Special Report

Chemosaturation With Percutaneous Hepatic Perfusion in Unresectable Hepatic Metastases Evan S. Glazer, MD, PhD, and Jonathan S. Zager, MD

Background: In patients with hepatic metastases from solid-organ malignancies, surgical resection may be a potentially curative option, but it is not possible in most cases. Chemosaturation with percutaneous hepatic perfusion was developed for the management of unresectable metastases to the liver. Methods: Relevant medical literature was summarized with regard to the outcomes and limitations of che- mosaturation with percutaneous hepatic perfusion. Results: Six articles were identified that contained data on 91 individuals who received chemosaturation with percutaneous hepatic perfusion. More than 60% of these study patients were diagnosed with ocular . The overall response rate was 48% and the rate of disease control was 90%. Chemosaturation with percutaneous hepatic perfusion improved the rates of overall and hepatic progression-free survival (PFS). The data are limited but suggest that the rate of PFS was improved in study patients with isolated melanoma hepatic metastases who received chemosaturation with percutaneous hepatic perfusion compared with those assigned to standard care. Conclusions: Our results suggest that chemosaturation with percutaneous hepatic perfusion produces favor- able tumor response rates in select individuals with unresectable hepatic metastases from multiple primary , particularly ocular and cutaneous . Data from a single randomized clinical trial have also shown that chemosaturation with percutaneous hepatic perfusion can affect hepatic PFS in certain patients.

Introduction primary .4 The rate of median survival in per- Hepatic metastases from solid-organ malignancies sons with hepatic metastases from ocular melanoma portend a poor prognosis in most cases and are an im- is less than 9 months.5 Hepatic metastases will devel- portant determinant to survival.1-3 Complete surgical op in 20% to 40% of individuals with colorectal can- resection might represent a curative option in patients cer; the median survival rate in these cases is approxi- with isolated hepatic metastases, but resection is not mately 2 years.2,6 Neuroendocrine tumors metastasize possible in most patients because of the number, loca- to the liver in 25% to 90% of individuals, and the tion, or size of the hepatic metastases. Likewise, poor 5-year survival rate of those with liver metastases is hepatic functional reserve limits standard therapeutic 40% compared with a rate that exceeds 75% in those options such as resection. whose cancer does not metastasize to the liver.1 He- The liver is the predominant site of for patic metastases can also co-occur with many other ocular melanoma, colorectal cancer, and gastrointes- malignancies, such as breast cancer, renal cell cancer, tinal neuroendocrine tumors.1-3 In persons with ocu- cutaneous melanoma, and soft-tissue sarcoma.7,8 For lar melanoma, the sole site of metastatic disease in up patients with unresectable liver-dominant disease, to 50% of individuals is the liver after treatment of the multiple strategies for liver-directed therapy have been developed, including hepatic artery chemoem- bolization, radioactive bead embolization, hepatic ar- From the Department of Cutaneous , H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida. terial infusion, isolated hepatic infusion, and chemo- 3 Submitted April 12, 2016; accepted June 24, 2016. saturation with percutaneous hepatic perfusion. Address correspondence to Jonathan S. Zager, MD, Department of In Europe, chemosaturation with percutaneous Cutaneous Oncology, Moffitt Cancer Center, 10920 North McKinley hepatic perfusion was commercially launched in 2012, Drive, Room 4123, Tampa, FL 33612. E-mail: Jonathan.Zager@Mof- and a second-generation, high-efficiency filter was ap- fitt.org proved for use in conjunction with a proprietary he- Dr Glazer is now affiliated with the Division of Surgical Oncology 3,9 in the Department of Surgery at the University of Tennessee Health patic delivery system that same year. However, the Sciences Center, Memphis, Tennessee. US Food and Drug Administration has not approved Dr Zager serves on the medical advisory board for Delcath. Dr Glaz- chemosaturation with percutaneous hepatic perfusion er reports no significant relationship with the companies/organiza- tions whose products or services may be referenced in this article. for use in the United States. This article describes therapy not approved by the US Food and Chemosaturation with percutaneous hepatic per- Drug Administration. fusion is a minimally invasive, repeatable, regional

