Background: there are no active treatment options available for patients (pts) diagnosed with melanoma brain metastases (MBM) that progress despite treatment with immune checkpoint inhibitors (ICI), and BRAF-/MEK-inhibitors (BRAF/MEKi) in pts with BRAF V600mutant melanoma. Regorafenib (REGO), an oral multi-target kinase inhibitor, including potent inhibition of RAF-dimers, has single-agent activity in pretreated melanoma (AS Vander Mijnsbrugge et al. SMR 2022). Here we report our single-institution experience with REGO in combination with BRAF/MEKi in pts with refractory MBM. Methods: pts were identified prospectively, medical records were reviewed retrospectively. Results: 13 pts with stage IV-M1d melanoma were included (6F, med age 54y [range 33-72], WHO PS: 0/1/2/3 resp. n = 1/4/6/2 pts, BRAF V600mut: n = 9, NRAS Q61mut: n = 4). Pts previously progressed on ICI (anti-PD-1 and -CTLA-4; 13 pts); BRAF-/MEK-inhibitor (all BRAF V600mut pts); chemotherapy (3 pts), T-VEC (2 pts), REGO mono (2 pts), and REGO+ICI (2 pts). At baseline, 11 pts had active MBM (8 pts on corticosteroids); 10 pts had extracranial evaluable disease. Pts were treated with REGO (40-80 mg QD) combined with BRAF/MEKi in 9 BRAF V600mut pts (dabrafenib/trametinib: n = 4; encorafenib/binimetinib: n = 5) or, in 4 NRAS Q61mut pts, with trametinib or binimetinib (+ low-dose BRAF-inhibitor if needed to mitigate skin toxicity). Treatment was stopped in 8 pts because of PD and is ongoing in 5 pts (3-49 wks after initiating therapy). There were no grade 4-5 TRAE. Grade 3 TRAE included: arterial hypertension (n = 4), hepatotoxicity (n = 2), diverticular bleeding (n = 1), and duodenal perforation, treated conservatively (n = 1). None of the 2 pts without active MBM at baseline progressed within the CNS. In 11 response evaluable pts (incl. 10 with evaluable extracranial metastases), best objective intracranial response (according to RANO-BM) was PR in 4 pts (36%; 3 BRAF V600mut pts), and SD in 4 pts (36%; 3 BRAF V600mut pts). Intracranial responses corresponded with extracranial responses in all but 3 pts (1 NRAS Q61mut pt with a PR intra-, had SD extracranially; 1 BRAF V600mut pt and 1 NRAS Q61mut pt with SD intra-, had PD extracranially); considering a potential clinical benefit of therapy beyond first PD, both pts continued treatment and remained clinically stable at the latest FU (resp. 15 and 49 wks after initiating therapy). Median time on REGO+BRAF/MEKi in BRAF V600mut pts was 15.9 wks (range 3-46), and on REGO+MEKi in NRAS Q61mut pts: 7.6 wks (range 3-49). Median PFS and OS: 12.9 and 26.4 wks (resp. 15.3 and 26.4 wks in BRAF V600mut pts; 8.6 and 10.1 wks in NRAS Q61mut pts). Conclusions: in heavily pretreated patients with MBM combination of REGO with BRAF-/MEK-inhibitors demonstrated promising anti-tumor activity. Further investigation of safety and clinical activity in a prospective trial is warranted.