32 | Tebentafusp in metastatic uveal melanoma: 3 years of real world data and insights from liquid biopsy

Luisa Piccin¹, Maria Chiara Scaini², Valentina Salizzato¹, Cristina Catoni², Elisabetta Di Liso¹, Alice Menichetti¹, Chiara Cacco^1|4, Chiara Gottardi^1|4, Francesca Di Sarra³, Marina Coppola³, Valentina Guarneri^1|4, Jacopo Pigozzo¹ | ¹Medical Oncology 2 - Veneto Institute of Oncology IOV-IRCCS; ²Immunology and Molecular Oncology Unit, Veneto Institute of Oncology IOV-IRCCS; ³Pharmacy Unit - Veneto Institute of Oncology IOV-IRCCS; ⁴Department of Surgery, Oncology and Gastroenterology, University of Padova, Italy.

Background: Real world data are needed to optimize the use of Tebentafusp, the first agent improving the overall survival in metastatic uveal melanoma (mUM) patients even if progressive disease (PD) is the best response.

Methods: Baseline clinical data, tumor assessments, toxicities, progression free survival (PFS) and overall survival (OS) were retrospectively collected in mUM patients treated with Tebentafusp between May 2022 and June 2025 at our center. Preliminary liquid biopsy data were also obtained for a cohort subgroup.

Results: This analysis reports our 3-years single institution experience: 21 HLA A*02:01 mUM patients treated with Tebentafusp and completing the dose-escalation phase. At the data cut-off 8 subjects were alive and 16 recorded PD. Median follow up was 12.6 months. Baseline characteristics: median age 66 years, ECOG-PS 0 52.4%, male gender 47.6%, no pretreatment 85.7%. Exclusive liver metastases in 33.3% and elevated LDH levels in 90.5% of cases. Instrumental evaluations, performed at least once in 18 patients, showed: disease progression as best response in 55.6% of patients (most of them continued Tebentafusp beyond oligoprogression due to clinical benefit) and disease control in 44.4% of cases. 1-year OS: 70.8%. Median PFS and OS: 3.8 (3.4-10;95%CI) and 13.9 (12.4-NA,95%CI) months, respectively. Adverse drug reactions included: cytokine release syndrome (81%), skin toxicities (66.7%) and hepatic enzyme increase (23.8%). Grade 3 toxicities (28.6%) never led to treatment discontinuation. Although performed preliminarily in 7 patients, liquid biopsy assessments suggest a potential role of this strategy in monitoring therapy benefits.

Conclusions: Activity, mPFS, 1-year OS and the manageable safety profile were comparable to the registrative trial. Instead, our shorter mOS could be explained by the unfavorable prognostic features of our population. Prospective real-world studies including ctDNA and CMC evaluations should be encouraged to improve treatment performance and its duration.