1125 Regional Perfusion with Hemofiltration (Chemofiltration) for the Treatment of Patients with Regionally Advanced Cancer Mordechai Gutman, M.D.' Subhi Abu-Abid, M.D.' Patrick Sorkine, M.D? Moshe Inbar, M . o . ~ Dina Lev, M.D.' Zipora Chen, M . D . ~ Dan Oron, M.D." Samario Chaitchik, M . D . ~ Joseph M. Klausner, M.D.' ' Department of Surgery B&C, Sourasky Tel Aviv Medical Center, Sackler Faculty of Medicine, Tel Aviv, University, Tel Aviv, Israel. Department of Intensive Care, Sourasky Tel Aviv Medical Center, Sackler Faculty of Medicine, Tel Aviv, University, Tel Aviv, Israel. Department of Oncology, Sourasky Tel Aviv Medical Center, Center, Sackler Faculty of Medicine, Tel Aviv, University, Tel Aviv, Israel. Department of lnvasive Radiology, Sourasky Tel Aviv Medical Center, Sackler Faculty of Medicine, Tel Aviv, University, Tel Aviv, Israel. Supported by the Nicholas & Elizabeth Slezak Fund for Experimental Surgery, Tel Aviv University, Israel. Address for reprints: Joseph M. Klausner, M.D., Professor and Head of Surgery, Sourasky Tel Aviv Medical Center, 6 Weizmann Street, Tel Aviv 64239, Israel. Received July 21, 1995; revision received November 24, 1995; accepted February 20, 1996 0 1996 American Cancer Society BACKGROUND. Regionally advanced cancer is a common, often unresolved problem. Effective local control with chemotherapy is limited by the toxicity following systemic administration. Chemofiltration (CF) is a form of regional perfusion that enables the administration of cytotoxic drugs into one body area while limiting systemic toxicity. The drug is infused into the artery supplying the involved area. The venous effluent of the same organ is pumped out into a hemofiltration unit at a high flow rate. The drug is then filtered away and the blood returned to systemic circulation. METHODS. Forty-one patients underwent 45 CF. Twenty-four patients had CF of the pelvis for advanced rectal carcinoma (101, malignant melanoma (6),and cancers of the uterine cervix (3),ovary (2), vulva (11, endometrium ( l ) ,and anus (1).Seventeen patients underwent CF of the liver for metastatic colon (101, breast (41, pancreas (I), ovary (l),and unknown primary (1) cancer. 5-fluorouracil (1 g/m2) and mitomycin-C (30 mglm2); cisplatinum (200 mglm2) alone or combined with bleomycin (50 mg/m2) and mitomycin-C (20 mg/m2); or melphalan (1 mg/kg) were the combinations used. RESULTS. Generally the procedure was well tolerated. Complications included transient leukopenia (18), paralytic ileus (2), hair loss (21, renal failure (1). Two patients died within 40 days following CF. Of 36 evaluable patients, 16 (44%) had partial response, 14 (38%)had stable disease, and 6 (18%) had disease progression. A decrease of at least 30% in carcinoembryonic antigen levels occurred in 12 of 24 patients (50%). Median time to progression was 7 months. Ten of 13 patients (77%) achieved good symptomatic palliation. CONCLUSIONS. The results of CF in our study are not superior to alternative methods of drug delivery to the liver and pelvis. However, considering that previous systemic chemotherapy had failed two-thirds of the patients, some benefit may be attributed to this regional delivery modality. Furthermore, pelvic CF afforded very significant symptomatic relief which was definitely superior to other methods. Cancer 1996; 7 8 1125-30. 0 1996 American Cancer Society. KEYWORDS: regional perfusion, chemofiltration, regionally advanced cancer, liver metastasis, pelvic cancer. R egionally advanced cancer (RAC) represents a common yet often unresolved problem: patients with large, nonresectable recurrent or metastatic tumor confined to one region causing disabling symptoms and eventually death. With colorectal cancer, 65% of the patients develop liver metastases, this being the only metastatic site in half of them.' The liver can also be the sole site of metastasis in patients with cancers of the stomach, pancreas, breast, etc. The pelvis is another common site of RAC. More than 30% of patients with rectal cancer, 1126 CANCER September 1, 1996 / Volume 78 / Number 