This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Allegra, C. Articles by Sargent, D. Search for Related Content PubMed PubMed Citation Articles by Allegra, C. Articles by Sargent, D.

End Points in Advanced Colon Cancer Clinical Trials: A Review and Proposal

University of Florida, Gainsville, FL

Charles Blanke

Oregon Health and Science University, Portland, OR

Marc Buyse

International Drug Development Institute, Louvain-La-Neuve, Belgium

Richard Goldberg

University of North Carolina, Chapel Hill, NC

Axel Grothey

Mayo Clinic, Rochester, MN

Neal J. Meropol

Fox Chase Cancer Center, Philadelphia, PA

Leonard Saltz

Memorial Sloan-Kettering Cancer Center, New York, NY

Alan Venook

University of California San Francisco, San Francisco, CA

Greg Yothers

University of Pittsburgh, Pittsburgh, PA

Daniel Sargent

Mayo Clinic, Rochester, MN

Colorectal cancer accounts for the death of more than 55,000 individuals in the United States each year.¹ The vast majority of these patients receive systemic chemotherapy with one or more of several agents presently approved for treatment in this setting. Given the relatively recent and rapid identification of new active therapeutics for patients with advanced colorectal cancer, the algorithm for patient management is undergoing a similarly rapid evolution. The traditional paradigm, dating from the initial use of fluorouracil in the 1960s, dictated that patients receive continuous systemic therapy until disease progression or unacceptable toxicity. This treatment model has generally been applied to therapeutic clinical trials and forms the foundation on which the currently accepted intermediate end points designed to assess the efficacy of new agents (eg, progression-free survival [PFS]) are based.

The high level of activity of newer regimens and a particular element of treatment with the agent oxaliplatin are challenging these standard paradigms. Although highly effective when used in combination with fluorouracil, oxaliplatin is associated with a cumulative sensory neuropathy that occurs at clinically significant levels after approximately 5 months (approximately 850 mg/m² total dose) of therapy.² This adverse event generally precludes long-term continuous therapy with this agent. In other tumor types (eg, lung, ovary), the use of platinum therapy has resulted in preplanned, fixed therapeutic windows of 4 to 6 months. Similar approaches are beginning to be explored in colorectal cancer, driven primarily by the desire to obviate or at least mitigate neurotoxicity. In addition, as three-, four-, and five-drug regimens become more common and a majority of patients respond to treatment (or at least establish stable disease), in many cases, toxicity exceeds tolerable limits before the effectiveness of a regimen is fully exploited. A recent study reported in abstract form may serve as an example of this phenomenon.³ In this investigation, bevacizumab did not seem to improve the PFS of oxaliplatin-based regimens to the extent expected based on prior trials.⁴ A possible explanation for this finding is that a substantial proportion of patients discontinued all treatment because of nonfluoropyrimidine/bevacizumab-related toxicity rather than because of disease progression, and thus, patients may not have had the opportunity to receive the full benefit possible from the combination of chemotherapy with the monoclonal antibody.

A recent series of European studies designed in the prebiologic era have included planned therapeutic interruptions not only for oxaliplatin, but also for irinotecan-based regimens. These studies suggest that certain forms of treatment interruptions/modifications do not seem to significantly lessen the overall efficacy of particular therapeutic interventions, although final outcomes associated with these trials are pending.^5-8 These studies support the utility of intermittent chemotherapy in advanced colorectal cancer and have provided confidence for the development and application of alternative treatment algorithms that include therapeutic holidays.

On the basis of current practice pattern data, many medical oncologists in the United States routinely provide partial or complete drug-free periods for patients with advanced colorectal cancer who have achieved stable disease or best response for at least a several-month interval.⁹ This evolution poses new complexities in the design and conduct of clinical trials, which must be aligned with prevailing clinical practice or the trials risk generating outcome information that is not relevant to current oncology practice.

Although overall survival is the traditional gold standard end point, with the virtues of being unambiguously defined and clearly critically important to the patient, this end point requires extended follow-up and may be confounded by subsequent therapy with any and all available agents, frequently including cross over to the agent under investigation. Intermediate end points are useful in isolating the clinical benefit associated with a specific regimen or strategy and for providing the information in a shorter time frame relative to overall survival. The end point of PFS, which is unencumbered by cross over or subsequent therapy, has recently been accepted by the US Food and Drug Administration as an end point to allow full first-line approval in advanced colorectal cancer. PFS is generally considered to have been reached when a patient experiences disease progression or death. PFS becomes problematic when patients and/or physicians desire the ability to incorporate partial or complete drug holidays into their management strategy because patients will ultimately have disease progression at some point during the holiday period. For these patients, who may eventually constitute the majority of patients in this setting, PFS does not provide an adequate reflection of clinical benefit for a strategy that a priori includes intermittent discontinuation because disease control may be re-established with the reintroduction of the initial regimen. Further supporting this notion is the fact that, in cross-trial comparisons, increases in PFS have not kept pace with recent improvements in overall survival. Prohibiting the use of drug holidays within a clinical trial in an effort to preserve the validity of the PFS end point will result in trial outcomes with little relevance to the practitioner or their patients and may limit trial enrollment by physicians who wish to have more flexibility in patient management.

