A revised report of findings from the phase II TRICOTEL study of atezolizumab, vemurafenib, and cobimetinib in patients with melanoma and central nervous system (CNS) metastases was published in The Lancet Oncology by Dummer et al. The first version of trial findings, published online in August 2022, was retracted. In the modified report, the study continued to show that the triplet demonstrated intracranial activity in this patient population.
A total of 65 patients were enrolled in a BRAF V600 mutation–positive cohort; a BRAF V600 wild-type cohort was closed early after enrollment of 15 patients (following primary analysis of the phase III IMspire170 study of cobimetinib plus atezolizumab in BRAF V600 wild-type melanoma). All patients had previously untreated metastatic melanoma and brain metastases of ≥ 5 mm in at least one dimension. Median follow-up was 9.7 months in the BRAF V600 mutation–positive cohort. Intracranial objective response was observed in 27 patients (42%) by independent review committee assessment, including complete response in 4 (6%). Median duration of response was 7.4 months. In post hoc analysis including 26 patients with symptomatic CNS metastases, intracranial objective response was observed in 9 (35%). Median intracranial progression-free survival was 5.3 months. Median follow-up was 6.2 months in the BRAF V600 wild-type cohort. Intracranial objective response on investigator assessment was observed in four patients (27%). Median intracranial progression-free survival was 2.2 months.
As stated by the editors of The Lancet Oncology, “[We] republish a corrected version along with a revised appendix, in which the findings have changed—for example, the intracranial response rate for patients with symptomatic CNS metastases, a post-hoc subgroup analysis, is lower than previously reported and thus the combination of atezolizumab, vemurafenib, and cobimetinib no longer appears to provide substantially improved efficacy over targeted therapy alone in this subset of patients. Of note, however, the primary endpoint has not changed and the overall message of the paper is broadly similar to the previous version.”