96 Cancer Control January 2017, Vol. 24, No. 1 therapy in which a high dose of is direct- flow from the proper hepatic artery. If the catheter is ly infused into the liver via the hepatic artery using a placed in the common hepatic artery, or there is back percutaneous approach.3,9,10 The liver is isolated from flow into the gastroduodenal artery, then the gastrodu- the systemic circulation by a double-balloon catheter odenal artery is embolized to avoid the systemic circu- inserted through the femoral vein. -in- lation of chemotherapy. Likewise, unnamed collaterals fused blood is then diverted through the arterial cath- or other noncritical arteries are also embolized. Iden- eter via an extracorporeal pump circulation and is then tification of accessory or replaced hepatic arteries are filtered and returned to the patient via venovenous by- identified (if not done so already) and used for hepatic pass and the jugular vein. chemoperfusion if needed. Blood from the hepatic veins then exits the liver Methods through a channel in the multilumen catheter and is Authors reporting on the utility of percutaneous hepat- extracorporeally filtered before it is returned to the cir- ic perfusion for any malignancy met our inclusion crite- culation to limit systemic toxicity from chemotherapy. ria. Only literature written in English was reviewed. If The second-generation extracorporeal filtration system data were reported in combination with other hepatic- extracts greater than 90% of melphalan before return- directed therapies, only the data reported for chemo- ing the filtered blood back to the patient through the saturation with percutaneous hepatic perfusion were internal jugular vein to minimize systemic drug deliv- included. Exclusion criteria included 5 or fewer case pa- ery and the associated adverse effects.15 tients or data not evaluated in a peer-reviewed fashion. The procedure is described elsewhere in detail, For literature from the same institution, original but high-dose melphalan 3 mg/kg is administered deidentified data were requested to remove duplicate unless dose-limiting toxicities have been previously study patients from multiple reports. Once all study pa- demonstrated.7,10 All procedures are performed under tients were determined to be unique, we analyzed the general anesthesia with heparin-induced systemic an- response and disease control rates. The authors of each ticoagulation performed after catheterizations. Prior individual report determined the response rate. We de- to introducing chemotherapy, contrast angiography fined the disease control rate as the summation of com- is obtained of all the isolated sections to confirm the plete response, partial response, and stable disease. absence of leakage and to verify isolation from the systemic circulation. Melphalan is administered to Literature Review the liver in a split fashion by advancing the hepatic We identified 6 articles from the medical litera- artery catheter into the left and right hepatic arteries. ture.7,8,10-13 The report from Abbot et al11 is an update Unless anatomical variations exist, 60% of the drug is of the study results from Forster et al,7 so these data are given via the right hepatic artery and 40% is given via combined in our review. Likewise, some study patients the left hepatic artery. Treatment is administered for reported in Hughes et al10 were included in Abbot 30 minutes, followed by an additional 30 minutes of et al,11 so we also omitted those case patients from in- venovenous bypass to allow for melphalan washout. clusion in our review of the results from Abbot et al11 During the treatment, patients undergo clinically sig- and Forster et al.7 No data from Pingpank et