Surg Today Jpn J Surg (1996) 26:305-313
SURGERYTODAY © Springer-Verlag 1996
Original Articles
Percutaneous Hepatic Venous Isolation and Extracorporeal Charcoal Hemoperfusion for High-Dose Intraarterial Chemotherapy in Patients with Colorectal Hepatic Metastases YONSON KU, l MASAHIROTOMINAGA,1 TAKESHIIWASAKI,1 TETSUSHIKITAGAWA,1 ICHIROMAEDA,~ MASAFUMINHIOTANI,~ NHINYAKUSUNOKI,I YOKO MAEKAWA,1 MASAHIRONAMIZO,1 TAKUMIFUKUMOTO,1 YOSHIKAZU KURODA,~ SHOZO HIROTA,a and YOICHI NAITOHI ~First Departmentof Surgery, and 2Department of Radiology,Kobe UniversitySchoolof Medicine,7-5-2 Kusunoki-cho,Chuo-ku, Kobe 650, Japan
Abstract: The results of treating 12 consecutive patients with unresectable colorectal hepatic metastases with a hepatic arterial infusion of high-dose Adriamycin, 100-120 mg/m 2, using hepatic venous isolation (HVI) and charcoal hemoperfusion (CHP) are reported herein. Adriamycin was administered over 5-15 rain under extracorporeal drug elimination by HVICHP. HVI was percutaneously accomplished by either the double-balloon technique using a Fogarty occlusion catheter (8/22F) or a balloon-tipped catheter (16F). During the infusion, isolated hepatic venous blood was filtered by CHP and pumped into the left axillary vein. There were no lethal complications, and good hemodynamic tolerance to HVI-CHP was confirmed. Tumor liquefaction accompanied by a sharp decrease in serum carcinoembryonic antigen levels by more than 50% of pretreatment levels was observed in 6 of the 12 patients 1 month after treatment. Apart from chemical hepatitis, which developed in 11 (92%) of the patients, the Adriamycin toxicities were well controlled following the development of nausea and vomiting in 2 patients (17%), leukopenia <2,000/mm 3 in 3 (25%), and gastric ulcer in 1 (8%). These results indicate that this method is a safe and useful procedure for otherwise hazardous high-dose intraarterial chemotherapy in patients with unresectable hepatic tumors. Key Words: hepatic venous isolation, charcoal hemoper-
fusion, high-dose chemotherapy, metastasis
colorectal
hepatic
Introduction
The low response rates of colorectal hepatic metastases to standard doses of systemic and regional chemotherapy have urged several researchers to try to develop other effective therapeutic modalities. A recent study
by Left et al.~ using a single infusion of melphalan with bone marrow rescue indicated that high-dose treatment yields the greatest response in less sensitive colorectal tumors. Although a steep relationship between dose and response has been demonstrated both in vivo and in vitro for cytotoxic agents used in cancer chemotherapy,; the dosage is generally limited by systemic toxicities. In view of this, various attempts such as hemofiltration3 and charcoal hemoperfusion (CHP) under nonisolated venous access 4 have been combined with high-dose intraarterial chemotherapy to control the adverse effects. Unfortunately, the clinical results of these methods have been unsatisfactory due to their limited efficacy in the prevention of systemic toxicitieso Aigner et al. 5 previously reported that isolated liver perfusion (ILP) with cytotoxic agents would further enhance the advantages of regional chemotherapy; however, this method has not yet gained wide acceptance, mainly because of its technical complexities. As an alternative to these techniques, we have developed a novel system consisting of percutaneous hepatic venous isolation (HVI) and extracorporeal CHP. Previous experimental studies have proven the feasibility and efficacy of HVI-CHP for reducing the systemic distribution of anticancer agents including Adriamycin, mitomycin C, 6 and cisplatin7 during hepatic arterial infusion. Accordingly, since May, 1989, we have applied this method clinically in the treatment of patients with unresectable hepatic tumors including hepatocellular carcinoma8-10 and colorectal metastases. Herein we report the results of administering this treatment to 12 patients with colorectal hepatic metastases. Patients and Methods
Patients"
Reprint requests to: Y. Ku (Received for publication on Mar. 30, 1995; accepted on Nov. 7, 1995)
A total of 12 consecutive patients with colorectal hepatic metastases were included in this study. The char-
306
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
Table 1. Clinical summary of the 12 patients who received high-dose chemotherapy under HVI-CHP
Patient no.