Given these considerations, we suggest a reconsideration of clinical trial end points in the setting of advanced colorectal cancer, a notion not unique to colorectal cancer or oncology in general. An investigation of the value of atypical antipsychotic drugs in the management of patients with Alzheimer's disease is a recent example of this phenomenon.¹⁰ Within the context of this highly structured, placebo-controlled, randomized trial that tested three different antipsychotic agents, physicians were permitted the flexibility to dose according to clinical judgment and use ancillary medications to acutely manage difficult behavior in an effort to encourage prescribing that reflected clinical practice. The primary end point of this trial was time to discontinuation of therapy for any reason including lack of efficacy or intolerable toxicity, as is the prevailing clinical practice in the management of these patients. In this way, the trial followed the "logic of clinical purpose," a postulate that clinical trials are "logically grounded in and ethically justified by the way in which they reflect and contribute to clinical practice."^11,12 In assessing an end point that reflected current clinical practice patterns, this trial succeeded in providing clinically useful guidance concerning the clinical benefit associated with the use of the atypical antipsychotic agents in the management of patients with Alzheimer's disease.

Because of the need for long-term, potentially toxic and expensive therapy, highly flexible treatment strategies have also been explored for the management of patients with HIV infections. A report from the Strategies for Management of Antiretroviral Therapy Study Group compared the value of traditional continuous antiretroviral therapy with a treatment strategy of episodic therapy where therapeutic interruptions/reinstitutions were guided by CD4⁺ counts or symptoms of HIV infection.¹³ Selection of the specific drug combinations used for each patient was determined by the treating clinician. The primary end points of an opportunistic disease or death were definitively classified as an event by a blinded end point review committee using pre-established criteria. A second investigation, the Flexible Initial Retroviral Suppressive Therapies Trial, randomly assigned 1,397 antiretroviral-naïve patients with HIV to one of three strategies, including doublets of different antiretroviral classes or a combination of all three therapeutic classes.¹⁴ The choice of which specific drugs to use was left to the treating clinician and patient. Physicians were free to discontinue specific drugs for toxicity and substitute an alternative agent of their choice as long as it was within the same therapeutic class. Primary end points included an AIDS-defining event, death, and CD4⁺ count decline to less than a prespecified level. All events were reviewed by an independent committee. These trials clearly demonstrate the feasibility of designing structured clinical trials that can provide clear answers to specific issues while allowing remarkable flexibility in patient management through the use of treatment strategies that are consistent with prevailing clinical practice patterns.

In the setting of advanced colorectal cancer, any clinically relevant new end point must allow and account for strategies that include partial or complete drug holidays and planned discontinuation of selected or all drugs for toxicities or other considerations such as surgical resection of metastases. These issues cannot be resolved through the PFS end point, regardless of the use of various censoring rules. Two standard approaches to censoring for PFS are as follows: (1) censor patients at the cessation of all (or some) study therapy; or (2) do not censor at all, crediting all progression-free time to the regimen under study. Option 2, although perhaps the most clinically relevant option, clearly may credit a regimen with progression-free time attributable to another agent. Option 1 fails to credit a regimen that may allow for an extended holiday (consider the extreme case of a complete response allowing permanent cessation of all therapy). Censoring rules also require unverifiable assumptions regarding the future disease progression profile of censored patients and reduce study power by reducing the number of events in the time-to-event analysis.

Therefore, we propose a new intermediate composite end point primarily designed to capture the full clinical benefit associated with a new agent. The end point may also be applied to the assessment of an extended clinical strategy, under investigation within the context of the highly flexible current practice patterns. We specifically use the term strategy as opposed to regimen because a protocol strategy may reflect any level of complexity, including intermittent use of specific drugs, complete cessation of therapy, or new therapy as part of the protocol at a protocol-defined event. This end point begins with the initiation of the strategy under investigation and ends with the first of the following events: death; patient requires the addition of a new therapeutic agent (ie, an agent not included in the original strategy; end point measured at the first of either the time of disease progression or the time of initiation of new agent); patient experiences disease progression while being treated with all agents that are components of the initial treatment strategy; or patient experiences disease progression during a partial or complete treatment holiday from initial treatment strategy and receives no further therapy within 1 month. This end point allows the full range of acceptable strategies (reducing/eliminating individual agents, maintenance therapy including the use of additional prespecified agents, therapy breaks for toxicity management or otherwise, and reintroduction of therapy on tumor growth after a therapy break) within which to manage patients while remaining on-study, thus providing a more complete assessment of a strategy's benefit. Patients would not be considered to have reached the end point if they experienced progression during a therapy holiday of at least 1 month in duration with subsequent re-establishment of disease control after reinitiation of full treatment per strategy and would be censored only if no new therapy was introduced and the patient did not experience disease progression at the time of analysis. Patients will be considered to have reached the end point only when they require the addition of a non–protocol-specified new agent for further management, have disease progression while on all of the planned agents, experience progression and receive no further therapy within 1 month, or die. Individual protocols should, as specifically as possible, define guidelines for the type and timing for the introduction of a new therapeutic agent as well as the definition and timing of therapy holidays. We propose that the name for this new end point be time to failure of strategy (TFS).

