The primary result of this randomised trial shows a modest survival benefit for marimastat treatment in the overall population of gastric cancer patients. The more interesting result, however, is seen in the sub-group of patients who had received prior chemotherapy. The survival benefit in this group, if true, is clinically significant with an absolute 2-year survival difference of 13%, and a 45% improvement in median survival. There are obvious concerns regarding the interpretation of a result from a sub-group even if, as in this case, it was pre-defined. For this reason the robustness of the result was explored further. Cox regression analysis indicated that imbalances of individual prognostic factors were unlikely to account for the survival difference. A survival benefit in favour of marimastat was also observed in both halves of the chemotherapy sub-group when divided arbitrarily on the basis of recruitment period. Finally, the result is not significantly reduced by the exclusion of the single patient in each arm of the sub-group who entered the study with potentially resectable disease (T1-2, N0-1).
It is important to note that the chemotherapy sub-group is a selected population, including as it does only those patients who had responded to chemotherapy or who had shown stable disease during treatment. The reason these patients responded to marimastat may be related to the fact that they responded in some way to the chemotherapy. Alternatively, the marimastat benefit may be related to the fact that this sub-group excludes those patients with more advanced or more rapidly progressive disease.
The safety and tolerability of marimastat was generally good, with only those events related to the musculoskeletal syndrome being significantly elevated above placebo. The musculoskeletal side effects clearly limit the duration of treatment for the majority of patients. However, the condition can be readily monitored and can be managed by treatment interruptions. In this study 21 placebo patients and 69 marimastat patients interrupted treatment, with a median duration of 14 days for both groups. Experience in this and other studies has shown that severe musculoskeletal side effects can usually be avoided by prompt treatment breaks. It is possible that improved treatment schedules and pre-medications could be employed to allow longer term dosing.
Early expectations for the use of MMP inhibitors in cancer treatment were high, but previous clinical trials with this new class of agent have been disappointing. Marimastat at a dose of 25
mg b.d. showed comparable 1-year survival to gemcitabine in patients with non-metastatic pancreatic cancer (Bramhall et al, 2001
). A follow-on study of 10
mg b.d. marimastat in combination with gemcitabine, however, failed to show any evidence of survival benefit compared to gemcitabine alone. Studies of marimastat in patients with glioblastoma (British Biotech, 2000a
), ovarian cancer (British Biotech, 2000b
) and small cell lung cancer (British Biotech, 2001
) also failed to show evidence of clinical benefit. Randomised, placebo-controlled studies with the MMP inhibitor AG3340 in patients with non-small cell lung cancer and prostatic cancer were stopped early on grounds of lack of efficacy (Agouron, 2000
). Of more concern was the early termination of a randomised placebo-controlled study of the MMP inhibitor BAY12-9566 in patients with small cell lung cancer. Patients receiving BAY12-9566 showed a significantly poorer survival than placebo patients (Hughes, 1999
). Reassuringly, an adverse effect on survival was not observed in the marimastat small cell lung cancer study (British Biotech, 2000c).
Against this background the current results are encouraging. This study of marimastat in patients with gastric cancer provides the first indication of a survival benefit for an MMP inhibitor. The reason that benefit is seen with this cancer and not others may be due to inadequate tissue concentrations of the drug. Biodistribution studies with 14C-marimastat in the rat have shown very high concentrations of marimastat in the stomach wall even after washing. By contrast the levels in lung were low (British Biotech – unpublished data).
In conclusion, these results support a possible role for marimastat as a maintenance treatment following a response or stable disease to chemotherapy. 5-FU based combination chemotherapy is gaining acceptance as the standard of care for gastric and gastro-oesophageal cancer patients with reasonable performance status, both in Europe and North America. In this setting, MMP inhibitor therapy may prove to be a valuable component of the anticancer armoury, and confirmatory trials of this novel agent are warranted.