Liver metastases (Number and distribution)
Age (years)
Sex
Primary tumor site
1
62
F
Ascending colon
Synchronous (Multiple unilateral)
2
61
M
Sigmoidcolon
3
42
M
Rectum
4
50
M
Rectum
5
64
M
Sigmoidcolon
6
61
F
Ascending colon
7
54
M
Sigmoidcolon
8
52
M
Sigmoidcolon
9
52
M
Rectum
Synchronous (Multiple bilateral) Synchronous (Multiple bilateral) Metachronous (Multiple bilateral) Synchronous (Multiple bilateral) Synchronous (Multiple bilateral) Synchronous (Multiple bilateral) Synchronous (Multiple bilateral) Synchronous (Multiple bilateral)
10
51
M
Transverse colon
Synchronous (Multiple bilateral)
11
62
F
Rectum
12
54
M
Ascending colon
Synchronous (Multiple bilateral) Synchronous (Multiple bilateral)
Extrahepatic disease
Time interval (months) of colorectal resection to high-dose chemotherapy under HVI-CHP
Abdominal lymph nodes and direct invasion to the duodenum Abdominal lymph nodes
Simultaneous
Abdominal lymph nodes and pulmonary metastases Abdominal lymph nodes
7
Abdominal lymph nodes and pulmonary metastases Abdominal lymph nodes
2
17 1 l
Abdominal lymph nodes
21
Abdominal lymph nodes
Simultaneous
Abdominal lymph nodes and direct invasion to the prostate gland and the bladder Abdominal lymph nodes and direct invasion to the pancreas Abdominal lymph nodes
12
Abdominal wall
HVI-CHP, hepatic venousisolationand charcoalhemoperfusion
acteristics and clinical features of these patients are summarized in Table 1. All of the patients except one (patient no. 4) had synchronous hepatic metastases and all had multiple metastases in the liver. Pulmonary metastases were also discovered in two patients during follow-up. Simultaneous treatment with high-dose chemotherapy under HVI-CHP was administered to 2 of the 11 patients with synchronous hepatic metastases during the initial operation for primary colorectal cancer, with concomitant cholecystectomy. In the remaining patients, chemotherapy was performed 1-21 months after the primary surgery. Postoperative examinations and follow-up studies included electrocardiogram (ECG), physical examinations, full blood counts, and liver function tests. These tests were performed at least every 2nd day during the 1st postoperative week, twice a week during the 2rid week, and once a week during the 3rd and 4th weeks. Thereafter, the tests were performed monthly. Computed axial tomographies (CT) of the liver, and serum carcinoembryonic antigen (CEA)
tests, were performed every 1-2 months after the treatment. Informed consent was obtained from all patients, and the study protocol was approved by the regional ethical committee of Kobe University Hospital.
Hepatic Arterial Catheter Placement Of the 11 patients with synchronous hepatic metastases, 2 underwent abdominal exploration for simultaneous resection of the primary lesion. In the first patient (patient no. 1), a hepatic arterial catheter was placed prior to the operation using te Seldinger technique under fluoroscopic guidance, because the right hepatic artery originated from the superior mesenteric artery and supplied most of the metastatic tumors. The other patient (patient no. 8) underwent intraoperative catheter placement through the gastroduodenal artery. In the remaining patients, the catheter was inserted into the proper hepatic artery using the Seldinger technique under fluoroscopic guidance just prior to HVI-CHP. In four
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
307
patients (patient nos. 6, 7, 10, and 11), a balloon-tipped catheter (Interventional balloon catheter, 5F, Herstellung und Vertrieb Medizinischer Produkte, Germany) was inserted using the Seldinger technique to perform balloon-occluded arterial infusion ~1,12 of the drug. The other patients underwent embolization of the gastroduodenal artery with steel coils (Occluding spring emboli, A. Cook Group, USA) during hepatic arterial catheter placement.
