Back to the drawing board: Melanoma vaccine ineffective, possibly harmful; GM2 tanks in major study

Key Points

Chicago - For patients with stage II melanoma, the adjuvant ganglioside GM2-KLH21 vaccine proved ineffective, and possibly even harmful, in a large multicenter European study (EORTC 18961) presented at the American Society of Clinical Oncology annual meeting.

The presentation of the EORTC 18961 findings by Alexander Eggermont, M.D., followed the announcement of negative results in the Sunbelt Melanoma Trial, which found no benefit for high-dose interferon in node-positive patients.

In reference to his predecessor, Dr. Eggermont began, "After the presentation of negative results, I now have the dubious honor to present significantly negative results."

Hopes had been pinned on GM2-KLH21 vaccine, since four weekly injections induce IgM and IgG responses in at least 80 percent of patients, as well as antibody-dependent cellular cytotoxicity. Previous studies had suggested that responding patients would have improved outcomes.

The phase 3 EORTC study included 1,314 patients with resected stage II melanoma and primary tumors thicker than 1.5 mm. Stratification was done according to sentinel node status (negative versus unstaged) and by primary tumor factors.

Patients were randomized to observation only or to the GM2 vaccine injected once weekly for four weeks, then every three months for two years, then every six months for a third year, for a total of 14 vaccinations.

Disease-free survival

The arms were equivalent in terms of disease-free survival, which at two years was 78 percent in the observation arm and 75 percent in the vaccinated arm.

"This was a perfectly overlapping curve that met the definition of futility, which was our primary endpoint, and gave a reason to stop the trial," he says.

By sentinel node status, non-staged patients experienced more relapses, mostly in the vaccine arm; however, the difference between the arms was not significant.

For distant metastasis-free survival in the intent-to-treat population, some differences emerged, again to the detriment of the vaccine. At two years, no distant metastases were observed in 88 percent of the observation arm and in 82 percent of the vaccinated arm.

"We see a trend emerging that is disquieting. We see the vaccine arm separating out, with 20 additional events in the vaccine arm, for a 33 percent increase (P = 0.08)," Dr. Eggermont says.

For overall survival, the difference became greater, with a 57 percent increase in mortality among vaccinated patients (P = 0.032), based on 35 deaths in the observation arm and 56 in the vaccinated arm.

"Obviously, this is not very reassuring, to say the least," he says.

"Therefore, the data monitoring committee decided not just to stop the trial but to stop the additional vaccinations scheduled to be given.

"The conclusion of the investigators was that the GM2 vaccine is not only ineffective, but may even be detrimental in stage II melanoma," Dr. Eggermont says.

Jeffrey S. Weber, M.D., head of the Melanoma Center of Excellence at H. Lee Moffitt Cancer Center, Tampa, Fla., says that the findings mirror the large trial in stage III melanoma patients with the similar Canvaxin (Serano) vaccine in stage III and IV patients, whose outcomes were worse than control patients receiving bacillus Calmette-Guerin.

In the large study of stage III patients (n = 1160, the vaccine was significantly inferior in terms of disease-free and overall survival (P = 0.04), he says.

"This raises significant concerns about vaccines. We may be raising the level of suppressor cells by vaccinating patients in the absence of other immunomodulation," Dr. Weber tells Dermatology Times.

Future studies are needed to detect and exclude induction of tolerance, to develop vaccines for stage IV patients, to evaluate and incorporate new means of immunomonitoring and to evaluate the consequences of trial design in all stages of patients.