the new method is compared to that of rule-based methods.

Phase I clinical trial design in cancer drug development

The past decade has seen the publication of a number of new proposals for the design of phase I trials of anticancer agents. The purpose of these proposals has been to address ethical concerns about treating excessive numbers of patients at subtherapeutic doses of a new agent and to increase the overall efficiency of the process while enhancing the precision of the recommended phase II dose. In early 1998, a workshop of phase I investigators was held under the sponsorship of Bristol-Myers Squibb Pharmaceutical Research Institute (Wallingford, CT) to review the experience to date with novel phase I methodologies, with a particular focus on their efficiency and safety. This report summarizes the material presented. It was concluded that for phase I trials of antineoplastics (cytotoxics), which begin at 0.1 mouse-equivalent LD10 doses, evidence to date suggests that the historic approach of using a modified Fibonacci escalation and three patients per dose level is not necessary and is seldom used. One patient per dose level and more rapid escalation schemes, both empirically based and statistically based, are commonly used with apparent safety. There remain questions, however: Which of the dose escalation schemes is optimal? Are there alternatives to toxicity as a phase I end point, and will these end points be reliable in defining active doses? Answering these questions in a reasonable time frame will be important if new anticancer agents are not to suffer undue delays in phase I evaluation.

Evaluation of rodent-only toxicology for early clinical trials with novel cancer therapeutics

Preclinical toxicology studies are performed prior to phase I trials with novel cancer therapeutics to identify a safe clinical starting dose and potential human toxicities. The primary aim of this study was to evaluate the ability of rodent-only toxicology studies to identify a safe phase I trial starting dose. In addition, the ability of murine studies to predict the quantitative and qualitative human toxicology of cancer therapeutics was studied. Data for 25 cancer drugs were collated for which the preclinical and clinical routes and schedules of administration were either the same (22/25), or closely matched. The maximum tolerated dose/dose lethal to 10% of mice (MTD/LD10) was identified for 24 drugs, and in patients the maximum administered dose (MAD) was associated with dose-limiting toxicity (DLT) in initial clinical trials with 20 compounds. In addition, for 13 agents, the toxicity of the drug at one-tenth the mouse MTD/LD10 was also investigated in rats, following repeated administration (20 doses). A phase I trial starting dose of one-tenth the mouse MTD/LD10 (mg m(-2)) was, or would have been, safe for all 25 compounds. With the exception of nausea and vomiting, which cannot be assessed in rodents, other common DLTs were accurately predicted by the murine studies (i.e. 7/7 haematological and 3/3 neurological DLTs). For two of the 13 drugs studied in rats, repeated administration of one-tenth the mouse MTD/LD10 was toxic, leading to a reduction in the phase I trial starting dose; however, one-tenth the mouse MTD/LD10 was subsequently tolerated in patients. For the 20 drugs where clinical DLT was reached, the median ratio of the human MAD to the mouse MTD/LD10 was 2.6 (range 0.2-16) and the median ratio of the clinical starting dose to the MAD was 35 (range 2.3-160). In contrast, in 13 subsequent phase I trials with 11 of the initial 25 drugs, the median ratio of the clinical starting dose to the MAD was 2.8 (range 1.6-56), emphasizing the value of early clinical data in rapidly defining the dose range for therapeutic studies. For all 25 drugs studied, rodent-only toxicology provided a safe and rapid means of identifying the phase I trial starting dose and predicting commonly encountered DLTs. This study has shown that the routine use of a non-rodent species in preclinical toxicology studies prior to initial clinical trials with cancer therapeutics is not necessary.

Continual reassessment methods in phase I trials of the combination of two drugs in oncology

Most phase I trials in oncology use standard methods for treating successive groups of patients with increasing doses in order to determine the maximum tolerated dose (MTD). These methods have been criticized because they treat many patients at suboptimal dose levels, and do not provide an accurate estimation of the best dose level. Continual reassessment methods for the study of toxicity in single agent phase I trials have recently been advocated since they present many advantages over traditional methods. Although the advantages of these methods are recognized by most clinical investigators, their use is not widespread and their advantages have not yet been universally accepted. A maximum likelihood continual reassessment method was conducted retrospectively and compared to the originally planned standard method in a two drug combination phase I trial in order to study its applicability in this setting. Calculations from the binomial distributions and simulations were used for identifying the MTD, for the proportion of patients treated at the MTD or at one dose level just below, and for the proportion of patients treated at doses above the MTD. If the new method had been applied in this study, the MTD would have been reached much earlier, since, most of the time, higher dose levels were recommended. This result shows the feasibility of the new method in a two-drug setting and its use should be encouraged since fewer patients are treated at suboptimal dose levels or at dose levels above the MTD.