Anesthetic Management All patients were positioned on a heating blanket set at 38°C. After induction of anesthesia by thiamylal and vencuronium, mechanical ventilation was begun with a nitrous oxide/oxygen mixture (60%/40%). Intermittent boluses of pentazocine were given when required. Immediately before HVI-CHP, the patient was rapidly given 10ml/kg of lactated Ringer's solution along with 500ml of 4.4% albumin. No inotropic agent was necessary throughout the procedures. The patients were monitored with ECG and radial artery pressure during anesthesia. Central venous pressure and cardiac output were also monitored using a Swan-Ganz catheter in five patients. All patients were extubated soon after completion of the surgical procedure, and intravenous 5% dextrose was infused during the first 12h. Antibiotics were given after treatment and oral intake was allowed the following morning.
Surgical Technique for HVI-CHP HVI-CHP was performed in the operating room in all patients. The extracorporeal system is schematically represented in Fig. 1. With the patient in the supine position and the left arm extended, bilateral groin incisions are made to explore each saphenofemoral junction. At the same time, the left axillary fossa is explored to access the left axillary vein. Once vascular access is established, anticoagulation is achieved by the intravenous administration of heparin at a dose of 100U/kg. First, a Fogarty occlusion catheter (Model 62-080-8/22F, Fogarty, American Edwards Laboratories, Santa Anna, CA, USA) is introduced into the right atrium through a major tributary of the right saphenous vein. A catheter (16F) to drain blood from the distal inferior vena cava is then placed through the right saphenous vein to the confluence of the common iliac veins. Finally, a specially-made balloon-tipped catheter (16F), 60cm in length, with an inflatable balloon at the distal end, the latter being 3 cm in maximum diameter, is introduced into the retrohepatic inferior vena cava through the left greater saphenous vein. Venography is then performed under transient occlusion of the cavoatrial junction to
Fig. 1. Schematic representation of the double-balloon technique used for administering high-dose intraarterial chemotherapy combined with charcoal hemoperfusion (CHP) under hepatic venous isolation (HVI). HVI is provided by balloon occlusions of the retrohepatic and suprahepatic inferior vena cava. Hepatic venous and total bypass flows are monitored by electromagnetic flowmeters at the points described above (F1 and F2). Arrows indicate the direction of blood flow in the circuit. The hepatic artery infusion (HA1) of Adriamycin was performed at dosages of 100-120mg/m 2 given over 5-15 min
determine the position of the retrohepatic catheter balloon. During venography, the Fogarty balloon in the right atrium is inflated with 30 ml of half-strength contrast agent and wedged at the cavoatrial junction by downward traction. Soon after occlusion, 50ml of fullstrength contrast agent is rapidly injected through the central lumen of the balloon-tipped catheter. A singleexposure venogram is taken and the occlusion is freed. Going by the venogram, the retrohepatic balloon is positioned between the short hepatic veins and the renal veins. The balloon-tipped catheter is connected to four CHP cartridges (DHP-1, Kuraray, Kurashiki, Japan) positioned in parallel, while another catheter is placed in the distal inferior vena cava connecting directly to the centrifugal pump (Biopump-80, Bio-Medicus, Minneapolis, MN, USA). A return catheter (16F) is placed in the left axillary vein and connected to the Biopump. All
308
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
of the tubing,the inside diameter of which ranges from 3/8-1/4 in, is made of polyvinyl chloride (Tygon, Norton Industrial Plastics, USA). Electromagnetic probes (Bioprobes DP-38 and DP-38P, Bio-Medicus) are used to monitor hepatic venous flow and total bypass flow as indicated by F1 and F2, respectively, in Fig. 1. The circuit is primed with 800ml of lactated Ringer's solution. After starting extracorporeal circulation via the CHP filter-excluded shunt (route A), the retrohepatic inferior vena cava is occluded by inflating the retrohepatic balloon with 15-20ml of normal saline. Subsequently, the Fogarty balloon is inflated with 30ml of halfstrength contrast agent while in the right atrium, and wedged at the cavoatrial junction to establish HVI. The position and contour of the balloon are intermittently monitored by fluoroscopy during the procedure. When hemodynamic stability is assured, the direction of flow, from the liver to the Biopump, is changed from route A to route B, which contains the CHP filters. Finally, the correct position of the arterial catheter and the selective perfusion of the liver are ensured under fluoroscopy, and the anticancer agents are administered via the hepatic arterial catheter. In this study, Adriamycin at doses ranging from 100 to 120mg/m 2, dissolved in 50-150ml of saline solution, were continuously infused over 5-15min, following which the HVI-CHP was maintained for 20-30±in. After completion, the vascular occlusions at the retrohepatic and suprahepatic inferior vena cava are released and the venovenous bypass is discontinued. After decannulation, the bilateral greater saphenous veins are ligated and the left axillary vein is repaired with a running suture using 5-0 prolene.