This end point must be distinguished from two other suggested end points, namely time to treatment failure and duration of disease control (DDC⁵; Table 1). Time to treatment failure registers an event at the first of treatment discontinuation, disease progression, or death. This end point is not preferred as a primary end point because it mixes adverse events with efficacy, and a regimen with low activity and low toxicity could theoretically perform better for this end point than a regimen with high activity but high toxicity. TFS does not include cessation of therapy as a result of toxicity as an end point. DDC, defined specifically as PFS or, if therapy is reintroduced, as "the addition of the initial PFS and the PFS of the reintroduction in case of no progression at the first evaluation on FOLFOX reintroduction,"⁵ is closer to TFS. However, DDC, as defined, lacks flexibility as a result of its definition within the context of a very specific protocol with planned intervals for starting and stopping therapy as well as the precise agents to be used at various points throughout the patient's therapeutic course. Conceptually, the definition of DDC could be liberalized, which may improve its general applicability. Furthermore, we feel that defining an event at initiation of a new anticancer therapy not included in the original protocol strategy is an important element of TFS and an aspect that is not captured by DDC. DDC specifically allows the inclusion of additional progression-free time with the reinstitution of the initial therapeutic regimen, potentially, even after exposure to a non–protocol-specified agent during a holiday. In our opinion, in practice, the introduction of a new therapy is an implicit recognition of a shortcoming of current therapy, and time after such an event should not be credited to a strategy. The specific agents that qualify as a next therapy will require specification in each protocol; for example, in a first-line protocol of infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX) plus bevacizumab, the substitution of irinotecan for oxaliplatin, or alternate fluoropyrimidine, may not classify as new therapy, but the introduction of cetuximab or panitumumab may classify as new therapy, depending on the goals and specifications of any given protocol.

Clinical trial methodology must continue to evolve with the clinical practice, which TFS allows. Unless we adopt new approaches to measuring success in clinical trials of new drugs and drug combinations in the setting of advanced colorectal cancer, we run the risk of either designing studies irrelevant to clinical practice and/or being unable to adequately and objectively evaluate new combinations of drugs and strategic approaches. We do not feel that returning to overall survival as a primary end point is feasible because virtually all patients now receive multiple lines of therapy and because median overall survival times now exceed 2 years. We feel that the new end point we propose, TFS, provides a clinically relevant yet rigorously defined end point that minimizes censoring, captures the full effect of a therapeutic strategy, and still allows the isolation of the agent's, regimen's, or strategy's unique contribution to control of disease. This end point eliminates the need for subtle and variable censoring rules for PFS, which have differed greatly among recent trials. We propose that this end point be included in future trials initially as an exploratory end point, which, if validated as useful, could serve in the future as a primary end point.

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

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: Carmen Allegra, Sanofi-Aventis (C), Genentech (C); Richard Goldberg, Sanofi-Aventis (C), Genentech (C), Amgen Inc (C), Pfizer Inc (C), GMS (C), Imclone Systems Inc (C); Axel Grothey, Sanofi-Aventis (C), Genentech (C); Leonard Saltz, Roche (U), Genentech (U), Sanofi-Aventis (U); Alan Venook, Amgen Inc (C), Pfizer Inc (C), Bayer (C), Lilly (C), YM Bioscience (C), Sanofi-Aventis (C); Greg Yothers, AstraZeneca (C), Amgen Inc (C), Pfizer Inc (C), Genentech (C), Novartis (C), GlaxoSmithKline (C), Bristol-Myers Squibb Co (C) Stock Ownership: None Honoraria: Richard Goldberg, Imclone Systems Inc; Axel Grothey, Sanofi-Aventis, Genentech; Leonard Saltz, Roche, Genentech, Sanofi-Aventis Research Funding: Neal J. Meropol, Genentech, Pfizer Inc, Novartis, Amgen Inc, GlaxoSmithKline; Leonard Saltz, Sanofi-Aventis; Alan Venook, Imclone Systems Inc, Bristol-Myers Squibb Co, Genentech, Pfizer Inc, Taiho, Roche; Greg Yothers, Genentech, Pfizer Inc Expert Testimony: Richard Goldberg, Imclone Systems Inc (C) Other Remuneration: None

