Abstract: Every new anti-cancer drug or drug combination is evaluated for safety and efficacy before being approved. Clinical development of cytotoxic anticancer drugs classically follows three main phases. Phase I trials represent the first administration of a new drug or combination to human beings. Their primary goal is to determine the recommended phase two dose and also to collect toxicity, pharmacokinetic and pharmacodynamic data. Phase II trials are screening studies aimed at identifying signals of anti-tumor activity in a specific tumor type and setting. Phase III trials aim to compare the efficacy of a new treatment with standard of care and can lead to regulatory approval when positive. The recent emergence of molecularly targeted agents has challenged the traditional developmental pathway for anti-cancer drugs. Using biomarker enriched patient populations has been successful for a few agents. Otherwise, new types of trials have been proposed for these agents in an attempt to elucidate their mechanism of action, such as phase 0 trials and "window of opportunity" trials. These two types of trials and the classical three phase trials are discussed in detail.
According to the Pharmaceutical Research and Manufacturers of America, there are more than 800 anti-cancer agents that are currently in active clinical development (http://www.phrma.org). This number represents a 143% rise in the number of oncology drugs being developed in the last decade alone. The increased rate of new anti-cancer therapy investigations can be attributed at least, in part, to improved comprehension of cancer biology mechanisms. As opposed to cytotoxic agents that target DNA, tubulin, or cell division machinery, novel agents have recently emerged that selectively target molecular pathways thought to be critical for tumor survival, growth, and metastases.
Every new drug or drug combination has to be evaluated for safety and efficacy before being approved. The evaluation of new anti-cancer drugs or drug combinations classically follows three main phases. The clinical drug development process has to be rigorous, as each phase of investigation can lead to permanent discontinuation of the studied agent. The failure rate for anti-cancer drugs is approximately 95%, which is high in comparison to drugs developed in other medical specialties (Di Masi and Grabowski, 2007).
In this review, we describe the different phases of the clinical development for a new anti-cancer drug or drug combination and summarize the current challenges in the era of molecularly targeted therapies (Figure 1). We also describe new strategies that have been recently suggested to help improve the development of molecularly targeted agents by providing early data about their mechanisms of action, including phase 0 trials and “window of opportunity” trials.
Phase I Trials
Phase I trials represent the first administration of a new drug or drug combination to human beings. Oncology phase I trials differ from non-oncology phase I trials in that they are usually performed in cancer patients rather than healthy volunteers, a practice that was derived due to the narrow therapeutic index of cytotoxic agents. Most of the time, these trials are also offered to heavily pre-treated patients where no other therapeutic options are available. Therefore, the selection of patients for phase I trials is critical. Patients are required to have good performance status (PS) and adequate organ function to minimize the chance of clinical deterioration from progressive disease that will require premature study withdrawal. Recent studies have demonstrated that serum albumin >35g/L, lactate dehydrogenase of more than upper limit of normal, >2 sites of metastases, and Eastern Cooperative Oncology Group performance status (ECOG PS) >0 influence clinical outcomes in an adverse manner in oncology phase I trial patients (Arkenau et al., 2009; Chau et al., 2010). These findings are useful in helping clinicians to further select patients in this particular setting. In addition, patients should not require receiving medications that may potentially interact with the experimental drug (e.g., drugs that interfere with the cytochrome P450 enzymatic pathway) so as to avoid exacerbating the toxicities of the experimental drug or altering its pharmacokinetics.
The primary goal of phase I cancer trials is to collect data on toxicity, pharmacologic and pharmacodynamic properties of a specific drug or therapy regimen, allowing determination of the recommended phase two dose (RPTD). For cytotoxic agents that target DNA, tubulin, or cell division machinery, efficacy is thought to be partly dose dependent. Therefore, the RPTD was traditionally derived from dose escalation strategies, aimed at identifying the highest possible dose that could be delivered (maximum tolerated dose, MTD). MTD has usually been defined as the dose producing a severe toxicity (dose-limiting toxicity, DLT) in approximately one third of patients. Each toxic event occurring during a phase I trial is graded, usually using the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE). Pharmacokinetic analyses are performed to determine the relationship of increasing dose to drug absorption, distribution, and metabolism. In addition, drug interactions are also investigated in combination regimen phase I trials. Pharmacokinetic parameters that are commonly evaluated include maximal drug concentration (Cmax), drug exposure (Area Under the Curve or AUC), half-life (t1/2), and clearance (CL) of the drug. Pharmacodynamic analyses aim at identifying predictive markers of biological activity. For instance, leucopenia is often a pharmacodynamic marker of clinical efficacy for cytotoxic agents.
