1141P - Percutaneous hepatic perfusion (chemosat or cs-php) of melphalan vs. Best alternative care (BAC) in patients (pts) with hepatic metastases from mela...

Date 01 October 2012
Event ESMO Congress 2012
Session Poster presentation III
Topics Anti-Cancer Agents & Biologic Therapy
Melanoma and other Skin Tumours
Presenter H. Alexander Jr.
Authors H.R. Alexander Jr.1, -. On Behalf Of Phase 3 Principal Investigators2
  • 1Surgery, University of Maryland Medical Center, Baltimore/US
  • 2-, -, Baltimore/US



CS-PHP (CHEMOSAT®; Delcath Systems Inc, New York, NY) is a regional therapy which isolates and perfuses the liver with high-dose chemotherapy. Unwanted systemic exposure is minimized by extracorporeal filtration of hepatic venous blood.


A prospective randomized multicenter phase 3 study compared CS-PHP of melphalan with BAC in pts with proven unresectable hepatic metastases from ocular or cutaneous melanoma. A survival update was performed on 31 March 2011. CS-PHP melphalan 3.0 mg/kg ideal body weight was infused into the hepatic artery over 30 min with concurrent extracorporeal filtration for 60 min. Up to 6 treatments were given every 4–8 wks. In the BAC group, crossover to CS-PHP melphalan was permitted after hepatic disease progression. The primary endpoint was investigator-assessed hepatic progression-free survival (hPFS). An exploratory post-hoc analysis of pts who crossed from BAC to CS-PHP vs. BAC-only pts was also performed.


93 pts were randomized to CS-PHP (n = 44) or BAC (n = 49). After hepatic disease progression, 28 pts crossed over to CS-PHP. Results are shown in the Table. The most common grade 3/4 toxicities in CS-PHP pts (n = 40) were hematological peri-procedural thrombocytopenia (73%) and anemia (55%) and post-procedural (beyond day 4 post-treatment) neutropenia (93%) or thrombocytopenia (83%). The safety profile in crossover pts was similar to that in pts randomized to CS-PHP melphalan.


CS-PHP melphalan significantly prolonged hPFS compared with BAC in pts with liver-dominant metastatic melanoma, thereby meeting the primary study objective. Efficacy was similar after hepatic disease progression in BAC-CS-PHP crossover pts as in those randomized initially to CS-PHP.

Treatment group n Median hPFS, mo Hazard ratio (95% CI) Median OS, mo Hazard ratio (95% CI)
CS-PHP 44 8.0 0.35 (0.23-0.54)P < 0.0001 9.8 1.08 (0.69-1.68)NS
BAC 49 1.6 9.9
BAC only 21 1.6 0.32 4.1 0.33
BAC → CS-PHP crossover 28 8.8 15.3


All authors have declared no conflicts of interest.