Statistical Analysis Paired and unpaired t-tests were used for statistical analysis of the hemodynamic results, which are expressed as means _+ the standard deviation (SD). A P value of less than 0.05 was considered statistically significant.
Results
Hemodynamic Data All patients had a smooth stepwise introduction of highdose chemotherapy under HVI-CHP, the hemodynamic data of which are listed in Table 2. The total bypass flow and the hepatic venous flow were remarkably stable throughout the procedures, the values during HVI-CHP being 1,467 _+ 288 and 470 +_ 155 ml/min, respectively. The fluid volume loading immediately prior to bypass resulted in a slight increase in the central venous pressure during the simple bypass period; however, the bypass alone did not affect the mean arterial pressure, heart rate, or cardiac index. After initiating HVI-CHP, the cardiac index and central venous pressure tended to decrease, whereas the mean arterial pressure and heart rate remained at similar levels. After the completion of bypass, the cardiac index promptly returned to the pretreatment level. The decrease in body temperature following bypass was about 1°C.
Tumor Response and Survival Changes in the serum CEA levels of individual patients are illustrated in Fig. 2. The CEA values were high in all
Table 2. Hemodynamic responses to HVI-CHP Before bypass (n)
Bypass without HVI-CHP (n)
Bypass with HVI-CHP (n)
End of bypass (n)
Total bypass flow (ml/min)
--
1,452 ± 273
1,467 ± 288
--
H e p a t i c v e n o u s flow (ml/min)
--
--
470 -- 155
--
36.5 +_ 0.5
B o d y t e m p e r a t u r e (°C) H e a r t rate ( b e a t s / m i n ) M e a n arterial p r e s s u r e ( m m H g ) Central venous pressure (mmHg) Cardiac i n d e x (1/min/m 2)
(12)
(12)
(12)
36.5 ± 0.5
35.5 ± 0,3"
(12)
(12)
(12)
(12)
90.3 _+ 18.8
84.3 -+ 16.2
86.2 ± 21.5
91.7 ± 17.3
(12)
(12)
(12)
(12)
96.5 +_ 8.3 (12)
95.2 ± 11.0 (12)
94.7 ± 18.5 (12)
103.9 ± 22.3 (12)
5.2 ± 1.8
7.2 ± 3.6
5.6 ± 2.4
5.8 ± 2.9
3.38 ± 0.96
3.47 ± 0.76
2.93 ± 0.66
3.50 ± 1.26
(5)
(5)
HVI-CHP, hepatic venous isolation and charcoal hemoperfusion Values are expressed as means 2 SD. * P < 0.05 versus I, II
(5) (5)
(5)
(5)
35.4 ± 0.2*
(5)
(5)
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
309 to progression of the hepatic disease. Thus, 3 patients remain alive and stable, but with persistence of disease, 13, 9, and 7 months after treatment, respectively.
