Georgina Long, MBBS, PhD
Long-term follow-up from the COMBI-AD trial continued to show a significant relapse-free survival (RFS) advantage for patients with resected stage III BRAF
-mutant melanoma treated with adjuvant dabrafenib (Tafinlar) plus trametinib (Mekinist).
Patients randomized to adjuvant dabrafenib plus trametinib had a 4-year relapse-free survival (RFS) rate of 54%, as compared with 38% for placebo-treated patients. A cure-rate analysis for RFS produced estimates of 54% for dabrafenib/trametinib and 37% for placebo, representing the proportion of patients expected to achieve long-term RFS, as reported at the 2018 ESMO Congress.
A biomarker analysis showed that MAP kinase gene alterations were not associated with outcome in the adjuvant setting but did suggest that a high tumor mutational burden (TMB) in association with high levels of interferon-gamma might be predictive of greater benefit.
“Longer-term follow-up in the COMBI-AD trial confirmed the relapse-free survival benefit with adjuvant dabrafenib and trametinib in patients with resected stage III BRAF
-mutant melanoma,” said Georgina Long, MBBS, PhD, of the University of Sydney in Australia. “As yet, we have no biomarker to select treatment; therefore, further investigation is warranted.”
Two randomized, phase III clinical trials, COMBI- and COMBI-v, showed that treatment with dabrafenib + trametinib improved overall survival in patients with unresectable or metastatic melanoma associated with BRAF
mutations. Whether the therapy improved survival in stage III (resectable) melanoma remained unknown, providing the rationale for the international COMBI-AD trial.
Investigators in 26 countries randomized 870 patients with resected, BRAF
-mutant melanoma to receive dabrafenib + trametinib or placebo. The trial had a primary endpoint of RFS. As reported last year in the New England Journal of Medicine
, the primary analysis showed a 3-year RFS of 58% with the combination therapy and 39% with placebo (HR 0.47, 95% CI 0.39-0.58, P
Updated and new analyses were performed to characterize more precisely the longer-term benefit of adjuvant dabrafenib + trametinib. The analysis included RFS and distant metastasis-free survival (DMFS) with extended follow-up; cure-rate modeling to estimate the percentage of patients expected to remain relapse-free long term; and a biomarker analysis involving baseline tumor DNA and RNA and gene expression profiling.
The analysis occurred after a median follow-up of 44 months, 11 months beyond the median follow-up at the time of the primary analysis.
“This was not a prespecified statistical analysis but was mandated by the EMA [European Medicines Agency],” said Long.
With regard to RFS, at the time of the primary analysis, relapses had occurred in 37% of the combination arm and 57% of the placebo group. The updated analysis showed relapse rates of 40% and 59% for the combination and placebo arms, respectively.
The updated analysis also showed that distant relapses accounted for 59% of all relapses in the combination arm, and locoregional relapses for 33%, similar to rates in the primary analysis. New primary melanoma and concurrent local/distant relapses accounted for 5% each. Distribution of relapse type in the placebo arm was 51% distant, 43% locoregional, 3% new melanomas, and 2% concurrent local/distant relapse.
Landmark analysis of RFS showed substantial absolute differences from the beginning of follow-up in favor of the dabrafenib + trametinib arm: 88% vs 56% at 1 year, 67% vs 44% at 2 years, 59% vs 40% at 3 years, and 54% vs 38% at 4 years. The hazard ratio for 4-year RFS was 0.49 (95% CI 0.40-0.59) DMFS for years 1-4 was 91%, 77%, 71%, and 67% for the combination arm and 70%, 60%, 57%, and 56% in the placebo arm. The hazard ratio after 4 years was 0.53 (95% CI 042-0.67).
Long described cure-rate modeling as a standard statistical technique that examines event rates in a preceding time period to predict future event rates. The analysis produced estimated cure rates, which demonstrated a statistically significant 17% absolute difference in favor of the combination arm.
The biomarker analysis included DNA samples for 368 study participants and RNA samples for 507 participants; 301 study participants had paired results. The genomic landscape at baseline proved to be similar to published data related to metastatic melanoma.
Investigators conducted additional analyses of genetic alterations in the MAP kinase pathway. BRAF
amplification and MEK1
mutation occurred in a small number of patients, but neither were associated with RFS. A combined analysis of all samples that had an aberration in the MAP kinase pathway also showed no association with RFS, similar to observations in the metastatic melanoma setting, said Long.