323 - Percutaneous Hepatic Perfusion (PHP) with Melphalan for Treatment of Liver Metastases from Uveal Melanoma: Initial Commercial Experience

Purpose: Approximately 50% of patients with uveal melanoma (UM) develop metastases with the liver as the primary site of disease in > 90% of cases. Control of hepatic tumors is critical to prolonging overall survival for patients with metastatic uveal melanoma (MUM). We report our initial single institution experience with commercially available PHP with melphalan (HEPZATO, Delcath, Queensbury, NY) for the treatment of UM hepatic metastases.

Materials and Methods: MUM patients with < 50% hepatic tumor burden underwent commercially available PHP at 6 to 8-week intervals. Best radiographic response was assessed with contrast-enhanced MRI using RECIST 1.1 criteria. CTCAE v5.0 was used to assess adverse events.

Results: Twenty-two patients (9 men; median age, 62; range, 31–72) including 10 patients with extrahepatic metastasis, underwent PHP (n=76 procedures) from March 2024 to September 2025. All patients had at least one prior liver-directed therapy (LDT), and 10 patients failed prior systemic treatment. Patients underwent a median of 3 PHPs (range 1-6), and 21 patients had ≥ 2 procedures. Median follow-up was 8.3 months (range, 0.4-18.1). Fifteen patients remain alive after initial PHP. Of the 20 patients who underwent follow-up MRI, 10 had stable disease (SD) and 4 demonstrated a partial response (70% disease control rate). Four patients with progression of disease (PD) after PHP underwent repeat treatment; one achieved SD. Two patients stopped treatment due to intrahepatic (n=2) and extrahepatic (n=1) progression. Median time to progression of metastases has not been reached.

Eighteen patients required transfusions following 1 (n=5), 2 (n=9), 3 (n=3) or 4 (n=1) procedures. One patient developed grade 4 sinusoidal obstruction syndrome prohibiting additional LDT and died 5.9 months post PHP due to PD. Two patients required hospitalization due to grade 4 thrombocytopenia and leukopenia requiring transfusions, antibiotics, and supportive care. Two patients developed grade 3 neutropenia requiring cancellation of a treatment. Two patients had strokes precluding further PHP treatments: one occurred 2 weeks post-treatment of unclear etiology; the other occurred post-procedurally with near-complete symptom resolution on discharge.

Conclusion: PHP remains a promising treatment for MUM patients. The durability of disease control is uncertain based on this early commercial experience, and it is unclear how PHP compares to other LDT for this disease. However, in our experience, procedural complications seem to be more significant and treatment-limiting compared to alternative LDT.