400 < "'
300
I ',183)
Adverse Effec~.s
:~
200
t-< uJ ~
~2 (9~) 4 (91) 2 (1263)
I00
9 (7) 6 (31)
The adverse effects observed are summarized in Table 4. All patients tolerated the surgical procedure without developing any lethal complications. However, macroscopic hemolysis and hematuria occurred immediately after treatment in 10 of 12 patients. Figure 4 illustrates the changes in laboratory results in individual patients. The platelet count transiently decreased in most patients immediately after HVI-CHP, and a decrease of less than 50,000/ram 3was observed in one patient on the 2rid posttreatment day. Although the HVI-CHP itself did not affect the leukocyte count in any patient, during the 2nd week three patients demonstrated a leukocyte count of less than 2,000/mm3, which returned to within the normal range within 1 week. Chemical hepatitis
0 I-Z
L~J n~ ~_ ~_ 0 ~J L9 < ~Z
'"
11 (695)
50
~r ~J ~_
10 (;8]2)
0
1
2
3
4
5
6
MONTHS POSTTREATMENT
Fig. 2. Percentage changes in posttreatment carcinoembryonic antigen (CEA) levels compared with the pretreatment levels. Values in parentheses indicate serum CEA levels before treatment in individual patients patients before treatment, and over 100ng/ml in 7 of the 12 patients, the normal value being <5.0ng/ml. However, 1 month after the treatment, 6 of the 12 patients showed a sharp decrease in CEA levels by between 57% and 93% of their pretreatment level. Although this value tended to rise again in most patients, two patients showed consistently low levels during the first 6 months. The posttreatment CT findings correlated well with the CEA results (Table 3). In patients showing a fall in the CEA value to below 50% of the pretreatment level, a significant density reduction in the hepatic tumors was seen on the posttreatment CT scan, as demonstrated in Fig. 3. In these patients, the tumor became well defined in contour. Table 3 outlines the survival and causes of death in the 12 patients. The inedian follow-up from the time of diagnosis of the hepatic metastases was 13 months, or 8 months from the date of high-dose chemotherapy under HVI-CHP. Of the 9 deaths, one patient committed suicide 10 months after treatment for reasons unrelated to cancer; 3 patients died from extrahepatic disease progression, being intestinal obstruction in 1, respiratory failure following intestinal hemorrhage in 1, and obstructive jaundice due to extrahepatic abdominal tumors in 1. The remaining 5 deaths were directly related
Fig. 3A,B. Plain computed tomography (CT) scans of patient no. 5 A before treatment and B 2 months after treatment. Note that multiple low-density tumors clearly appeared after treatment
310
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
Table 3. Survival and causes of death after high-dose chemotherapy under HVI-CHP
Dose (mg/m 2) of Adriamycin
BOAI
CT findings of liver metastases (1 to 2mo after treatment)
1
100
No
Liquefaction
2
100
No
No change
3
120
No
Liquefaction
4
110
No
No change
5
110
No
Liquefaction
6
110
Yes
Liquefaction
7
110
Yes
Liquefaction
8
110
No
Liquefaction
9
100
No
No change
10
110
Yes
Liquefaction
11
100
Yes
No change
12
100
No
No change
Patient no.
Status (time from treatment) Died (6mo) Died (5 too) Died (7 mo) Died (10mo) Died (11 mo) Died (16 too) Died (6too) Died (5 mo) Died (8 mo) Alive with disease (13 mo) Alive with disease (10too) Alive with disease (7too)
Cause of death Intestinal obstruction Progression of disease Respiratory failure Suicide Progression of disease Progression of disease Obstructive jaundice Progression of disease Progression of disease --
Survival from diagnosis of liver metastases (mo) 7 7 12 13 13 17 27 6 19 18
--
10
--
9
BOAI, balloon-occluded arterial infusion of adriamycin; HVI-CHP, hepatic venous isolation and charcoal hemoperfusion; too, months
AST 50
1
[
2000-
TOTAL BILIF1UBIN
5.0
4O
4.0
3O
3.0
E x
1000-
~- zo
2.0
k
10
0 PRE
1.0
1-~D
2W
~PFIE
8-1~1D
2W
0" PRE
1-~D
1W
0 PFIE
/
1-3D
2W
Fig. 4. Changes in laboratory results pre- and posttreatment. Each line represents one patient. Patients treated with balloon-occluded arterial infusion are represented by open circles, and patients without, by solid circles
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
311
Table 4. Adverse effects
Adverse effects HVI-CHP-related Thrombocytopenia <50,000/mm3 Macroscopic hemolysis/hematuria Adriamycin toxicities Hematologic Leukocyte count <2,000/mm3 Gastrointestinal Chemical hepatitis Sclerosing cholangitis Nausea/vomiting Ulcer/gastritis Hair loss Renal Creatinine elevation >25% Cardiac ECG abnormalities Catheter-related Thromboembolism Intimal tears
No. of patients (n = 12)
(%)
1 10
(8) (83)
3
(25)
11 0 2 1 3
(92) (0) (17) (8) (25)
0
(0)
0
(0)
1 0
(8) (0)
HVI-CHP, hepatic venous isolation and charcoal hemoperfusion; ECG, electrocardiogram
resulting from the high-dose infusion of Adriamycin was found in almost all patients, as indicated by elevations in the aspartate aminotransferase (AST) level, ranging from 2 to 26 times above the baseline. The AST reached its peak on the 2nd day in most patients, thereafter decreasing to the pretreatment levels within 1 week after treatment. Apart from one patient who showed a moderate AST elevation before treatment, the AST elevation appeared to be more marked in those patients treated with balloon-occluded arterial infusion of the drug. The total bilirubin levels were elevated in most patients during the 1st week, which was directly related to a rise in the indirect bilirubin due to hemolysis caused by HVI-CHP. None of the patients showed elevated bilirubin levels to indicate sclerosing cholangitis throughout the follow-up period. One patient developed a gastric ulcer which was proven by endoscopy. No sign of cardiac toxicity, as evidenced by hypotension or ECG abnormalities, was observed in any patient.