Phase I trials consist of successive patient cohorts receiving increasing doses of treatment. Preclinical toxicology studies are used to determine a relatively safe starting dose for first-in-human studies, as well as to anticipate potential toxicities and their reversibility. European guidelines require the evaluation of new drugs in two species, one rodent and one non-rodent (dogs or monkeys) model. For cytotoxic agents, one tenth of the dose that was lethal to 10% of mice (LD10) has been classically considered as a safe starting dose in humans. However, if the non-rodent species tested appears to be more sensitive than the rodent species, then the recommended starting dose is one sixth to one third of the lowest dose that results in toxicity (toxic dose low or TDL) in the non-rodent species (Rozencweig et al., 1981).
The guiding principle for dose escalation in phase I trials is to avoid unnecessary exposure of patients to very low doses of an agent while preserving safety and maintaining rapid accrual. Dose escalation methods are divided into two main categories: the rule-based designs, which include the traditional “3+3″ design; and the model-based designs (Le Tourneau et al., 2009). Rule-based designs assign patients to dose levels according to pre-specified rules based on the occurrence of DLT. The traditional “3+3″ design is by far the most commonly used rule-based dose escalation design. It proceeds with successive cohorts of three patients. The first cohort is treated with the starting dose and if none of the three patients in a cohort experiences DLT, three additional patients will be treated at the next dose level. If one of the first three patients experiences a DLT, three additional patients will be treated at the same dose level. The dose escalation continues until two or more patients among a cohort of three to six patients experience a DLT. Inter-cohort dose escalation usually follows a modified Fibonacci sequence in that each successive dose increment becomes smaller as progressively higher doses are used. It is worth noting that there are some differences in terminology in that some call the dose level where 2 or more patients (out of 3-6 patients) experienced a DLT as the MTD whereas others call the preceding dose level as the MTD. As such, we prefer the term recommended phase two dose (RPTD), which refers to the dose level below that where two or more patients (out of 3-6 patients) experienced a DLT. The traditional “3+3″ design is easy to implement and safe. However, the early dose levels involve treating patients at very low doses which are almost certainly biologically inactive, while few patients actually receive doses at or near the recommended phase two dose. An alternate rule-based strategy, known as the accelerated titration design attempts to minimize this problem. A single patient is treated per cohort, with the dose doubling between each cohort in the absence of any grade two toxicities. As soon as a grade 2 toxicity occurs, dose escalation is switched to the slower and more traditional “3+3″ design. In some of these designs, intrapatient dose escalation is also allowed.
Model-based designs also attempt to reduce the number of patients treated at very low doses. The best-known example of this type of design is the continual reassessment method (O’Quigley et al., 1990; Piantadosi et al., 1998). Model-based designs use statistical models to actively seek a dose level that produces a pre-specified probability of a DLT, usually between 20 and 30%. These model-based designs use all the available data on DLT observed at previous dose levels to model the dose-toxicity curve. These designs may allow faster dose escalation with fewer patients being treated at sub-therapeutic doses. However, the use of this design is challenging, as it needs biostatistical expertise and available software on site to perform model fitting in real time, as well as expedited collection of data from each cohort of patients to fit the model.