5 YlOTOXlC DRUG ULTRAFILTRATE FIGURE 1. Liver chemofiltration. The drugs are injected into an hepatic arterial catheter, which is placed at laparotomy or angiographically. A 16F catheter is inserted into the inferior vena cava, and connected to a hemofiltration unit. The detoxified blood and substitution fluid are infused back to the patient via a catheter in the superior vena cava. -5V85TITUTION FLUID I section, attempts at local control mainly include irradiation or chemotherapy. Radiotherapy is of some benefit especially with rectal cancer. Currently, systemic chemotherapy has a low therapeutic index since dosage is limited by toxicity to normal tissue^.^ To increase the therapeutic index of chemotherapy, regional perfusion techniques were developed. Intra-arterial administration of cytotoxic drugs (CD) has slightly increased the response rate. However, since the drug is still distributed systemically from the venous effluent of the infused organ and the dosage cannot be significantly increased, systemic toxicity is not Attempts to limit systemic toxicity by means of vascular isolation led to the development of isolated limb perfusion (ILP) which enables delivery of very high doses of CD with total vascular isolation of the perfused organ using tissue oxygenation by a heartlung machine. This method is especially suitable for cancer confined to the limbs, mainly malignant melanoma or sarcoma, since the main blood vessels can be easily isolated. As expected, the response rate is superior to systemic treatment and systemic toxicity is virtually eliminated.8 Obviously it is difficult to implement this technique in the trunk or abdominal viscera due to the complexity of their blood h up ply.^ Recently a method has been developed which enables the delivery of high dose CD to organs such as the liver or pelvis, while reducing systemic toxicity by concomitant detoxification of the venous effluent of the infused organ by hemofiltration.'0-'2 Presented is our initial experience with 41 cancer patients with RAC in the liver or pelvis who were treated by chemofiltration (CF1. MATERIALS AND METHODS FIGURE 2. Pelvic chemofiltration. The drugs are injected into a catheter placed angiographically in the aorta just belowthe renal arteries. Pneumatic tourniquets are used to occlude the blood flow to the limbs. The venous pelvic blood is pumped out to the hemofiltration unit. The detoxified blood and substitution fluid are infused back to the patient via a central venous catheter. as well as those with cancers of the urinary bladder or female genitalia, develop pelvic RAC.2,3These patients are usually highly symptomatic, with pain, tenesmus, and discharge being the most common complaints. Since most RAC are not amenable to surgical re- During the past 2 years, 41 patients underwent 45 CF. There were 26 females and 15 males. The average age was 56 years (range: 21-81 yrs). Seventeen patients had metastatic cancer confined to the liver, originating in the colorectum (lo), breast (41, pancreas (11, ovary (l),and unknown (1). Twenty-four patients had RAC confined to the pelvis, the origin being the rectum ( lo), lymph node metastases of malignant melanoma ( 6 ) )ovary (2) endometrium ( l ) ,uterine cervix (31, vulva ( I ) , and anus (1). Twenty-seven of the 41 patients (67%) had received previous systemic chemotherapy and were referred to CF due to treatment failure and tumor progression. Treatment Protocol Liver chemofiltration An hepatic arterial catheter was placed angiographically (10 patients) or during laparotomy (7 patients). Chernofiltration/Gutrnan et al. 1127 TABLE 1 Drugs Used in Chemofiltration Primary tumor Cytotoxic drugs No. of patients Colon-rectum Uterine, cenix and endometrium Malignant melanoma Mitomycin-C 30 mglm’ & 5-FU 1000 mglm’ Cisplatinum 200 mg/m2 & mitomycin-C 20 rnglm’ Melphalan 0.8-1.2 mglkg or cisplatinum 200 mg/mZ Melphalan 1.2 rnglkg, mitomycin-C 25 mglm’ & 5-FU 750 mglm’ Mitomycin-C 20 mglm’, bleomycin 50 mglm’ & cisplatinum 200 mglm2 Mitomycin-C 30 mglm2 & 5-FU 750 mglm‘ Mitomycin-C 30 mglm’, cisplatinum 100 