Discussion
The use of ILP to accomplish the short-term perfusion of anticancer agents at high dosage levels has been proposed by several investigators.~,~3,14 However, despite its potential efficacy, ILP is a cumbersome procedure requiring exposure and cannulation of all the major vessels adjacent to the liver. Furthermore, an extracorporeal heart-lung circuit is essential to establish corn-
plete isolation of the liver. In contrast, HVI-CHP is a simpler technique which can be performed percutaneously. As illustrated in Fig. 1, a laparotomy is not essential for this method to be carried out if the hepatic arterial catheter is inserted preoperatively by the Seldinger technique. In this study, the use of Adriamycin was decided upon because it is a potent cytotoxic agent with a dose-dependent action, and it has well-documented pharmacokinetics for our system.6 Moreover, although it is not a standard drug used in the treatment of colorectal cancer, it was chosen because it has a lower hepatic extraction ratio than 5-fluorodeoxyuridine,z5 and the pharmacokinetic advantage under HVI-CHP would theoretically be greater in drugs with a relatively low hepatic extraction ratio. With regard to the duration of drug administration and extraction by the filters, a 5- to 15-rain infusion was combined with 20- to 30-rain HVICHP in this study. As we reported in our previous canine studies, 6 plasma Adriamycin levels at the inlet side of the CHP filter after a 10-rain infusion promptly decreased from the time of completion and became similar to the values obtained at other sampling points at 20rain. Taking this into consideration, 20-30min seems to be a reasonable duration of HVI-CHP for a 5- to 15min infusion. The adverse effects observed in the present study primarily consisted of HVI-CHP-related injury to the blood, complications caused by the arterial catheter, and drug-related toxicities, but there were no treatment-related deaths. However, HVI-CHP did entail injury to the blood corpuscles, as manifested by hemolysis and thrombocytopenia immediately after treatment. Considering the high rate of these adverse effects, we have currently undertaken a modification of the extracorporeal circuit design, in which the CHP filters are placed in the outflow line from a centrifugal pump, whereby the hepatic effluent passes through the filters under a pump-generated positive pressure, rather than under a negative pressure as in the present study. This modification has in fact significantly reduced the incidence and extent of hemolysis and hematuria by 20% (unpublished data). A gastric ulcer requiring cimetidine treatment developed in one patient who had received Adriamycin through an infusion catheter placed in the proper hepatic artery by the Seldinger technique. However, it is well-documented that extrahepatic drug perfusion is frequently associated with serious gastrointestinal complications during regional chemotherapy. Therefore, it may be preferable to perform surgical ligation of the regional collateral arteries including the right gastric and gastroduodenal arteries. Despite the high-dose infusion of up to 120mg/m2 of Adriamycin, cardiac toxicity, as indicated by ECG abnormalities or hypotensive shock, was not observed in
312 any of our patients, although a leukocyte count of below 2,000/mm 3 was observed in three patients, who also suffered alopecia. According to the previously published data, alopecia and severe myelosuppression almost always occurred with a one-shot dosage of Adriamycin exceeding 60mg/m2.16 Thus, we consider that the efficacy of this system is substantial in reducing the systemic toxicities. The liver is still the primary organ to be affected by a high-dose infusion of cytotoxic agents. Chemical hepatitis, 17,18as indicated by an increase in the AST value to more than twice the baseline, occurred in almost all patients, although it subsided promptly after treatment. Furthermore, most of the patients showed a transient rise in total bilirubin due to H V I - C H P related hemolysis immediately after treatment; however, none presented a clinical picture consistent with sclerosing cholangitis. 