Phase Ib studies are designed to investigate combinations of two or more agents (as opposed to phase Ia trials that investigate single agents). The evaluation of these combinations should be based on a strong scientific rationale rather than simple empiricism. Determining the recommended dose for phase II trials of agents to be administered in combination may appear easier than that for single agents, given that the recommended dose and the toxicity profile of each drug are already known. For this reason, phase I combination trials usually explore only a limited number of dose levels. The choice of the dose levels to be tested may be based on several factors including preclinical data, the current standard treatments in tumor types for which the combination is intended, and the expected control arm if the combination under evaluation is planned to be benchmarked in subsequent randomized trials. Several strategies for dose escalation are possible: alternate escalation of the agents, simultaneous escalation of both agents, escalation of one agent to the recommended phase two dose while holding the other agent at a fixed (generally high) dose, or escalation of one agent to the recommended phase two dose while holding the other agent at a low dose. Overall the choice of the dose escalation scheme for drug combination trials is as relevant as it is for first-in-human monotherapy trials. The recommended dosage and schedules can have a critical role in the success or failure of a combination regimen during its subsequent clinical development. Despite this, there are no accepted guidelines for the investigational design of these combinations.
Phase II Trials
Phase II trials are screening studies aiming to identify signals of anti-tumor activity in a homogenous population of patients with a particular tumor type. It is a critical step in drug development, as a positive phase II trial can lead to a large and expensive phase III trial.
The primary goal of anti-cancer treatments in general is to improve overall survival and/or quality of life. As overall survival has limited utility in phase II trials due to the limited length of observation and the confounding effects of subsequent therapies, other endpoints such as progression free survival, biomarkers, and composite (multi parameters) endpoints may be preferable (Dhani et al., 2009). The choice of the most appropriate primary endpoint should be tailored to the disease and drug(s) under investigation. Tumor shrinkage and progression-free survival (PFS) are two commonly used endpoints in phase II trials (Seymour et al., 2010).
Tumor evaluation is usually based on standard criteria using the repeated measurements of selected target lesions, usually utilizing computerized tomography imaging. The first standardized criteria to be commonly used were defined by the World Health Organization (WHO) which assesses tumor size change by adding the products of the two dimensions of each target lesion on imaging. This sum is compared to the smallest sum observed since the start of treatment. Disease progression is defined as a >25% increase of this sum, while a response is defined as a >50% decrease. Since 2000, the WHO criteria have been largely replaced by the RECIST criteria, for which only one dimension measurements are evaluated (Eisenhauer et al., 2009). Disease progression is defined as >20% increase of the sum of the target tumor diameters (long axis for non-nodal lesions and short axis for nodal lesions), while a response is defined as >30% decrease in the sum of diameters. The basic assumption in using response rate as the primary endpoint is that tumor shrinkage correlates with efficacy. Specific imaging techniques or biomarkers changes may also be used to identify signs of anti-tumor activity early on.
PFS, on the other hand, not only considers tumor shrinkage but also tumor stabilization in its definition. The main limitation of PFS in terms of defining its clinical relevance includes the need for frequent and standardized disease assessments, although the latter can be overcome by considering a PFS rate as a prespecified time point as the primary endpoint.
There is currently a passionate debate on whether phase II trials should be randomized or not (Gan et al., 2010). Results of single arm phase II trials need to be compared with data from previous studies, called historical controls (Hunsberger and al., 2009). However, comparisons with historical data may not be valid for several reasons. First, patient selection may differ between studies resulting in selection bias. Second, stage migration may have occurred because of earlier detection, systematic screening, and, more importantly, improvements in imaging. Third, improvements in survival may occur as a result of improvements in supportive care over time even if there are no changes in the efficacy of anti-cancer treatment. Finally, primary endpoint data may have been collected differently or inconsistently in the past or be absent from historical data sets. There has been therefore increased interest in randomized designs for phase II studies in oncology. With randomization, there is less selection bias inherent to single arm studies, although one may argue that the small size of randomized phase II trials actually prevents adequate neutralization of biases between arms. However, randomized phase II trials cannot establish the superiority of one treatment over the other, because there are not enough patients for such comparison, and superiority should be confirmed in a phase III trial (Rubinstein et al., 2009). The Clinical Trial Design Task Force of the National Cancer Institute Investigational Drug Steering Committee recommends the use of randomization for combination trials, as well as for monotherapy trials to optimize dose and schedule or to benchmark activity against known active therapies (Seymour et al., 2010). The need for randomization in other phase II trials should be guided by the characteristics of the phase II trial in question (Gan et al., 2010).
