Commentary (Pritchard): Current Status of Vinorelbine For Breast Cancer

Dr. Smith's excellent article summarizes the preclinical and clinical data obtained to date on the role of vinorelbine (Navelbine) in women with breast cancer. Introducing a new agent into the arena of breast cancer is fraught with logistic difficulties. Once an agent appears relatively safe in phase I studies and relatively active in small phase II trials, it is nonetheless difficult to decide which route to pursue in terms of further development. Larger phase II studies using the drug as a single agent or in combination are useful in establishing more exact confidence intervals around the response rate and duration, but they do not provide the information necessary to "place" the new agent in the complex algorithm of breast cancer therapy.

Phase III Trial Design Problematic

Phase III (randomized, comparative) trials are far more useful, but the strategy to be used in their design is problematic. Should the new agent be studied in combination, or as a single agent? Should it be compared to the "best" single agent, which most would agree is an anthracycline (doxorubicin or epidoxorubicin [Epirubicin]), or should it first be tested against a second-line drug or combination? Should it be tested initially in first- or second-line therapy for metastatic disease? When can a new agent be moved into the arena of adjuvant therapy?

Answers to these questions are not obvious, but it seems that the sooner a drug can be moved into randomized testing, the more rapidly its position can be established. Patient selection, particularly in breast cancer, can so strongly influence outcome that intertrial comparison of phase II data is perilous. Once it is clear that the new agent is relatively active and safe, randomized trials should begin.

In a previous era, combination chemotherapy for breast cancer was so firmly established that testing single agents for first-line therapy of metastatic disease was felt to be almost unethical. Now, however, it is increasingly accepted that the best agents, used alone in optimal doses, may be as active as and no more toxic than combinations of several drugs. Thus, testing a new agent used alone vs a standard agent used alone, with the same criteria for dose de-escalation and for use of colony-stimulating factors, prophylactic antibiotics, and other adjuvant agents, has become an ideal trial design.

Which Outcome Measures?

And what should be the outcome measures? Response rate and duration, while standard, convey only limited information. A'Hern et al have suggested in an overview that a certain increase in response rate will usually translate into a certain survival benefit (for example, an increase in response rate from 20% to 30% will translate into a median survival increase from 18 to 21.4 months) [1], but this concept is far from universally accepted. Similarly, there is some acceptance of the concept that in symptomatic patients with metastatic disease, complete or even partial responses will almost certainly result in improvement in at least some aspects of quality of life. Thus, unless severe toxicity is produced by the regimen in question, an overall benefit relative to quality of life should be seen.

Of course, reliable and valid measures of quality of life in the metastatic setting are now widely available and should probably be made standard outcome measures in trials of metastatic disease, since an improvement in quality of life is the most probable and desired outcome in this setting. Survival, the most important outcome, should be measured, but in the metastatic setting limited responses as well as contamination by subsequent therapy have generally made improvement in this outcome measure difficult to obtain or to demonstrate. Furthermore, if improvements in survival are small, or if one wishes to definitively rule out any difference, large sample sizes will be required in order to produce adequate statistical power.

Where Does This Leave Vinorelbine?

Preclinical studies have shown a broad spectrum of activity for vinorelbine and have suggested that it might be substantially less toxic than the standard vinca alkaloids (vincristine and vinblastine). Phase I/II trials have shown the efficacy of vinorelbine against a variety of tumors and have demonstrated a very acceptable toxicity profile. Neutropenia is the dose-limiting toxicity, accompanied by frequent mild anemia and a rare (< 10%) incidence of thrombocytopenia below 100,000/mm³. Nonhematologic toxicity is usually fairly minor, consisting of mild elevations of liver function tests; mild nausea, with some vomiting (23%), constipation (38%), and diarrhea (20%); mild, reversible peripheral neuropathy (31%); reversible dyspnea for a few hours (9%); and alopecia (12%). Vinorelbine also is a mild vesicant.

There are now a number of fairly large phase II trials of vinorelbine available, perhaps too many and too large. However, these trials demonstrate single-agent response rates of 40% to 44% in first-line treatment of metastatic breast cancer and rates of 17% to 36% in second-line treatment, results that appear to be very similar to those seen with doxorubicin or paclitaxel. Trials of combination therapy have shown response rates of 54% to 74% with vinorelbine plus doxorubicin, 67% with vinorelbine plus fluorouracil infusion, and 56% with vinorelbine and mitoxantrone (Novantrone), all in first-line studies; a response rate of 35% with vinorelbine and mitomycin (Mutamycin) was obtained in one second-line study.

Response durations with single-agent therapy range from 17 to 38 weeks, while combination therapy has produced response durations of 36 to 52 weeks. Overall survival durations range from 20 to 27.5 months in first-line trials to 8.8 months in one second-line study.

The Need for Comparisons

Exact comparisons with more standard agents remain pivotal, however. One trial of vinorelbine vs melphalan (Alkeran) as second- or third-line therapy for metastatic disease demonstrated that vinorelbine produced a superior time to treatment failure (13 vs 8 weeks), longer time to disease progression (12 vs 8 weeks), higher response rate (16% vs 9%), and longer survival (35 vs 31 weeks) in a heavily pretreated group of women. The second/third-line setting and the use of a less than current standard as a control arm limit the value of this trial. Rather, the phase III trials still underway are more likely to place this drug more clearly in the therapeutic hierarchy. For example, an active study of single-agent doxorubicin vs doxorubicin plus vinorelbine in first- or second-line metastatic therapy has just been completed and will be analyzed short-ly (personal communication, Brian Norris, MD). A third arm with vinorelbine alone would have been a useful addition to this study.

In the meantime, the reasonable activity and minimal nonhematologic toxicity of vinorelbine recommend its use in older or debilitated women as second- or third-line therapy, or in those who wish a less toxic but relatively active chemotherapy alternative. The low proportion of patients suffering alopecia will recommend this drug to women for whom that side effect is particularly unacceptable. One disadvantage is the weekly injection schedule, but the proposed availability of oral vinorelbine would avoid this problem, making this drug an extremely attractive alternative for palliation. Whether it will emerge as a first or second choice for front-line therapy, or in the adjuvant setting, however, awaits the results of further randomized trials.

References:

1. A'Hern RP, Ebbs SR, Baum MD: Does chemotherapy improve survival in advanced breast cancer? A statistical overview. Br J Cancer 57:615-618, 1988.