Maintenance Therapy for First-Line Metastatic Colorectal Cancer: Activity and Sustainability

Abstract

Results of a study, published in this issue of The Oncologist, comparing capecitabine and oxaliplatin plus bevacizumab with the same regimen for only six cycles followed by bevacizumab monotherapy in metastatic colorectal cancer patients are examined.

The role of maintenance therapy for metastatic colorectal cancer is controversial. There are many biological and clinical questions around maintenance therapy and no single study can answer all of them. In addition, there is also likely some degree of semantic confusion around what is meant by maintenance therapy versus simply treatment to progression using whatever components of the initial regimen are tolerable. The MACRO (Maintenance in Colorectal Cancer) study team is to be congratulated on the conduct of the first of many “maintenance” studies that can be expected in the near future. The MACRO study was designed to compare standard capecitabine and oxaliplatin (XELOX) plus bevacizumab, adjusted as appropriate, until disease progression, with the same regimen for only six cycles followed by bevacizumab monotherapy [1]. The authors noted that, although the study did not formally meet its prespecified endpoints for noninferiority, a difference >3 weeks in the median progression-free survival time is unlikely.

Several considerations are worth noting. As is well known to clinicians, first-line chemotherapy for metastatic colorectal cancer is not “sustainable” for prolonged periods of time without treatment modifications or treatment breaks. This results not only from the development of debilitating neuropathy from oxaliplatin but also from mild to moderate nausea, diarrhea, and asthenia, which often become intolerable because of their persistence. This dilemma was the rationale for the OPTIMOX1 and OPTIMOX2 (Optimized Leucovorin [LV]–Fluorouracil [FU]–Oxaliplatin) studies [2, 3], as well as the MACRO study and most other “maintenance” approaches. This is also the reason that many first-line studies suffer from nearly half of all patients stopping protocol-defined therapy before the primary study endpoint of disease progression. Historically, early discontinuation of therapy has been attributed to patient or physician unwillingness to comply with the trial protocol. However, rather than blaming the patient or physician for being “noncompliant,” perhaps we should consider that the protocol itself is not “compliable” for a large fraction of patients, particularly over prolonged periods of time. Most protocols require patients to stay on an every 2 or 3 week schedule indefinitely and discourage repeated delays or even short holidays. Even the MACRO study limited treatment breaks for bevacizumab to 3 weeks, which is only one additional half-life for that drug. During the initial few months of most treatments, many if not most side effects are acceptable, particularly in the setting of gratifying tumor responses. However, by 6 months of treatment, particularly when tumor responses have plateaued, the accumulated burden of toxicities can become intolerable. In addition, there are cumulative social and financial hardships associated with treatment, particularly for patients and families for whom repeated visits to the clinic, admissions to the hospital, and copayments for many supportive care medications can stress limited family resources. When patients take prolonged breaks from scheduled treatments, there can be concerns about compliance. When there is uncertainty about the value of protocol-defined treatment, the default position is to simply stop protocol treatment to allow greater flexibility in managing the whole patient. Like many other first-line studies, the MACRO study had nearly half of all patients stop the protocol-defined treatment before progression, even in the bevacizumab monotherapy group. This suggests that treatment breaks are not just related to classical chemotherapy toxicities but may also be a result of the inconvenience of prolonged treatment itself. The importance of this inconvenience to the patient—and how this barrier to continued treatment is overcome—may depend upon many patient- and physician-specific factors as well as the types of available social support, which may vary considerably within and across countries. There is tension between protocol consistency (so results are interpretable) and protocol flexibility (so the trial is practical). The MACRO study highlights the importance of these issues, particularly for the progression-free survival outcome, and especially for studies in which “maintenance” therapy is a potentially major part of the total treatment algorithm.

Another consideration is that the MACRO study may have been “underpowered” for its primary endpoint. The study was designed to test for a noninferiority hazard ratio of 1.32. This design would pick up large differences, when the progression-free survival interval between the two arms differed by >32%, but it was also not able to detect smaller differences. Whether smaller differences are clinically meaningful is in the eye of the beholder. It should also be noted that the issue of patient dropout may dilute or obscure treatment differences or limit the generalizability of such conclusions. Similarly, although there was not a statistically significant difference in survival between the two study arms, it should be remembered that overall survival can be affected by postprotocol therapies, including the use or reuse of the agents that were in the protocol-defined regimens.

In addition, most patients in the XELOX–bevacizumab arm dropped oxaliplatin after 4 months, consistent with routine care. Thus, the primary comparison of the MACRO study was, to a large degree, capecitabine plus bevacizumab versus bevacizumab alone. An important historical benchmark for the MACRO study is the second-line E3200 study by Giantonio and colleagues, which randomized patients to 5-fluorouracil, leucovorin, and oxaliplatin (FOLFOX) versus FOLFOX plus bevacizumab versus bevacizumab monotherapy [4]. Bevacizumab monotherapy was inferior to both FOLFOX and FOLFOX plus bevacizumab in terms of the response rate and progression-free survival time, and the overall survival duration was longer for patients receiving FOLFOX plus bevacizumab. These findings supported two conclusions at the time: that all drugs are better than a subset of drugs and that there is no role for bevacizumab monotherapy per se. However, the lack of benefit from bevacizumab monotherapy during the “induction” part of second-line treatment does not preclude benefit during “maintenance” therapy, particularly in the first-line setting. The MACRO investigators are to be congratulated for appreciating these important distinctions, and for conducting a clinical trial to formally test the assumptions.

Given the limitations of a single study, the MACRO trial did not address the role of other possible treatment comparators, such as capecitabine alone and observation without treatment. Additional maintenance studies that will provide additional context to the MACRO study are now in progress. The CAIRO-3 (Capecitabine, Irinotecan, and Oxaliplatin in Advanced Colorectal Cancer) study is randomizing patients to maintenance capecitabine plus bevacizumab or observation alone after completing six cycles (4 months) of XELOX plus bevacizumab [5]. The DREAM (Double Inhibition Reintroduction Erlotinib Avastin in Metastatic Colorectal Cancer) study is evaluating the role of maintenance therapy with bevacizumab and erlotinib versus bevacizumab monotherapy after completing six cycles of FOLFOX (or XELOX) plus bevacizumab [6]. At this time, given the limitations of the MACRO study and the lack of other maintenance studies in first-line colorectal cancer patients, it seems most reasonable to continue the components of first-line therapy that are tolerable, including bevacizumab monotherapy, for patients who might not tolerate lower doses of 5-fluorouracil or capecitabine. As the MACRO study illustrates, ensuring this treatment is tolerable and sustainable is an issue for both clinicians and clinical trialists.

See the accompanying article on pages 15–25 of this issue.

Author Contributions

Conception/Design: Herbert I. Hurwitz, John H. Strickler

Provision of study material or patients: Herbert I. Hurwitz, John H. Strickler

Collection and/or assembly of data: Herbert I. Hurwitz, John H. Strickler

Data analysis and interpretation: Herbert I. Hurwitz, John H. Strickler

Manuscript writing: Herbert I. Hurwitz, John H. Strickler

Final approval of manuscript: Herbert I. Hurwitz, John H. Strickler