Phase III Trials
Phase III trials are a critical step before licensing and consist of benchmarking the efficacy of a new treatment against the best current standard therapy. Phase III studies are the most rigorous and extensive form of clinical investigation for a new treatment (Thall, 2008). A scientifically valid comparison between two treatment groups depends on the groups being alike as much as possible, with the only exception being the specific treatments under investigation. Without such an assurance, the treatment groups would not be really comparable and any conclusions drawn might be biased. The best way to achieve a valid comparison is through the use of randomization in which a chance mechanism determines the treatment assignment. Randomization will ensure that either the clinician or the patient does not know a specific treatment assignment in advance. The primary benefit of randomization is that it will eliminate both conscious and unconscious biases associated with the selection of a treatment for a given patient. If any of the outcome measures of a phase III trial is subjective, then it is important that the trial be designed as a double-blind, placebo controlled study. Only when both the patient and clinician are unaware of the treatment assignment can their desire for a favorable outcome not potentially bias the results of the trial. Another benefit of the use of placebos is the objective assessment of toxicities. If a randomized controlled trial of a new drug is conducted in a blinded manner, then all unexpected toxicities on the new drug arm are often ascribed to the new drug, frequently a lower rate that would be reported by any unblinded study.
In most oncology phase III trials, the outcome is a time-to-event variable such as PFS or overall survival. Quality of life can also be a primary endpoint, as improvements in quality of life event may still be valuable even if survival is not prolonged, but this is much less common that the use of survival as a primary endpoint. Randomized controlled trials can be classified by their goals. Difference (superiority) trials aim to determine if sufficient evidence exists that one treatment arm is different from another. These trials are by far the most frequent. Equivalence trials aim to determine that two treatment arms are equivalent (or nearly so) and are conducted less often than difference trials, partly because a larger number of patients are required. Equivalance trials are performed to demonstrate that a less expensive or less toxic new treatment provides similar clinical benefit to standard therapy.
A new drug or drug combination might be approved if a rigorously conducted phase III trial met a clinically relevant primary endpoint against current standard of care. After licensing, phase IV post-marketing trials may be undertaken to explore the long-term safety/morbidity of the treatment. These phase IV trials are unfortunately infrequently done.
Challenges for Molecularly Targeted Agents
The emergence of molecularly targeted agents in oncology not only has revolutionized the care of cancer patients, but also changed the daily practice of medical oncologists. Molecularly targeted agents often differ from traditional cytotoxic agents by their administration schedules and routes, their toxicity profiles, and/or the assessment of their anti-tumor activity (Le Tourneau et al., 2010a). Therefore, the emergence of molecularly targeted agents has challenged the traditional clinical development of anti-cancer drugs (Table 1).
In respect to phase I trials, there are no guidelines about how preclinical models and data should be used to define the starting dose for molecularly targeted agents (Le Tourneau et al., 2010b). Molecularly targeted agents have a mode of action different from that of cytotoxic drugs and their toxicity profiles based on off-target effect are often not as easy to predict from their basic pharmacology. The occurrence of non-hematologic DLT is common with molecularly targeted agents, while hematologic toxicity is usually dose-limiting for most cytotoxic agents. One study suggested that non-rodent models may better predict the human maximum tolerated dose than rodent models for cytotoxic agents and may also more optimally assess non-hematologic toxicities (Le Tourneau et al., 2010b). This observation reinforces the recent recommendation by regulatory agencies to perform preclinical safety analysis in both rodent and non-rodent models for molecularly targeted agents.
As efficacy is supposed to increase with dose for cytotoxic agents, the conventional primary endpoint of cytotoxic phase I cancer clinical trials has been toxicity. Molecularly targeted agents modulate specific aberrant pathways in cancer cells while sparing normal tissues, such that toxicity, efficacy, and dose may not have the interdependence that they have for cytotoxic agents. Alternative endpoints besides toxicity have been proposed for phase I trials that evaluate molecularly targeted agents, including assessment of target inhibition in tumors or surrogate tissues, and/or detection of biologically relevant pharmacokinetic levels. The assessment of target inhibition may be one of the most challenging aspects of clinical designs for several reasons. First, tumor tissue or a valid surrogate tissue must be easily accessible. Second, there must be a reliable assay for measuring the effect of the drug on the target. Third, the optimal extent of target inhibition must be known. Because these three conditions are rarely all met, a recommended phase two dose in phase II/III trial that is based solely on the measurement of target inhibition in phase I trial may be suboptimal. Pharmacokinetic endpoints, such as the attainment of plasma drug concentrations that were shown to correlate with biological activity in preclinical studies, may aid dose selection of some molecularly targeted agents in phase I trials. However, these endpoints are appropriate only if sufficient preclinical data exist that demonstrate a convincing pharmacokinetic-pharmacodynamic relationship. The Task Force on Methodology for the Development of Innovative Cancer Therapies, which recently published its recommendations on phase I studies of targeted anticancer therapy, has not provided guidance on dose-escalation specifically for molecularly targeted agents (Booth et al., 2009).
