Abstract
INTRODUCTION: Percutaneous hepatic perfusion with melphalan (PHP-M) for hepatic metastasis of uveal melanoma (LMUM) achieves high local response rates, but the individual clinical benefit is poorly defined. We aimed to determine cofactors of response and clinical outcomes including the probability of long-term (5-years) overall survival (OS) in PHP-M-treated patients with LMUM. Patients and Methods. We retrospectively reviewed clinicopathological, radiological, and outcome data of 19 patients with unresectable LMUM treated with 43 PHP-M (median 2 PHP-M) between 2014 and 2019. Tumor response and adverse events were evaluated using RECIST 1.1 and the Clavien-Dindo classification. Kaplan-Meier methods and Cox regression hazard proportional models were used. RESULTS: Of 19 patients, 10 (53%) achieved a partial response (PR) and 9 (47%) had stable disease (SD). There was no progressive disease (PD) and no adverse events exceeding Clavien-Dindo grade IV. Median OS was 16.7 months after the first PHP-M treatment and 26.4 months after initial diagnosis. Low hepatic tumor volume (median of 10 mL vs. 150 mL) was an independent predictor of favorable OS (hazard ratio (95% confidence interval): 0.190 (0.041, 0.893); p < 0.05), and female patients were at a lower risk compared with males (0.146 (0.017, 1.240)). Estimates of the overall survival were 0.213 (0.0449, 1) from first imaging (95% confidence interval) to 5 years and 0.793 (0.609, 1) and 0.604 (0.380, 0.960) for 1 and 2 years after chemosaturation, respectively. Discussion. PHP-M for nonresectable LMUV provides a safe and locally efficient liver-directed procedure that offers patients a chance for long-term OS, especially for patients with a low hepatic tumor burden.