AUTHOR CONTRIBUTIONS

Conception and design: Carmen Allegra, Charles Blanke, Richard Goldberg, Axel Grothey, Neal J. Meropol, Leonard Saltz, Alan Venook, Greg Yothers, Daniel Sargent

Administrative support: Carmen Allegra, Daniel Sargent

Collection and assembly of data: Carmen Allegra, Richard Goldberg, Neal J. Meropol, Greg Yothers, Daniel Sargent

Data analysis and interpretation: Carmen Allegra, Charles Blanke, Marc Buyse, Richard Goldberg, Alan Venook, Greg Yothers, Daniel Sargent

Manuscript writing: Carmen Allegra, Charles Blanke, Richard Goldberg, Axel Grothey, Neal J. Meropol, Leonard Saltz, Alan Venook, Greg Yothers, Daniel Sargent

Final approval of manuscript: Carmen Allegra, Charles Blanke, Marc Buyse, Richard Goldberg, Axel Grothey, Neal J. Meropol, Leonard Saltz, Alan Venook, Greg Yothers, Daniel Sargent

REFERENCES

1. National Center for Health Statistics, Centers for Disease Control and Prevention: U.S. mortality public use data tapes, 1969-2003. http://seer.cancer.gov

2. Goldberg RM, Sargent DJ, Morton RF, et al: Randomized controlled trial of reduced-dose bolus fluorouracil plus leucovorin and irinotecan or infused fluorouracil plus leucovorin and oxaliplatin in patients with previously untreated metastatic colorectal cancer: A North American Intergroup Trial. J Clin Oncol 24:3347-3353, 2006[Abstract/Free Full Text]

3. Saltz LB, Clarke S, Diaz-Rubio E, et al: Bevacizumab in combination with XELOX or FOLFOX4: Efficacy results from XELOX-1/NO16966, a randomized phase III trial in the first-line treatment of metastatic colorectal cancer. J Clin Oncol 25:204s, 2007 (suppl; abstr 238)

4. Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004[Abstract/Free Full Text]

5. Tournigand C, Cervantes A, Figer A, et al: OPTIMOX1: A randomized study of FOLFOX4 or FOLFOX7 with oxaliplatin in a stop-and-go fashion in advanced colorectal cancer—A GERCOR study. J Clin Oncol 24:394-400, 2006[Abstract/Free Full Text]

6. Maindrault-Goebel F, Lledo G, Chibaudel L, et al: OPTIMOX2, a large randomized phase II study of maintenance therapy or chemotherapy-free intervals after FOLFOX in patients with metastatic colorectal cancer: A GERCOR study. J Clin Oncol 24:147s, 2006 (suppl, abstr 3504)

7. Labianca R, Floriani E, Cortesi L, et al: Alternating versus continuous "FOLFIRI" in advanced colorectal cancer: A randomized GISCAD trial. J Clin Oncol 24:147s, 2006 (suppl, abstr 3505)

8. Maughn TS, James RD, Kerr DJ: Comparison of intermittent and continuous palliative chemotherapy for advanced colorectal cancer: A multicentre randomized trial. Lancet 361:457-464, 2003[CrossRef][Medline]

9. Denlinger CS, Collins MA, Wong YN, et al: Metastatic colorectal cancer patterns of care: Implications for clinical trial design. J Clin Oncol 25:225s, 2007 (suppl; abstr 281)

10. Schneider L, Tariot P, Dagerman K, et al: Effectiveness of atypical antipsychotic drugs in patients with Alzheimer's disease. N Engl J Med 355:1525-1538, 2006[Abstract/Free Full Text]

11. Karlawish J: Alzheimer's disease: Clinical trials and the logic of clinical purpose. N Engl J Med 355:1604-1606, 2006[Free Full Text]

12. Freedman B: Placebo-controlled trials and the logic of clinical purpose. IRB 12:1-6, 1990[Medline]

13. Strategies for Management of Antiretroviral Therapy (SMART) Study Group, El-Sadr WM, Lundgren JD, et al: CD4+ count-guided interruption of antiretroviral treatment. N Engl J Med 355:2283-2296, 2006[Abstract/Free Full Text]

14. MacArthur RD, Novak RM, Peng G, et al: A comparison of three highly active antiretroviral treatment strategies consisting of non-nucleoside reverse transcriptase inhibitors, protease inhibitors, or both in the presence of nucleoside reverse transcriptase inhibitors as initial therapy (CPCRA 058 FIRST Study): A long-term randomized trial: Lancet 368:2125-2135, 2006[CrossRef][Medline]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Allegra, C. Articles by Sargent, D. Search for Related Content PubMed PubMed Citation Articles by Allegra, C. Articles by Sargent, D.