The basic assumption for cytotoxic agents that tumor shrinkage is a surrogate endpoint for efficacy might not hold true for molecularly targeted agents, where tumor shrinkage may be modest and may not meet the empiric criteria defined in studies of cytotoxic agents. Therefore, the use of overall response rate might not be relevant for phase II trials of molecularly targeted agents. The use of overall response rate plus tumor stabilization or PFS might be more relevant in that setting (El-Maraghi and Eisenhauer, 2008). In that sense, randomized discontinuation phase II trials are probably more appropriate for molecularly targeted agents to determine if the drug has a cytostatic effect or if the disease progresses slowly (Ratain et al., 2006). All patients are initially being treated with the drug under investigation, classically for three months. Responding patients continue treatment, while progressing patients are taken off the study. Patients with stable disease are randomized between placebo and continuation of treatment. Drawbacks of randomized discontinuation studies include the necessity of large sample size, in particular if not enough patients have stable disease, and potential ethical problems if regaining disease control cannot be achieved once patients are randomized off the drug.
Primary endpoint for most phase III trials involving molecularly targeted agents should remain overall survival (or an appropriate surrogate endpoint). However, if a target has clearly been identified, a molecularly targeted agent is available to inhibit this target, and a high level of clinical activity has been observed in phase II trials, one may ask whether large phase III trials are still necessary. High level of efficacy observed in a phase II trial in a select patient population could potentially serve as a proof of principle. For example, imatinib, a tyrosine kinase inhibitor targeting the BCR-ABL fusion gene, was approved for the treatment of chronic myeloid leukemia caused by this fusion gene based on phase II trial data only. There are currently more and more smaller trials which enroll patients selected for a particular biologic abnormality and report high levels of anti-tumor activity, such as inhibition of the Hedgehog pathway with GDC-0449 in patients with basal cell carcinoma (Rudin et al., 2009), or inhibition of the V600E BRAF mutation with PLX4032 in patients with BRAF-mutated melanoma (Flaherty et al., 2010).
Phase 0 Trials and “Window of Opportunity” Trials
Phase 0 trials represent an innovative but controversial attempt to change our traditionally conservative approach to drug development by accelerating the clinical testing of molecularly targeted agents. Phase 0 trials are conducted prior to phase I trials that ordinarily initiate clinical testing. Phase 0 trials in oncology may be conducted in either healthy volunteers or cancer patients. Treatment dose is typically low, and duration of treatment short. The FDA has established guidelines for phase 0 trials (Lorusso et al., 2010; Takimoto, 2009). There are broadly three types of phase 0 trials (Kinders et al., 2007). The first is phase 0 microdose studies of pharmacokinetics or imaging, which are designed to evaluate the pharmacokinetics, metabolism, and/or imaging distribution of specific agents in the absence of any planned pharmacological effect. The second type of phase 0 trials has a pharmacodynamic endpoint. These trials are designed to examine the pharmacological effects of the drug under investigation. The third type of phase 0 trials is designed to evaluate the specific mechanisms of action of a novel agent. These trials are similar to the second type of phase 0 trials, but focus on the target tissue and on an agent’s proposed mechanism of action. These studies have the greatest relevance for the clinical development of molecularly targeted agents.
