<h3>Objectives:</h3> Pembrolizumab/lenvatinib demonstrated impressive response rates in advanced/recurrent endometrial cancer (EC) in a recent phase II clinical trial, leading to its accelerated FDA approval. However, tolerability of the recommended 20 mg starting dose of lenvatinib may pose clinical challenges. To comprehend how this regimen translates into daily clinical practice, we evaluated our preliminary data regarding toxicity and efficacy of pembrolizumab/lenvatinib in the treatment of advanced/recurrent EC at the University of Texas MD Anderson Cancer Center. <h3>Methods:</h3> In this single-institution, retrospective cohort study, we evaluated recurrent/advanced EC patients who received ≥1 cycle of pembrolizumab/lenvatinib treatment. Toxicity was evaluated through rates of hospitalization and treatment interruptions or discontinuations, as well as number of dose reductions. Clinical efficacy was evaluated by best radiologic response to treatment. Clinical benefit was defined as having stable disease or partial/complete response. <h3>Results:</h3> From October 2019 to July 2020, 83 EC patients were identified but 16 were excluded for unconfirmed administration of pembrolizumab/lenvatinib or lack of toxicity or response data. Thus, 67 EC patients were included in the analysis with a median follow-up of 5 months (range 1 - 11). The median age and number of prior lines of therapy were 67 (range 31–77) and 2 (range 1–9), respectively. The majority had an ECOG performance status of 0 to 1 (91%). The predominant histologic subtype was endometrioid and serous (n=19 each) followed by carcinosarcoma (n=13), mixed (n=8), clear cell (n=4), and other (n=4). The most common starting dose of lenvatinib was 14 mg (n=43; 64.2%), followed by 20 mg (n=15; 22.4%). Treatment-related hospitalizations occurred in 26.9% of patients but did not differ between lenvatinib dosage groups (p=0.37). Compared to those with a lower starting dosage level, patients started on 20 mg doses had significantly higher treatment discontinuations (100% vs 69.2%, p=0.01) and numbers of lenvatinib dose reductions (1.0 vs 0.38, p=0.004), with a trend of higher treatment interruptions (93.3% vs 73.1%, p=0.09). Among the 56 patients with evaluable radiologic response, there were 21 (37.5%) patients with either a partial (n=18) or complete response (n=3), while 18 patients (32.1%) had stable disease. Of the patients with complete response, all patients had a starting lenvatinib dose of 14 mg. Based on a starting lenvatinib dose of 20 mg vs < 20 mg, there were no differences in response to therapy (23.1% vs 41.9%; p=0.56; respectively) or clinical benefit rates (53.8% vs 74.4%, p=0.18; respectively). There were no differences in response rates among histologic subtypes (p=0.76). Patients with carcinosarcoma histology had a response and clinical benefit rate of 25% (3 of 12) and 58.3% (7 of 12), respectively. <h3>Conclusions:</h3> In treatment of recurrent/advanced EC, lower starting dosage of lenvatinib was associated with fewer treatment interruptions/discontinuations and dose de-escalations compared to the FDA recommended 20 mg dosage level while maintaining equivalent response and clinical benefit rates. Furthermore, patients with carcinosarcoma histology also appeared to benefit from pembrolizumab/lenvatinib. While larger studies are needed to validate these safety and efficacy findings, our preliminary data supports starting a lower dosage of lenvatinib in clinical practice.