mglm’ & 5-FU 1500 mglm’ Mitomycin-C 30 rnglm’ & 5-FU 750 mglm’ Mtomycin-C 30 mglm’ & 5-FU 1000 mglm* 20 Ovary Vulva Breast Anus Pancreas Unknown The tip of the arterial catheter was situated distal to the bifurcation of the gastroduodenal artery to avoid infusion of the duodenum and jejunum. A large bore (16F) catheter was inserted into the inferior vena cava via a cutdown in the saphenous vein (13 patients), femoral vein (31, or directly during laparotomy (1). Using fluoroscopic guidance, the tip of the catheter was positioned at the level of the diaphragm distal to the entrance of the hepatic veins. The venous catheter was then connected to a Gambro hemofiltration unit (Gambro Instruments, Lund, Sweden) using a 1.2 m2 hollow tube filter (Fig. 1). A balanced hemofiltration was established at a rate of 500 to 750 mL per minute. The detoxified (filtered) blood was infused to the superior vena cava via a large bore (Vygon 8F) central venous catheter. It was supplemented by Ringer’slactate solution (preheated to 38 “C) infused at the same rate as the efflux of the ultrafiltrate, thus assuring an essentially zero balance. The anesthetized, fully heparinized patients were carefully monitored, and further fluid replacement was given to maintain normal hemodynamics. Once stable extracorporeal hemofiltration was established, the CD was infused into the arterial catheter using an automatic infusion pump over 10 to 30 minutes, depending on the drug. After CD administration, CF was continued for an additional 30 minutes to ascertain further CD clearance. Pelvic chemofiltration An arterial catheter was angiographically placed in the aorta just distal to the renal arteries. The venous large bore catheter was inserted into the inferior vena cava via a saphenous cutdown (22 patients) or femoral venotomy (2 patients), its tip reaching the same level as the arterial catheter. Blood flow to both lower limbs was occluded using a pneumatic tourniquet. In melanoma patients, when metastases were confined to one leg and the pelvis, the pneumatic tourniquet was ap- 4 5 1 3 1 4 1 1 1 plied only to the contralateral leg. The remaining details are identical to liver CF (Fig. 2). Drugs Cytotoxic drug combinations were mostly based on high dose mitomycin-C. They included mitomycin-C and 5-fluorouracil (5-FU) for cancer of the colon, rectum, breast, pancreas, and ovary, and a combination of bleomycin, cisplatinum, and mitomycin-C for vulvar cancer. Phenylalanin-mustard (melphalan),at the same dose used for isolated limb perfusion (0.8-1.2 mg/kg), or high dose cisplatinum were used for malignant melanoma. The doses are detailed in Table 1. Response Criteria All patients had measurable lesions. The lesions were measured prior to, and 4, and 8 weeks following CF. Complete response (CR) was defined as the disappearance of all known disease determined by 2 observations not less than 4 weeks apart. Partial response (PR) was defined as a 250% decrease in total tumor load of the lesions lasting > 4 weeks.I3 RESULTS Forty-one patients successfully completed CF of the liver (17) or pelvis (24). Two patients underwent an additional CF of the liver (1)and pelvis (11, and a third patient had two additional liver CF. A filtration flow rate of at least 500 mL per minute was achieved, with an average ultrafiltration of 11,500 mL (9400- 16000 mL). The extensive fluid exchange was well tolerated in all but two patients; one developed hypotension and hypoxemia at the beginning of the procedure which was terminated before injection of CD. The patient was offered a second attempt to be performed after extensive hydration, but he refused and therefore is not included in this series. A second patient developed extreme bradycardia (20/min) and hypotension. 