Percutaneous hepatic perfusion with melphalan in uveal melanoma: A safe and effective treatment modality in an orphan disease

Background Metastatic uveal melanoma (UM) carries a poor prognosis; liver is the most frequent and often solitary site of recurrence. Available systemic treatments have not improved outcomes. Melphalan percutaneous hepatic perfusion (M‐PHP) allows selective intrahepatic delivery of high dose cytotoxic chemotherapy. Methods Retrospective analysis of outcomes data of UM patients receiving M‐PHP at two institutions was performed. Tumor response and toxicity were evaluated using RECIST 1.1 and Common Terminology Criteria for Adverse Events (CTCAE) v4.03, respectively. Results A total of 51 patients received 134 M‐PHP procedures (median of 2 M‐PHPs). 25 (49%) achieved a partial (N = 22, 43.1%) or complete hepatic response (N = 3, 5.9%). In 17 (33.3%) additional patients, the disease stabilized for at least 3 months, for a hepatic disease control rate of 82.4%. After median follow‐up of 367 days, median overall progression free (PFS) and hepatic progression free survival (hPFS) was 8.1 and 9.1 months, respectively and median overall survival was 15.3 months. There were no treatment related fatalities. Non‐hematologic grade 3‐4 events were seen in 19 (37.5%) patients and were mainly coagulopathic (N = 8) and cardiovascular (N = 9). Conclusions M‐PHP results in durable intrahepatic disease control and can form the basis for an integrated multimodality treatment approach in appropriately selected UM patients.


INTRODUCTION
Metastatic uveal melanoma (UM) carries a dismal prognosis with 1 year survival rates reported at 10-25%. [1][2][3] Unlike cutaneous melanoma (CM) where recent developments in the field of immunotherapy 4 and targeted therapy 5 have transformed the outlook in the metastatic setting, there are no established effective systemic treatments for metastatic UM. Activating BRAF mutations are rare 6 and so far no alternative molecular targeted agents have demonstrated significant activity. 7 Immunotherapy of UM to date has been extremely disappointing with response rates of <10%, much lower than those seen in CM. [8][9][10] This is especially true in the context of progressive liver disease, 11 which is common in metastatic UM as the liver is involved in >85% of cases of metastatic spread. 3 The liver microenvironment is known to facilitate immune escape 12 and the specific mechanisms involved may account both for the predilection of UM for liver metastases and the reduced efficacy of immunotherapeutic agents in patients with progressive liver disease. As the liver is the sole site of metastatic involvement in around 50% of UM cases, 3 adopting a liver-directed treatment approach can result in clinically meaningful periods of disease control while minimizing systemic toxicity. Resection or ablation of metastatic deposits is associated with prolonged survival in preselected patient groups. 13,14 For unresectable or multifocal small volume disease, arterially delivered methodsm-such as chemo-, radio-, and immunoembolization and isolated hepatic perfusion (IHP) have been devised. 15 The concept is to exploit the differential blood supply to the metastatic deposits, derived almost exclusively from the hepatic artery, 16 as opposed to the supply of the healthy parenchyma primarily from the portal vein. 17 IHP was established as a surgical procedure involving temporary surgical isolation of the hepatic circulation and delivery of high dose of cytotoxic chemotherapy through the hepatic artery. 18 While IHP was shown to have response rates of approximately 40-50% in UM, 19