al14 were nificant changes in pulse and blood pressure that re- utilized because this report included preliminary re- quire aggressive support with intravenous fluids and sults from a randomized control trial eventually pub- vasoactive drugs such as phenylephrine, norepineph- lished by Hughes et al.10 rine, and vasopressin. This is especially the case early on during the procedure when the filters are opened, Procedural Details and Management and vasoactive support is weaned off as the procedure Chemosaturation with percutaneous hepatic perfu- continues. After the procedure is completed, prot- sion is performed in an interventional radiology suite amine is used to reverse heparin. or operating room with imaging capability under gen- Catheters and intravascular sheaths are removed eral anesthesia (Fig). A catheter is first manipulated after baseline laboratory blood values (complete blood into the hepatic artery via the femoral artery followed count, platelet count, activated clotting time, prothrom- by catheterization of the internal jugular and femo- bin time, and partial thromboplastin time) have nor- ral veins utilizing ultrasonography and fluoroscopic malized, with transfused fresh-frozen plasma, plate- guidance. A multilumen, double-balloon catheter is lets, cryoprecipitate, and fibrinogen (blood proteins inserted via the femoral vein into the inferior vena removed by the filtration system) given as needed. All cava and positioned across the hepatic veins. The bal- patients receive 10 units of cryoprecipitate, because loons are inflated to isolate hepatic venous outflow the real-time measuring of individual clotting proteins from the systemic circulation. found in cryoprecipitate can be difficult, cumbersome, Hepatic artery angiography is obtained to confirm and is not routinely performed. However, platelets, red the absence of collateralization or extrahepatic blood blood cells, and fresh-frozen plasma are transfused

January 2017, Vol. 24, No. 1 Cancer Control 97 Connect to Drug Delivery System

Fig. — Illustration of the isolation of the retrohepatic inferior vena cava for the return of melphalan-infused blood via a double-balloon catheter and cath- eterization of the proper hepatic artery. Shown is a complete setup for percutaneous hepatic perfusion, including drug delivery (yellow via left femoral artery catheterization into the hepatic proper artery); normal saline for priming (A; tubing 1 and 2); removal of melphalan-infused blood from the retrohepatic inferior vena cava double-balloon catheter (B and C; tubing 3 and 10); bypass and filtration (D to E; tubing 5–8); and the return line (F; tubing 9 and 11). This figure describes therapy not approved by the US Food and Drug Administration. Reprinted with permission by Delcath Systems, Inc. based on postprocedure laboratory values. Pressure is er protocol. Positron emission tomography/computed held at each vascular access point for 45 minutes, and tomography is used to evaluate disease response and patients are then are observed for 1 day in the inten- for other sites of metastatic disease and functional sive care unit. Typically, patients are discharged home disease burden. Magnetic resonance imaging of the on postoperative day 1, provided that laboratory val- brain is also obtained in cases of melanoma to rule ues (complete blood count, renal and hepatic function out central nervous system metastases. tests) have returned toward baseline. Patients are postoperatively monitored with Outcomes twice-weekly laboratory studies (complete blood Increased Experience count, renal and hepatic function tests) for up to In 2004, van Etten et al12 published a report of a small 4 weeks. Granulocyte-stimulating factor is used as series of 10 study patients with colorectal and ocular needed for neutropenia, and, if present, anemia and melanoma hepatic metastases who underwent chemo- thrombocytopenia are treated with transfusions. saturation with percutaneous hepatic perfusion. Over- At our institution, we obtain repeat liver imaging all, these patients had a disease control rate of 80% and 6 weeks following treatment with a triple-phase liver a 30-day mortality rate of 5%.12 The median number of protocol for computed tomography or contrast-en- treatments per study patient was not reported. The re- hanced magnetic resonance imaging, also per the liv- searchers compared chemosaturation with percutane-