19 It should also be noted that the extent of AST elevation seemed to be more marked in those patients treated by balloon-occluded arterial infusion as this increases the hepatic extraction of the drug 12 and thereby augments parenchymal injury of the liver. Thus, dosage reduction may be required in patients with impaired liver function who are treated with balloon-occluded arterial infusion. After the treatment, 6 of the 12 patients showed a rapid decline in the C E A value to below 50% of the pretreatment value by 1 month. In most of these patients, the hepatic tumors had been identified only as an area of slightly decreased density on the pretreatment CT scan. After treatment, a marked reduction in density was observed in the tumor area, which was not enhanced by contrast media. Consequently, the tumor became well-defined in contour. Although these findings do not meet the standard oncologic criteria of an objective response, they are noteworthy and invariably associated with the sharp decrease in serum C E A levels after treatment. Aigner et al. reported that even when a complete response has been achieved, colorectal hepatic metastases tend to develop liquefaction and hypodensity as a sign of tumor necrosis. 5 Thus, it may be speculated that the mode of response of colorectal hepatic metastases to a high-dose cytotoxic agent is somewhat different from the objective responses observed in standard regional chemotherapy, such as the continuous infusion of cytostatic agents like 5fluorodeoxyuridine. 2° These observations prompted us to consider that a significant reduction in tumor density on the CT scan after treatment indicated an objective response when associated with a sharp decreased in the serum C E A levels to less than 50% of the pretreatment value after i month. This type of response was observed in 6 of our 12 patients, or 50%, the duration of which lasted from 3 to 9 months. According to the results of regional 5-fluorodeoxyuridine infusion by Niederhuber
Y. Ku et al.: Percutaneous Hepatic Venous Isolation et al. 21 the median duration of hepatic response in patients with extrahepatic disease was 6 months, which was much shorter than that observed in patients without extrahepatic spread. In our series, almost all the patients demonstrated evidence of extrahepatic disease consisting of abdominal lymph node or visceral involvement, pulmonary metastases, or both, prior to the commencement of treatment. In addition, the majority had an extensive hepatic tumor burden. Although a direct comparison is difficult due to the limited number of patients in the present study and the lack of uniform response criteria, one massive dose of Adriamycin under H V I - C H P appears to be almost comparable to the reported results of continuous regional infusion chemotherapy with multiple cycles of 5fluorodeoxyuridine. 18,;1-~3 Nevertheless, it should be noted that extrahepatic disease was still the major cause of death in this study, and therefore, as a future direction, a repeated high-dose chemotherapy regimen using H V I - C H P may be of particular value to prolong the survival of those patients with disease confined only to the liver. In conclusion, the results of the present study indicate that H V I - C H P is a safe and useful procedure for the administration of otherwise hazardous high-dose intraarterial chemotherapy to the liver, which is far less complicated than the ILP technique.
Acknowledgments. This study was supported in part by Grants-in-Aid for Scientific Research nos. 0280721, 04670781, and 06454386, from the Ministry of Education, Science and Culture, Japan.
