Background: Few platinum resistant ovarian cancer (PROvC) patients respond to anti-PD1 monotherapy (ORR 7.6%) with little impact on survival (OS 10.1 mo). Among responders the median duration of response is impressive (18.7 mo) (Hamanishi 2021). Methods: We have evaluated the combination of pembrolizumab (200mg), with bevacizumab (400mg) for 6 cycles plus minus peglyated liposomal doxorubicin (PLD) q3w in PROvC patients with no limit in previous treatment lines, allowed to be previously treated with bevacizumab. An initial safety run evaluated the dual combination of pembrolizumab plus PLD (cohort A). The triple combination was evaluated at MTD-1 and at MTD of PLD (30mg/m² q3w) (cohort C). The dual combination of pembrolizumab + bevacizumab was run in parallel (cohort B). This is an open label phase I trial with a modified toxicity probability interval design. The evaluation criteria endpoints were safety and efficacy. Pharmacokinetics of bevacizumab were evaluated. NCT03596281 Results: A total of 47 patients (pts) were enrolled between January 2019 and February 2021. Median age was 70 years (38-77). 30/12 pts (63.8/25.5%) had an initial FIGO stage III/IV, 44 pts (93.6%) had a HGSOC. 40 pts (85.1%) underwent surgery, out of which 13 pts (32.5%) had a primary debulking. BRCA mutations were present in 9 pts (19.1%). Pts had a median of 3 previous treatment lines (0-13), including pretreatment with antiangiogenic agents in 36 (76.6%) and PARP inhibitors in 21 pts (44.7%). No DLT was reported. Grade 3/4 treatment-related adverse events were reported in 2 pts (30%), 4 (20%) and 11 (50%) in cohorts A, B and C respectively. The ORR was 0, 26.3 (95% CI 6.5-46.1) and 30% (9.9-50.1) with a DCR of 0, 78.9 and 75% in cohorts A, B and C respectively. According to investigator assessment, the median PFS was 2.1, 4.7 and 4.8 mo (table). The blinded independent central review is currently under evaluation. A large inter-patient variability in bevacizumab plasma concentrations was observed among patients. The 400 mg flat dosing achieved residual concentrations similar to that of 5 mg/kg Q2W or 7.5 mg/kg q3w (51± 30 μg/ml in cohort B and 63 ± 55 μg/ml in cohort C (p>0.05) after C1). Overall, 22 % of pts of cohort B and 18 % of cohort C showed trough levels below the targeted threshold (i.e. < 25 μg/ml). Correlative studies are ongoing. Conclusions: Short-term flat dose bevacizumab potentiates the response to anti-PD1 therapy even in the absence of chemotherapy in heavily pre-treated PROvC patients. The long-term treatment with bevacizumab could potentially improve the outcome. The combination of anti-PD-1 plus anti-angiogenic agents should be a backbone for the treatment of PROvC patients. Clinical trial information: NCT03596281.