98 Cancer Control January 2017, Vol. 24, No. 1 ous hepatic perfusion and open isolated hepatic per- tivariate analysis, this difference was significant for fusion, which has significantly higher morbidity and chemosaturation with percutaneous hepatic perfu- mortality rates.9,12 Because open isolated hepatic perfu- sion compared with yttrium alone (P = .03) and not sion has risks and, due to the development of percu- for chemosaturation with percutaneous hepatic per- taneous intravascular techniques, van Etten et al12 rec- fusion compared with chemoembolization (P = .20).11 ommended percutaneous hepatic perfusion without Neither ECOG score nor tumor burden was a signifi- abdominal incision as a potential therapeutic option cant predictor of overall survival (OS). for patients with unresectable hepatic malignancies.12 Because 7 study patients included in the retro- Vogl et al8 published a small series in 2014 of spective review from Forster et al7 and Abbott et al11 13 study patients who underwent chemosaturation were also included in clinical trial results published by with percutaneous hepatic perfusion for hepatic me- Hughes et al,10 it is worth noting that these data were tastases predominately from melanoma. The mortal- removed from the summary analysis and the Table that ity rate in this series was similar to previously report- describes the unique patients.8,12,13 ed series, and the median number of treatments per Hickson et al13 described the experience of treat- study patient was 1.8 The disease control rate for this ing 18 study patients who received chemosaturation cohort was 92%, thus demonstrating that chemosatu- with percutaneous hepatic perfusion. The primary dis- ration with percutaneous hepatic perfusion has a sig- ease of the cohort was ocular melanoma, and the me- nificant effect.8 dian number of treatments was 2 per study patient.13 Abbott et al11 built upon the data from Forster et al,7 A disease control rate of 94% was reported, confirming updating it to investigate the use of chemosaturation the work of previous groups who demonstrated signifi- with percutaneous hepatic perfusion in 13 individu- cant responses to chemosaturation with percutaneous als with ocular and cutaneous melanoma. These re- hepatic perfusion in individuals whose disease failed searchers evaluated the outcomes of chemosaturation to respond to other treatments.13 with percutaneous hepatic perfusion at a single insti- The results of a phase 3 randomized trial with a tution in study patients with unresectable hepatic me- cross-over design published by Hughes et al10 vali- tastases who received liver-directed therapy between dated the results of previously published retrospec- 2008 and 2014. Their data suggested that chemosatu- tive studies. In 44 study volunteers assigned to che- ration with percutaneous hepatic perfusion was suc- mosaturation with percutaneous hepatic perfusion, cessful in this cohort.11 The researchers compared the the disease control rate was reported to be 89%, the results of chemosaturation with percutaneous hepatic mortality rate was 7%, and the complication rate was perfusion against other liver-directed therapies, such 90%, with the majority being neutropenic without sig- as yttrium and chemoembolization, in 31 individuals. nificant sequelae.10 The median number of treatments No differences were observed between the groups was 3 per participant. Hepatic PFS and OS were signifi- receiving yttrium, chemoembolization, and chemo- cantly improved in this population. The rate of median saturation with percutaneous hepatic perfusion in re- OS improved from 46 days in the best available care gard to age, adjuvant therapy, prior regional hepatic group to 186 days in the chemosaturation with percu- treatment, or major complications following treat- taneous hepatic perfusion group (P < .0001).10 Because ment for both melanoma and non–melanoma primary cross-over was allowed, no difference in OS rate was malignancies.11 Extrahepatic