Patient eligibility
All patients with histologically confirmed UM who underwent M-PHP in our institutions between December 2008 and October 2016 were included in this retrospective study. Approval for retrospective analysis of treatment outcomes was obtained from the institutional review boards of participating centres. Previous systemic or liverdirected treatments other than M-PHP were allowed provided any related adverse events (AEs) had either resolved or were not expected to impact the safety or efficacy of the procedure. Patients with known or suspected extrahepatic disease were included if disease was nonprogressive following previous treatments or amenable to ablative treatment modalities.
Generally, several weeks prior to M-PHP, angiography is performed to delineate the arterial supply to the liver and a strategy for chemotherapy infusion is formulated. Occasionally coil embolization of vascular variants, such as the gastruoduodenal or right gastric arteries, that may predispose the patient to inadvertent flow of chemotherapeutic drugs into branches supplying the gastrointestinal tract may be required. The catheter is then connected to an extracorporeal circulation system consisting of a centrifugal pump and two drug filtration activated carbon filters ( Figure 1). Blood is aspirated through catheter fenestrations in a segment between the two balloons, actively pumped through Once correct positioning of the two balloons is confirmed, the filtration of blood by the two cartridges is started in a stepwise fashion. A centrifugal pump is used to achieve appropriate flow rates. The hemofiltration filters are brought online and once the cartridges are completely filled with blood, the bypass line is closed. When the hemofiltration circuit is running adequately and the patient is hemodynamically stable intra-arterial infusion of melphalan is started.
The dose of melphalan was calculated at 3 mg/kg, corrected for the patient's ideal body weight (maximum dose: 220 mg). After the infusion, extracorporeal filtration is continued for a 30 min ("washout period") to allow clearance of melphalan from the liver. 20 Post M-PHP procedure, protamine sulphate is infused to reverse heparinization, and blood products are transfused to replace clotting factors as needed. The vascular sheaths are left in place until coagulation is sufficiently corrected. Once the patient coagulation profile normalizes, the vascular sheaths are removed and pressure is held on the sites for 45 min. Once stable, patients were transferred to the intensive care unit for monitoring and most received G-CSF within 72 h of melphalan administration.  disease progression was felt to be attributable to differential perfusion of liver parenchyma due to anatomic constraints, subsequent attempts with M-PHP would be made to preferentially target these areas.

Response assessment
Either a dedicated liver MRI or a triple phase liver CT was performed to assess tumor response following the guidelines set forth in RECIST 1.1. 22

Data capture and analysis
Data was collected retrospectively from the electronic medical record.
GraphPad Prism Version 6.01 was used for survival curve graphing and analysis using the Kaplan-Meier method; log-rank test was used to compare curves and determine the P value. SPSS version 23.0.0 was used for Cox regression analysis.

Response analysis
Radiological assessments took place as clinically indicated, typically 6-8 weeks after each treatment. progressed systemically at the time of the first assessment. Figure 2 illustrates by way of a waterfall plot the magnitude of observed responses.
In 17/51 (33.3%) patients, the best hepatic response was stable disease (SD) for a minimum of three months; in 11 this was

Survival analysis
After a median follow-up of 12.2 months, at the time of data cut off, median OS was 15.3 months. Seventeen patients were still alive, five on-treatment, 32 had passed away and two were lost to follow-up.
One year OS rate is 64.6% ( Figure 3A). Patients who responded had significantly improved survival as opposed to non-responders ( Figure 3B, P < 0.01). Two-year OS for the responders was 50.2%

Safety analysis
M-PHP was well tolerated in this study population with frequency and adverse events types commensurate with those reported in the original Phase III study 21 (Table 3). There were no treatment related fatalities. Nineteen patients (37.5%) experienced grade 3-4 nonhematologic treatment related toxicity. Cardiovascular toxicity was primarily observed peri-procedurally-three cases of ventricular tachycardia and one case of supraventricular tachycardia were seen.
There were five cases of post-operative troponin elevation, one with associated ECG changes suggestive of non-ST elevation myocardial infarction and one associated with pulmonary oedema. In addition there were two episodes of pulmonary oedema without documented associated myocardial ischaemia.
Bleeding events were common peri-operatively and seen in 19  "high" disease burden implies more than 10 lesions or more than 50% parenchymal involvement.  Table 3.  Finally, we need to consider that improved outcomes may simply be a "stage migration"-like effect due to selection of patients with earlier disease. Historically presentation was late and driven by chance findings of deranged liver function tests or symptoms relating to liver capsule pain or biliary tract obstruction. As more treatment modalities are becoming available, there has been increased recognition of the importance of early diagnosis and routine biannual liver imaging in high-risk patients is now considered standard practice. 23 The debate regarding the merit of regional therapy in melanoma and in particular of liver directed therapy in UM is longstanding. 24