1128 CANCER September 1, 1996 / Volume 78 / Number 5 The procedure was suspended and reperformed successfully 48 hours later, after assuring intact cardiac and neurologic function. Two patients died within 40 days following CF. The first, a male age 68 years, who had multiple liver metastases of rectal cancer with marked weight loss and severely impaired liver function, developed pneumonia with respiratory insufficiency which led to his death 28 days after the procedure. The second was a female, age 55 years, with a large, nonresectable pancreatic cancer and multiple liver metastases. During CF, CD were infused to the common hepatic artery to deliver CD to the pancreatic head via the gastroduodenal artery. Afterwards, the patient had severe gastrointestinal (GI) toxicity and developed respiratory failure necessitating prolonged ventilation, leading to her death after 40 days. Complications attributed to the surgical procedure were observed in 9 of 45 CF (20%)and included wound hematoma (41, infection (21, deep vein thrombosis (11, pulmonary embolism (11, and pneumothorax (1). Complications attributed to CD were observed in 22 of 45 CF (48%) and included mild leukopenia (>lo00 mmi) in 9 patients, severe (<lo00 mm3) in another 7, and thrombocytopenia in 1. Of the 16 patients who developed leukopenia, 15 (94%) had received previous systemic chemotherapy. Two patients developed a prolonged paralytic ileus after liver CF. Partial hair loss was observed in another two patients, and one, who received cisplatinum in two sequential pelvic CF, developed moderate renal failure. Measurable disease could be evaluated in the follow-up of 36 of the 41 patients. Of the remaining five nonevaluable patients, two died, another had his entire tumor mass (pelvic malignant melanoma) resected, and in two, insufficient time elapsed. There was no CR. A PR was observed in 16 of 36 patients (44%).Time to progression in this group was 7 months (3-22 mos). Fourteen patients (39%)showed stabilization of their disease (SD). Six (16%) experienced disease progression (PD). Of 14 evaluable patients who underwent liver CF, 7 (50%)had PR and another 5 (37%)SD. Of 22 patients who underwent pelvic CF and had a measurable tumor, 9 (41%) had PR and 9 (41%) SD (total 82%). PR was notable in one patient with an inoperable rectal cancer who underwent CF as a first line treatment. The tumor’s size was markedly reduced and she underwent an abdominoperineal resection 4 weeks after CF. The patient is disease-free 20 months afterwards. PR was demonstrated also in two of four melanoma patients, and in each of the patients with ovarian, cervical, or vulvar cancer. Marked symptomatic relief occurred in 10 of 13 (77%) highly symptomatic patients, 2 of whom underwent pelvic CF despite pulmonary metastases due to huge pelvic masses, causing agonizing pain, discharge, and tenesmus. In both cases, only SD was achieved but quality of life was significantly improved. DISCUSSION The main theoretical advantage of CF is the ability to deliver higher doses of CD directly to the tumor, while the concomitant hemofiltration limits side effects to an acceptable range. The close relationship between dosage and response rate has long been established, and even a small increment in dose may result in a significant improvement in the response rate.14-” When relatively low doses are administered, drug resistance and DNA repair commonly occur. This phenomenon has been demonstrated in a variety of CD, especially alkylating agents and antibiotics, using in vitro assays and in human studies.’8-20 Indeed, CF achieved high local concentrations of CD and although pharmacodynamic measurements in the infused organ, systemic blood, and ultrafiltration were not performed in the current study, it has been well demonstrated that CD levels in the perfused tissues during CF were 4 to lox greater compared with other nonperfused This would explain the feasibility of administering mitomycin-C at a dose of 30 mg/m2 which is about 3-fold the usual dose” or melphalan which is rarely administered systemically due to its toxicity. The high doses of CD delivered in CF led to a response rate of 44% (PR), and if SD are included, 83%,in patients with advanced regional cancer, many of whom (67%) experienced disease progression after previous systemic chemotherapy. Although inferior to ILP, where a 6 to 20-fold increase in the administered