disease was observed to observed as more than 50% of the study participants be more prevalent in the chemoembolization group who received best available care crossed over to per- compared with the groups receiving yttrium and cutaneous hepatic perfusion. Among the study vol- those receiving chemosaturation with percutaneous unteers originally randomized to best available care, hepatic perfusion (P = .004).11 The median rate of he- those who crossed over to chemosaturation with per- patic progression-free survival (PFS) was significantly cutaneous hepatic perfusion had a median OS rate of longer among those who received chemosaturation 396 days compared with 124 days for study partici- with percutaneous hepatic perfusion compared with pants assigned to the best available care group who those assigned to yttrium (P < .001) and chemosat- did not cross-over after disease progression (P = .01).10 uration with percutaneous hepatic perfusion com- pared with chemoembolization (P = .008).11 A higher Response Rates Eastern Cooperative Oncology Group (ECOG) score In total, 91 individuals received a total of 182 treat- (P = .01) and a greater tumor burden also correlated ments of chemosaturation with percutaneous he- with a shorter duration of hepatic PFS (P = .03).11 patic perfusion.7,8,10-13 Although it was not precisely Median overall Kaplan-Meier survival was also lon- reported in all studies, ocular melanoma was the ger with chemosaturation with percutaneous hepat- primary malignancy in more than 60% of the cases ic perfusion (736 days) than with yttrium (285 days) studied (see Table).7,8,10-13 When investigating these and chemoembolization (265 days); however, on mul- studies, based on Response Evaluation Criteria In

January 2017, Vol. 24, No. 1 Cancer Control 99 Table. — Selected Literature on Outcomes Data for Use of Chemosaturation With Percutaneous Hepatic Perfusion No. of No. of Median Rate of Type of Primary Selected Study Study Treatments No. of Disease Comment Tumor, n Outcome Patients Received Treatments Control, % First-Generation Filter CR: 0 Phase 3 randomized trial Cutaneous melanoma (5) PR: 16 Hughes10 44 120 3 89 Deaths associated with Ocular melanoma (39) SD: 23 treatment: 3 (7%) PD: 5 CR: 0 Colorectal cancer (9) PR: 2 van Etten12 10 10 Not reported 80 30-d mortality rate: 5% Ocular melanoma (1) SD: 6 PD: 2 Second-Generation Filter

Ocular melanoma (4) CR: 1 PFS improved with therapy Abbott11 Sarcoma (1) PR: 5 compared with other treatments (update of 10 30 3 100 (P < .04) Forster7)a Unknown primary mela- SD: 4 noma (1) PD: 0 No reported mortalities CR: 2 PR: 13 2 study patients NT due to technical Hickson13 18 Ocular melanoma (18) 34 2 94 SD: 2 anatomical reasons PD: 1 Breast cancer (1) CR: 1 1 death resulted in NE Cholangiocarcinoma (1) PR: 6 Vogl8 13 18 1 92 1 study patient NT due to intra- Cutaneous melanoma (3) SD: 5 procedural bleeding Ocular melanoma (8) NE: 1 Disease control rates are CR + PR + SD. aSeven patients not listed here because they are included in Hughes et al.10 CR = complete response, NE = not evaluable, NT = not treated, PD = progressive disease, PR = partial response, SD = stable disease. This table describes therapy not approved by the US Food and Drug Administration.

Solid Tumors (when available), the complete response in those receiving chemosaturation with percutane- rate was 4%; the overall response rate was 47%; the dis- ous hepatic perfusion compared with those assigned ease control rate was 90%; the median mortality rate to best alternative care (7.03 vs 1.64 months; hazard was 2.5% (range, 0%–7.7%); and, when reported, the ratio 0.34; P < .0001).10 The rate of OS was comparable median number of chemosaturation with percutane- among groups (9.79 vs 9.89 months for chemosatura- ous hepatic perfusion treatments per study patient tion with percutaneous hepatic perfusion and best was 3 (range, 1–3).7,8,10-13 alternative care, respectively).10 This may be because of the design of the trial. Because it was a cross-over Hepatic Function and Overall Survival design, the trial allowed patients assigned to best al- As of publication, the best available data on hepat- ternative care to cross-over to