Background: Combination of vemurafenib plus cobimetinib is approved for the treatment of BRAF-mutated advanced melanoma. Although patients initially respond to treatment, resistance emerges before 18 months in most cases. One of the key pre-clinical observations that supported an intermittent schedule was that resistant tumors suffer a fitness deficit in the absence of the drug, so modulation of the drug pressure through an intermittent dosing could delay the emergence of resistance. Methods: GEM1501 is a randomized phase 2 study comparing the activity of the combination of vemurafenib 960 mg every 12 h/d plus cobimetinib 60 mg/d in a standard (arm A) versus intermittent schedule (arm B). Arm A: four-week (w) cycles of daily vemurafenib for 4w plus cobimetinib for 3w-on and 1w-off-treatment. Arm B: first three cycles according to the standard schedule, followed by 6w-cycle with 2w-off vemurafenib & 3w-off cobimetinib. Primary endpoint was progression free survival (PFS) and secondary were objective response (OR) and treatment-related adverse events (TAEs). Results: 70 treatment-naïve patients were included. Results in arms A and B: median PFS 16.2 (95%CI 9.5, 24.1) vs 6.9 months (95%CI 5.2, 9.3) (p = 0.079); OR in 25 (71.4%) (8 complete -23%-) vs 21 (60%) patients (5 complete -14%-); G3-4 TAEs 42.8% vs 40.0%, respectively. Analysis of BRAF^V600 mutation in tumoral cell free DNA (cfDNA) was performed in serial plasma samples in 41 patients. Twenty-one (51%) patients had detectable BRAF^V600 mutation in pretreatment cfDNA (preBRAF+). Significant differences in PFS were found according to preBRAF^V600: 8.2 months (95%CI 5.2, 13.6) in preBRAF+ vs non-reached (NR) (95%CI 2.8, NR) in preBRAF- (p = 0.017). In arm A, median PFS was 13.3 months (95% CI 4.6, NR) in preBRAF+ vs NR (95% CI 2.3, NR) in preBRAF-. In arm B, median PFS was 6.2 months (95% CI 0.3-8.3) in preBRAF+ vs NR (95%CI 2.8, NR) in preBRAF- (p = 0.003).BRAF^V600 mutation became undetectable in cfDNA after treatment initiation in all preBRAF+ patients. Different kinetic of BRAF^V600 mutation in cfDNA was found according to treatment arm. At progression, BRAF^V600 reappeared in cfDNA in all (5/5) cases treated in arm B, but only in 50% (3/6) of cases in arm A. NGS analysis of cfDNA at progression suggested different resistance mechanisms. Conclusions: The results of this study do not support the use of an intermittent schedule of vemurafenib plus cobimetinib in advanced melanoma. BRAF^V600 detection in pretreatment cfDNA is a prognostic factor of poor survival that it is independent of treatment schedule, although most striking differences favoring continuous arm vs intermittent arm were found in patients with detectable BRAF^V600 mutation on pretreatment cfDNA. Further research is required to determine the clinical value of the analysis of resistance mechanisms in cfDNA. Clinical trial information: 2014-005277-36.