CD dosage leads to a 10 to 20fold increment in tissue CD levels, and hence a much greater response CF still demonstrates some advantages over conventional systemic therapy. The usual systemic dosage of mitomycin-C is 10 mg/m2.Even at this dosage it is considered to be toxic and therefore its use is limited. It is considered one of the drugs of choice in the regional setting and indeed was the base for the treatment combination in 35 of the 41 patients in our series. Only those patients with melanoma did not receive mitomycin-C. The precise influence of 5-FU remains speculative but given the fact that most of the patients in this series had previously failed with 5-FU based therapy, it can be assumed that the addition of another drug, i.e., mitomycin-C, was of major impact. Furthermore, the dosage of mitomycin-C was two to three times greater than the usual systemic one. ChemofiItration/Gutman et al. The most impressive effect of CF was the palliative one observed among the pelvic CF patients. Seventyseven percent of patients suffering from disabling symptoms, such as pain, secretions, and tenesmus reported a significant improvement in their quality of life. This effect was not related to the magnitude of the objective response, leading us to conclude that it was not necessarily due to a cytotoxic response of CF, but rather to a plausible form of neurotoxicity. However, CF does have several disadvantages: it is an invasive procedure, associated with general anesthesia, angiography, operation, heparinization, and massive fluid exchange. Despite filtration of the CD, related complications are still encountered. Since CF is a semiclosed technique, unlike isolated limb perfusion which enables total vascular isolation, one cannot expect to pump out all the venous effluent from the liver or the pelvis into the hemofiltration unit, and therefore a significant portion of the infused CD “leaks” into the systemic circulation. Furthermore, serum proteins may bind to the CD and due to their large molecular weight are not filtered thus contributing to systemic toxicity. Previous studies estimated the “leakage” to the systemic circulation at 30 to 70%.” Considering the fact that the infused dose of CD is sometimes 3fold the conventional one, even a 30% leak can be significant. Hemofiltration is most effective in recapturing CD which are of small molecular weight and low ability to bind to serum proteins. Fortunately, many CD are in this category, and as shown by Aigner et al.,’”commonly used CDs, such as mitomycin-C, 5-FU, cisplatir u m , and adriamycin are suitable for the procedure. Antimetabolites such as methotraxate and actinomycin-D are considered to be less suitable for CF. The 44% response rate which consisted of PR is not better than that obtained by other locoregional methods. Coupled with the relatively short time to progression (7 m), it seems that CF is not effective enough. There are several factors which could explain the lack of significant improvement in response rate to be expected with a 2 to 3-fold increase in dose. Infusion of high dose CD into an organ does not guarantee high tumor concentrations. There are two major obstacles to the distribution of drugs in a tumor. One is the nonuniform blood supply resulting in areas of hypoperfusion within the tumor.24The other is the abnormally high interstitial pressure within tumors, measuring up to lox more than the adjacent normal l i s ~ u e sThe . ~ ~efficacy of CD which to penetrate tumor cells may also be hindered by the development of drug resistance in patients previously treated with systematically administered CD. The results of CF in our study were not superior 1129 to alternative methods of drug delivery. On the surface, CF might not be justified. However, it should be emphasized that these results were obtained in a selected group of patients, two-thirds of whom failed systemic delivery of chemotherapy. In addition, CF is a single procedure compared with prolonged systemic treatments. Based on our results, we cannot state that CF is superior to other methods. However, its palliative effect, especially in pelvic cancer patients, merits its use in selected cancer patients. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Bengmark S, Hofstrom L. The natural history of primary and secondary malignant tumors in the liver: 1. The prognosis for patients with hepatic metastases from colonic and rectal carcinoma by laparotomy. Cancer 1969;23:198-202. Dobrowsky W, Schmid AP. Radiotherapy of presacral recurrence following radical surgery for rectal carcinoma. Dis Colon Rectum 1985;28:917-9. Wanebo HJ, Whitehill R, Gaker DL, Wang GJ, Morgan R, Constable W. Pelvic recurrence of rectal cancer. Ann Surg 1987;205:482-95. DeVita VT. Principles of chemotherapy. In: DeVita VT, Hellman s, Rosenberg SA, Editors. Cancer, Principles & Practice of Oncology. Philadelphia: JB Lippincott Co., 1989:276-300. Chang AE, Schneider PD, Sugarbaker PA. A prospective randomized trial of regional versus systemic continuous 5-Fluorodeoxyuridine chemotherapy in the treatment of colorectal liver metastases. Ann Surg 1987;201:685-93. Balch CM, Urist MM, McGregor ML. Continuous regional chemotherapy for metastatic colorectal carcinoma using a totally implantable infusion pump. Am J Surg 1983;145:285-90. Ensminger WD, Rosowsky A, Raso V, Levin DC, Glade M, Come S, et al. A clinical pharmacological evaluation of hepatic arterial infusions of 5-Fluoro-i-deoxyl1ridine and 5Fluorouracil. Cancer Res 1978;38:3784-92. Muchmore JH, Carter RD, Krementz ET. Regional perfusion for malignant melanoma and soft tissue sarcoma: A review. Cancer Invest 1985;3:129-42. Krementz ET. Regional perfusion: Current sophistication, what next? Cancer 1986;57:416-32. Aigner KR, M d e r H, Walther H, Lurk KH. Drug filtration in high dose regional chemotherapy. In: Eckhardt S, Holzner JH, Nagel GA, editors, Contributions to Oncology: Basel, Karger, 1988. Muchmore JH, Krementz ET, Carter D, Preslan JE, George W. Treatment of abdominal malignant neoplasms using regional chemotherapy with hernofiltration. Arch Surg 1991; 126:1390-6. Fiorentini G, Davitti B. Tienghi A, Guglielminetti D, Sanguinetti F, Carosi V. et al. Intra-arterial-hepatic chemotherapy alone and combined with intra-peritoneal chemotherapy associated with central venous drug filtration: a two step pilot study. Reg Cancer-Trent 1988; 1:28-32. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Caiicer 1981;47:207- 14. Frei E 111, Canellos GP. Dose: A critical factor in cancer chemotherapy. Am 1 Med 1980;69:585-94. Pinkel D, Hernandez K, Borella L, Holton C, Aur R, Sanioy G, et al. Drug dosage and remission duration in childhood lymphocytic leukemia. Cancer- 1971;27:247-56. 1130 CANCER September 1,1996 I Volume 78 I Number 5 16. DeVita VT. Dose-response is alive and well. J Clin Oncol 1986;4:1157-9. 17. Hyrniuk WM, Bonadonna G, Valagussa P. The effect of dose intensity in adjuvant chemotherapy. In: Salmon SE. Adjuvant Therapy of Cancer V. Orlando: Grune and Statton, 1987:13-23. 18. Crooke ST, Bradner WT. Mitomycin-C: A review. Cancer Treat Rev 1976;3:121-39. 19. Dorr RT. New findings in the pharmacokinetic, metabolic and drug-resistance aspects of mitomycin-C. Semin Oncol 1988; 1 5 ( S ~ p p4):32-41. I 20. Schroy DC 111, Cohen A, Winawer SJ, Friedman EA. New chemotherapeutic drug sensitivity assay for colon carcinomas in monolayer culture. Cancer Res 1988;48:323644. Weinberg MC, Leenen 21. Turk PS, Belliveau JF, Darnowski JW, L, Wanebo JH. Isolated pelvic perfusion for unresectable cancer using a balloon. Occlusion technique. Arch Surg 1993; 128:533-9. 22. Lejeune FJ, Lienard D, El Douaihy M, Vadoud Seyedi J, Ewalenko P. Results of 206 isolated limb perfusions for malignant melanoma. Eur J Surg Oncol 1989; 15:510-9. 23. Minor DR, Allen RE, Alberts D, Peng Y-M, Tardelli G, Hutchinson J. A clinical and pharmacokinetic study of isolated limb perfusion with heat and melphalan for melanoma. Cancer 1985;552638-44. 24. lain RK. Determinants of tumor blood flow: A review. Cancer Res 1988;48:2641-58. 25. lain RK. Transport of molecules in the tumor interstitium: A review. Cancer Res 1987;47:3039-51.