Phase 0 trials may be difficult to carry out in cancer patient populations as this type of trial obviously lacks any therapeutic intent. Furthermore, many of the molecularly targeted therapies currently under investigation do not have valid pharmacodynamic endpoints that correlate to clinical activity, reflecting perhaps on our limited understanding of “on” and “off” target effects of studied agent. Despite these limitations, phase 0 trials may be valuable in reducing the time and resources required to test promising agents. The early termination of less promising compounds allows more time and resources to be allocated to compounds with a better chance of reaching regulatory approval by having an early go/no-go decision point during the phase 0 trial (Kummar et al., 2009).
“Window of opportunity” trials represent another elegant way to elucidate the mechanism of action of a new drug by using the pre-operative setting (Ratain, 2010). This setting first allows administering a treatment to chemo-naïve patients whose tumors have not developed tumor resistance due to previous therapies. Second, it brings a unique opportunity for translational research from tissue obtained from initial biopsy and from surgical resection specimen. In these trials, the new drug is classically administered for a short period of time, in order not to expose these potentially curable patients to disease progression. The primary endpoint in these trials should be a biological endpoint rather than a clinical activity endpoint that will likely not be met. This kind of trials requires a close collaboration between surgeons, medical oncologists, and research teams.
The classical sequence for developing anti-cancer drugs with phase I, II, and III trials evolved during the era of cytotoxic drugs. However, the recent emergence of molecularly targeted agents challenges this traditional methodology. In case clinical development of a new agent cannot be envisioned in a relevant enriched patient population, phase 0 trials and “window of opportunity” trials represent two ways to accelerate the clinical development of these latter agents by elucidating their mechanism of action. In addition, high levels of antitumor activity observed in small phase II studies might render it unnecessary to perform large randomized phase III trials.
No conflicts of interest to disclose.
Arkenau HT, Barriuso J, Olmos D, Ang JE, de Bono J, Judson I, Kaye S. Prospective validation of a prognostic score to improve patient selection for oncology phase I trials. J Clin Oncol 27(16):2692-2696, 2009.
Booth CM, Calvert AH, Giaccone G, Lobbezoo MW, Seymour LK, Eisenhauer EA. Endpoints and other considerations in phase I studies of targeted anticancer therapy: Recommendations from the task force on Methodology for the Development of Innovative Cancer Therapies (MDICT). Eur J Cancer 44:19-24, 2008.
Chau NG, Florescu A, Chan KK, Wang L, Chen EX, Oza AM, Siu LL. Improving patient selection for oncology phase I trials: Predictors of 90-day mortality and overall survival. J Clin Oncol 28(suppl; abstr 2588):15s, 2010.
Dhani N, Tu D, Sargent DJ, Seymour L, Moore MJ. Alternate endpoints for screening phase II studies. Clin Cancer Res 15(6):1873-1882, 2009.
Di Masi JA, Grabowski HG. Economics of new oncology drug development. J Clin Oncol 25(2):209-216, 2007.
Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, Dancey J, Arbuck S, Gwyther S, Mooney M, Rubinstein L, Shankar L, Dodd L, Kaplan R, Lacombe D, Verweij J. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 45(2):228-247, 2009.
El-Maraghi RH, Eisenhauer EA. Review of phase II trial designs used in studies of molecular targeted agents: outcomes and predictors of success in phase III. J Clin Oncol 26(8):1346-1354, 2008.
Food and Drug Administration. Guidance for industry. Investigators, and reviewers: Exploratory IND studies. US Department of Health and Human Services. http://www.fda.gov/CDER/guidance/7086fnl.htm, accessed on Aug. 15, 2010.
Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, O’Dwyer PJ, Lee RJ, Grippo JF, Nolop K, Chapman PB. Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med 363(9):809-819, 2010.
Gan HK, Grothey A, Pond GR, Moore MJ, Siu LL, Sargent D. Randomized phase II trials: inevitable or inadvisable? J Clin Oncol 28(15):2641-2647, 2010.
Hunsberger S, Zhao Y, Simon R. A comparison of phase II study strategies. Clin Cancer Res 15(19):5950-5955, 2009.
Kinders R, Parchment RE, Doroshow JH, Murgo AJ, Gutierrez M, Collins J, Rubinstein L, Pickeral O, Steinberg SM, Yang S, Hollingshead M, Chen A, Helman L, Wiltrout R, Simpson M, Tomaszewski JE, Doroshow JH. Phase 0 clinical trials in cancer drug development: from FDA guidance to clinical practice. Mol Interv 7(6):325-334, 2007.