chemosaturation with ic function and OS come from Hughes et al,10 who percutaneous hepatic perfusion after progression. published the results of a phase 3, randomized, The most common adverse events observed in study cross-over trial that investigated chemosaturation patients receiving chemosaturation with percutane- with percutaneous hepatic perfusion using melpha- ous hepatic perfusion with melphalan were thrombo- lan vs best alternative care for patients with metastatic cytopenia, anemia, and neutropenia.10 ocular or cutaneous melanoma to the liver. Limited extrahepatic metastatic melanoma was allowed, pro- Use in Multiple Tumor Types vided that the hepatic metastases represented the Several noncomparative studies of chemosaturation life-limiting tumor burden. Patients were required to with percutaneous hepatic perfusion using melphalan have adequate hepatic and renal function. No ana- for the treatment of unresectable hepatic metastases tomical variations were allowed that would limit che- were published prior to the randomized trial conduct- mosaturation with percutaneous hepatic perfusion. ed by Hughes et al.8,11-13 Study patients in these trials The researchers discovered that PFS was prolonged had a variety of primary tumor types, including ocu-

100 Cancer Control January 2017, Vol. 24, No. 1 lar melanoma, cutaneous melanoma, cholangiocarci- cutaneous hepatic perfusion with best alternative care is noma, leiomyosarcoma, and colorectal, breast, and now recruiting patients (NCT02678572). gastric cancers.8,11,13 In these studies, rates of overall hepatic response ranged from 50% to 75%.8,11,13 Hepat- References ic responses were seen in those with cholangiocarci- 1. Chamberlain RS, Canes D, Brown KT, et al. Hepatic neuroendo- crine metastases: does intervention alter outcomes? J Am Coll Surg. noma, colorectal cancer, ocular melanoma, cutaneous 2000;190(4):432-445. melanoma, and leiomyosarcoma. Complete responses 2. Goldberg RM, Rothenberg ML, Van Cutsem E, et al. The continuum of care: a paradigm for the management of metastatic colorectal cancer. were observed in 1 study patient with cholangiocarci- Oncologist. 2007;12(1):38-50. noma and 2 with ocular melanoma.8,11,13 3. Agarwala SS, Eggermont AM, O’Day S, et al. Metastatic melanoma to the liver: a contemporary and comprehensive review of surgical, sys- temic, and regional therapeutic options. Cancer. 2014;120(6):781-789. Adverse Events 4. Egan KM, Seddon JM, Glynn RJ, et al. Epidemiologic aspects of uveal melanoma. Surv Ophthalmol. 1988;32(4):239-251. Observed toxicities following chemosaturation 5. Gragoudas ES, Egan KM, Seddon JM, et al. Survival of patients with with percutaneous hepatic perfusion are consistent metastases from uveal melanoma. Ophthalmology. 1991;98(3):383-390. 6. Norstein J, Silen W. Natural history of liver metastases from among reports of the therapy used in conjunction colorectal carcinoma. J Gastrointest Surg. 1997;1(5):398-407. with melphalan.7,14,16-18 Overall, approximately 90% 7. Forster MR, Rashid OM, Perez MC, et al. Chemosaturation with percutaneous hepatic perfusion for unresectable metastatic melanoma of patients have experienced a treatment-related ad- or sarcoma to the liver: a single institution experience. J Surg Oncol. verse event.7,8,10 Thrombocytopenia, anemia, and 2014;109(5):434-439. 8. Vogl TJ, Zangos S, Scholtz JE, et al. Chemosaturation with percuta- neutropenia are the most common adverse events of neous hepatic perfusions of melphalan for hepatic metastases: experience chemosaturation with percutaneous hepatic perfu- from two European centers. Rofo. 2014;186(10):937-944. 7,8,10,14,16-18 9. Yamamoto M, Zager JS. Isolated hepatic perfusion for metastatic sion. In the majority of cases, these adverse melanoma. J Surg Oncol. 2014;109(4):383-388. events were managed with granulocyte colony-stim- 10. Hughes MS, Zager J, Faries M, et al. Results of a randomized con- trolled multicenter phase III trial of percutaneous hepatic perfusion com- ulating factor, blood-product transfusions, or both, as pared with best available care for patients with melanoma liver metasta- necessary.7,8,10,14,16-18 ses. Ann Surg Oncol. 2016;23(4):1309-1319. 