The investment in therapeutic oncology studies in North America is enormous. Every year, approximately 20,000 potential anticancer agents are screened by the United States National Cancer Institute (NCI) in vitro human tumor cell line assay, and in 1999 approximately 30 new agents were taken to clinical trial.<sup>22 Klausner R.D.   The Nation's Investment in Cancer Research Bethesda, MD: National Cancer Institute (1999).  NIH Publication No. 99-4373..

Cliquez ici pour aller à la section Références, 37 Rubinstein L.V., Ungerleider R. Cooperative cancer trials Encyclopedia of Biostatistics Chichester, UK: John Wiley & Sons (1998).  933

Cliquez ici pour aller à la section Références</sup> In North America there are currently 10 government-funded cancer cooperative clinical trials groups (American College of Surgeons Oncology Group, Cancer and Leukemia Group B, Children's Oncology Group, Eastern Cooperative Oncology Group, Gynecologic Oncology Group, National Cancer Institute of Canada, National Surgical Adjuvant Breast and Bowel Project, North Central Cancer Treatment Group, Radiation Therapy Oncology Group, and Southwest Oncology Group) and approximately 60 cancer centers, primarily large academic medical centers. Among them, they represent approximately 8000 investigators at 1700 institutions and place more than 20,000 patients on therapeutic studies annually.<sup>22 Klausner R.D.   The Nation's Investment in Cancer Research Bethesda, MD: National Cancer Institute (1999).  NIH Publication No. 99-4373..

Cliquez ici pour aller à la section Références, 37 Rubinstein L.V., Ungerleider R. Cooperative cancer trials Encyclopedia of Biostatistics Chichester, UK: John Wiley & Sons (1998).  933

Cliquez ici pour aller à la section Références</sup>

By 1960, the current clinical pathway for the development of a potential anticancer agent or regimen had been defined<sup>14 Gehan E.A., Schneiderman M.A. Historical and methodological developments in clinical trials at the National Cancer Institute Stat Med 1990 ;  9 : 871

Cliquez ici pour aller à la section Références</sup>: phase I trials of approximately 20 patients to determine the appropriate dose and schedule for further testing, usually the maximal tolerated dose (MTD); followed by phase II trials of approximately 30 to 50 patients to establish an indication of clinical effect, usually tumor shrinkage; followed by randomized phase III trials involving hundreds or thousands of patients to test clinical efficacy, usually defined as the ability to prolong survival, compared with a control therapy. Since the early 1960s, the cooperative groups have accrued patients according to standardized clinical protocols, with standardized criteria of diagnosis, treatment, and measurement of effect and with a prospective statistical design and collaborative analysis and reporting. The early development of the cooperative group statistical centers was led by NCI statistician Marvin Schneiderman, <sup>14 Gehan E.A., Schneiderman M.A. Historical and methodological developments in clinical trials at the National Cancer Institute Stat Med 1990 ;  9 : 871

Cliquez ici pour aller à la section Références</sup> and since then statisticians from the National Institutes of Health (NIH) and from cooperative group and cancer center statistical centers have developed most of the new statistical methodology for the design, conduct, and analysis of cancer clinical studies.<sup>37 Rubinstein L.V., Ungerleider R. Cooperative cancer trials Encyclopedia of Biostatistics Chichester, UK: John Wiley & Sons (1998).  933

Cliquez ici pour aller à la section Références, 41 Simon R. A decade of progress in statistical methodology for clinical trials Stat Med 1991 ;  10 : 1789

Cliquez ici pour aller à la section Références</sup>

A significant body of literature reviews clinical trials methodology. For cancer trials in particular, possibly the most inclusive recent work is that of Piantadosi.<sup>34 Piantadosi S.   Clinical Trials: A Methodologic Perspective New York: John Wiley & Sons (1997). 

Cliquez ici pour aller à la section Références</sup> Two shorter, but still comprehensive, studies are that by Leventhal and Wittes, <sup>26 Leventhal B.G., Wittes R.E.   Research Methods in Clinical Oncology New York: Raven Press (1988). 

Cliquez ici pour aller à la section Références</sup> which gives a clinician's perspective, and that of Green et al, <sup>16 Green S., Benedetti J., Crowley J.   Clinical Trials in Oncology London: Chapman & Hall (1997). 

Cliquez ici pour aller à la section Références</sup> which gives a statistician's perspective. Simon<sup>43 Simon R. Clinical trials in cancer Cancer: Principles and Practice of Oncology, ed 5 Philadelphia: Lippincott-Raven (1997).  513

Cliquez ici pour aller à la section Références</sup> has written a useful chapter in the text by DeVita et al and has also written a review of clinical trials methodology<sup>41 Simon R. A decade of progress in statistical methodology for clinical trials Stat Med 1991 ;  10 : 1789

Cliquez ici pour aller à la section Références</sup> in the active decade of the 1980s. The author has drawn on these sources liberally in this article and has also referred to original methodology papers. This article attempts, as much as possible, to keep the discussion on an intuitive and nontechnical level. It elucidates the problems in basic design, conduct, and interpretation associated with phase I, phase II, and phase III trials and explains how the various statistical approaches have arisen as solutions to these problems. The fundamental problem common to all three trial types is that of achieving a correct and precise answer to the question posed by the trial, to inform future testing and treatment better, while protecting the trial patients from receiving treatment that has demonstrated excessive toxicity or lack of clinical efficacy. This shared problem gives rise to statistical designs with basic similarities across the three trial types.