Kummar S, Doroshow JH, Tomaszewski JE, Rubinstein L, Parchment RE, Phillips LR, Ji J, Monks A, Low JA, Chen A, Murgo AJ, Collins J, Steinberg SM, Eliopoulos H, Giranda VL, Gordon G, Helman L, Wiltrout R, Tomaszewski JE, Doroshow JH. Phase 0 clinical trials: Recommendations from the task force on methodology for the development of innovative cancer therapies. Eur J Cancer 45(5):741-746, 2009.
Le Tourneau C, Lee JJ, Siu LL. Dose escalation methods in phase I cancer clinical trials. J Natl Cancer Inst 101(10):708-720, 2009.
Le Tourneau C, Diéras V, Tresca P, Cacheux W, Paoletti X. Current challenges for the early clinical development of anticancer drugs in the era of molecularly targeted agents. Targ Oncol 5(1):65-72, 2010.
Le Tourneau C, Stathis A, Vidal L, Moore MJ, Siu LL. Choice of starting dose for molecularly targeted agents evaluated in first-in-human phase I cancer clinical trials. J Clin Oncol 28(8):1401-1407, 2010.
LoRusso PM, Boerner SA, Seymour L. An overview of the optimal planning, design, and conduct of phase I studies of new therapeutics. Clin Cancer Res 16(6):1710-1718, 2010.
O’Quigley J, Pepe M, Fisher L. Continual reassessment method: a practical design for phase I clinical trials in cancer. Biometrics 46:33-48, 1990.
Pharmaceutical Research and Manufacturers of America (PhRMA). Medicines in development for cancer. http://www.phrma.org, accessed on Aug. 15, 2010.
Piantadosi S, Fisher JD, Grossman S. Practical implementation of a modified continual reassessment method for dose-finding trials. Cancer Chemother Pharmacol 41:429-436, 1998.
Ratain MJ, Eisen T, Stadler WM, Flaherty KT, Kaye SB, Rosner GL, Gore M, Desai AA, Patnaik A, Xiong HQ, Rowinsky E, Abbruzzese JL, Xia C, Simantov R, Schwartz B, O’Dwyer PJ. Phase II placebo-controlled discontinuation trial of sorafenib in patients with metastatic renal cell carcinoma. J Clin Oncol 24(16):2505-2512, 2006.
Ratain MJ. Bar the windows but open the door to randomization. J Clin Oncol 28(19):3104-3106, 2010.
Rozencweig M, Von Hoff DD, Staquet MJ, Schein PS, Penta JS, Goldin A, Muggia FM, Freireich EJ, DeVita VT Jr. Animal toxicology for early clinical trials with anticancer agents. Cancer Clin Trials 4:21-28, 1981.
Rubinstein L, Crowley J, Ivy P, Leblanc M, Sargent D. Randomized phase II designs. Clin Cancer Res 15(6):1883-1890, 2009.
Rudin CM, Hann CL, Laterra J, Yauch RL, Callahan CA, Fu L, Holcomb T, Stinson J, Gould SE, Coleman B, LoRusso PM, Von Hoff DD, de Sauvage FJ, Low JA. Treatment of medulloblastoma with hedgehog pathway inhibitor GDC-0449. N Engl J Med 361(12):1173-1178, 2009.
Seymour L, Ivy SP, Sargent D, Spriggs D, Baker L, Rubinstein L, Ratain MJ, Le Blanc M, Stewart D, Crowley J, Groshen S, Humphrey JS, West P, Berry D. The design of phase II clinical trial testing cancer therapeutics: Consensus recommendations from the clinical trial design task force of the National Cancer Institute Investigational Drug Steering Committee. Clin Cancer Res 16(6):1764-1769, 2010.
Takimoto CH. Phase 0 clinical trials in oncology: a paradigm shift for early drug development? Cancer Chemother Pharmacol 63(4):703-709, 2009.
Thall PF. A review of phase 2-3 clinical trial designs. Lifetime Data Anal 14(1):37-53, 2008.
[Discovery Medicine; ISSN: 1539-6509; Discov Med 10(53):355-362, October 2010.]