11. Abbott AM, Doepker MP, Kim Y, et al. Hepatic progression-free and Two studies with the highest toxicity rates — van overall survival after regional therapy for the liver for metastatic melanoma. Etten et al12 and Hughes et al10 — utilized the first-gen- Am J Clin Oncol. 2017. Epub ahead of print. 12. van Etten B, Brunstein F, van IMG, et al. Isolated hypoxic hepatic eration filter. A second-generation filter was utilized perfusion with orthograde or retrograde flow in patients with irresectable in the other studies, which had lower rates of grade 3 liver metastases using percutaneous balloon catheter techniques: a phase 7,8,11,13 I and II study. Ann Surg Oncol. 2004;11(6):598-605. or higher toxicities. Furthermore, pharmacoki- 13. Hickson G, Karydis I, Wheater MJ, et al. Single centre experience netics have shown that more than 93% of melphalan of chemosaturation percutaneous hepatic perfusion in the treatment of metastatic uveal melanoma. J Clin Oncol. 2015;33(15 suppl):e20000. is removed by the second-generation filter, thus mini- 14. Pingpank JF, Hughes MS, Alexander HR, et al. A phase III ran- mizing systemic exposure.19 Likewise, a general trend dom assignment trial comparing percutaneous hepatic perfusion with melphalan (PHP-mel) to standard of care for patients with hepatic me- is associated with the second-generation filter that has tastases from metastatic ocular or cutaneous melanoma. J Clin Oncol. demonstrated that the newer-generation filter is as- 2010;28(18):LBA8512. 15. Pingpank JF, Libutti SK, Chang R, et al. Phase I study of hepatic sociated with a decreased rate of hematological tox- arterial melphalan infusion and hepatic venous hemofiltration using per- icities, an increased number of treatments per patient, cutaneously placed catheters in patients with unresectable hepatic malig- nancies. J Clin Oncol. 2005;23(15):3465-3474. and a decreased rate of treatment-associated mortality 16. Pingpank JF, Royal RE, Kammula US, et al. High dose intra-arterial (see Table).7,8,10-13 melphalan delivered via percutaneous hepatic perfusion (PHP) for pa- tients with unresectable hepatic metastases from primary neuroendocrine tumors. Presented at: Americas Hepato-Pancreato-Biliary Association An- Conclusions nual Meeting; Fort Lauderdale, FL; March 27–30, 2008. 17. Lillemoe HA, Alexander HR. Current status of hepatic perfusion Results from a randomized trial and several retrospec- as regional treatment for patients with unresectable hepatic metastases: tive studies demonstrate that chemosaturation with a review. Am Oncol Hematol Rev. 2014:15-23. http://delcath.berkman- tech.com/workspace/media/bibliographies/Alexander%20article.pdf. percutaneous hepatic perfusion achieves good tumor Accessed January 13, 2017. response rates in patients with unresectable hepatic me- 18. Pingpank J, Royal RE, Kammula US, et al. Percutaneous hepatic perfusion (PHP) with melphalan for patients with unresectable liver metas- tastases from ocular or cutaneous melanoma, colorectal tases of neuroendocrine tumours (MNET)-NCT00096083. Eur J Cancer. cancer, or cholangiocarcinoma.7,8,10,11,13 In 1 study, che- 2011;47(suppl):S478-S478. 19. Burgmans MC, de Leede N, den Hartigh J, et al. Safety and ef- mosaturation with percutaneous hepatic perfusion was ficiency of the Delcath 2nd generation filter in percutaneous hepatic per- significantly associated with prolonged rates of hepatic fusion with melphalan for unresectabel hepatic metastases of colorectal cancer and uveal melanoma. Presented at: Cardiovascular and Interven- progression-free survival compared with 2 other region- tional Radiology Society of Europe Annual Meeting and Postgraduate ally delivered, liver-targeted therapies.11 More random- Course; Lisbon, Portugal; September 26–30, 2015. ized trials are needed to confirm these initial findings, to investigate the potential benefit in overall survival, and to study the use of chemosaturation with percuta- neous hepatic perfusion in combination with current immunotherapy-based treatments for melanoma. To this end, a phase 3, multicenter, international, randomized trial without a cross-over design that is comparing per-

January 2017, Vol. 24, No. 1 Cancer Control 101