Background: Mucosal melanoma is a rare malignant melanoma in Caucasians but ranks the second most common subtype in the Asian population. It is more often diagnosed at an advanced stage and responds poorly to current PD-1/PD-L1 inhibitors. Here we report the interim analysis results of ML41186, an open-label, multicenter, single-arm phase II study, aiming to evaluate the efficacy and safety of atezolizumab in combination with bevacizumab in patients (pts) with advanced mucosal melanoma. Methods: Eligible pts aged 18 to 75 years with histologically confirmed unresectable locally advanced or metastatic mucosal melanoma had at least one measurable lesion per RECIST version 1.1 at baseline, with an ECOG PS 0 or 1 and adequate hematologic and organ function. ML41186 is a Simon two-stage design study, if 22 pts completed ORR evaluation and more than 3 pts respond in stage I, the study then continue to Stage II. Atezolizumab and bevacizumab were administered at a fixed dose of 1200 mg and 7.5 mg/kg Q3W respectively (on day 1 of each 21-day cycle) until unacceptable toxicity or loss of clinical benefit. The primary endpoint is the objective response rate (ORR). The secondary endpoints include progression-free survival (PFS), duration of objective response (DoR), disease control rate (DCR), and safety. Results: By the cut-off date of 9^th September 2020, 35 pts has been enrolled, among whom 22 pts in the stage I analysis set has completed two efficacy evaluation, while 28 pts (full analysis set) has completed at least one efficacy evaluation. In ITT populations (n=35), mean age was 58.9 years with 10 (28%) pts had ECOG PS of 1. LDH level elevated in 9 (25.7%) pts. More than half pts (19, 54.3%) had metastatic mucosal melanoma, of whom 3 (15.8%) pts had more than 3 metastasis sites and 4 (21.1%) pts had liver metastasis. In stage I analysis set (n=22), the best confirmed ORR was 36.4% (95% CI, 17.0%-59.3%). Median progression-free survival was 5.32 months (95% CI, 1.58-not reached), and the best confirmed DCR was 59.1% (95%CI, 36.4%-79.3%). The median confirmed DoR was not reached (95% CI, 2.76-NR). In the full analysis set (n=28), the unconfirmed ORR was 42.9% (95%CI, 24.5%-62.8%). In ITT populations (n=35), 28 pts (80%) experienced at least one adverse event (AE) and 5 pts (14.3%) experienced at least one grade 3-4 AEs. Only one patient experienced AE leading to treatment discontinuation. One patient died of autoimmune lung disease. Conclusions: The combination of atezolizumab plus bevacizumab showed promising benefit and was tolerable in pts with advanced mucosal melanoma. At the time of this interim analysis, the primary endpoint did not cross the futility boundary, thus the study will run into Stage II. Clinical trial information: NCT04091217