References
1. Left RS, Thompson JM, Johnson DB, Mosley KR, Daly MB, Knights Wa III, Ruxe RL Jr, Messerchmidt GL (1986) Phase II trial of high-dose melphalan and autologous bone marrow transplantation for metastatic colon carcinoma. J Clin Oncol 4:15861591 2. Frei III E, Canellos GP (1980) Dose: a critical factor in cancer chemotherapy. Am J Med 69:585-594 3. Aigner KR, M(illerH, Walther H, Link KH (1988) Drug filtration in high-dose regional chemotherapy. Contrib Oncol 29:261280 4. Kamidono S, Fujii A, Hamami G, Nakano Y, Umezu K, Oda Y, Ishigami J (1984) New preoperative chemotherapy for bladder cancer using combination hemodialysis and direct hemoperfusion: preliminary report. J Urol 131:36-40 5. Aigner KR, Walther H, Link KH (1988) Isolated liver perfusion with MMC/5-FU: surgical technique, pharmacokinetics, clinical results. Contrib Oncol 29:22%246 6. Ku Y, Saitoh M, Nishiyama H, Fujiwara S, Iwasaki T, Tominaga M, Maekawa Y, Ohyanagi H, Saitoh Y (1990) Extracorporeal removal of anticancer drugs in hepatic artery infusion: the effect of direct hemoperfusioncombined with veno-venousbypass. Surgery 107:273-281 7. Maekawa Y, Ku Y, Saitoh Y (1993) Extracorporeal cisplatin removal using direct hemoperfusionunder hepatic venous isolation
Y. Ku et al.: Percutaneous Hepatic Venous Isolation
8.
9. 10.
11. 12.
13. 14. 15. 16.
for hepatic arterial chemotherapy: an experimental study on pharmacokinetics. Surg Today 23:58-62 Ku Y, Saitoh M, Iwasaki T, Tominaga M, Maekawa Y, Shiki H, Samizo M, Fukumoto T, Kuroda Y, Sako M, Okada M, Saitoh Y (1993) Intraarterial high-dose chemotherapy for unresectable hepatocellular carcinoma using direct hemoperfusion under hepatic venous isolation. Eur J Surg Oncol 19:387-392 Ku Y, Saitoh Y (1994) Extracorporeal carbon chemofiltration under hepatic venous isolation for high-dose intraarterial chemotherapy of the liver. Surgery 116:941 Ku Y, Fukumoto T, Iwasaki T, Tominaga M, Samizo M, Nishida T, Kuroda Y, Hirota S, Sako M, Obara H, Saitoh Y (1995) Clinical pilot study on high-dose intraarterial chemotherapy using direct hemoperfusion under hepatic venous isolation in patients with advanced hepatocellular carcinoma. Surgery 117:510 519 Weber J, Novak D (1980) Occlusion arteriography: diagnostic and therapeutic applicability of balloon catheters. Cardiovasc Intervent Radiol 3:81-96 Kawabata M, Takashima S, Mitsuzane K, Tanaka M, Maeda C, Hamachi J, Nomura S, Matsuoka T, Kobayashi N, Sato M, Nakatsuka H, Yamada R (1984) Balloon-occluded arterial infusion therapy for malignant hepatic tumors (in Japanese). Jpn J Cancer Chemother 11:806-813 Creech O Jr, Krementz ET, Ryan RF, Winblad JN (1958) Chemotherapy of cancer: regional perfusion utilizing an extracorporeal circuit. Ann 8urg 148:612-632 Mulcare RJ, Solis A, Fortner JG (1973) Isolation and perfusion of the liver for cancer chemotherapy. J Surg Res 15:87-95 Chert HG, Gross IF (1980) Intraarterial infusion of anticancer drugs: theoretic aspects of drug delivery and review of responses. Cancer Treat Rep 64:31-40 Hirose H, Aoyama M, Oshima K, Udo K, Kobayashi S, Adachi N, Nonaka T, Yamada T (1982) Chemotherapy for patients with
313
17.
18.
19.
20.
21. 22.
23.
hepatocellular carcinoma: clinical evaluation of high-dose Adriamycin regimen with intraarterial infusion (in Japanese). Jpn J Cancer Chemother 9:216-2221 Kemeny N, Daly J, Oderman P, Shike M, Chun H, Petroni G, Geller N (1982) Hepatic artery pump infusion: toxicity and results in patients with metastatic colorectal carcinoma. J Clin Oncol 2:595-600 Sail F, Bittner R, Roscher R, Schnmacher K, Gaus W, Gunter H, Beget H (1980) Regional chemotherapy for hepatic metastases of colorectal carcinoma (continuous intraarterial versus continuous/ intravenous therapy). Cancer 64:379-387 Kemeny MM, Battifora H, Blayney DW, Cecchi G, Goldberg DA, Leong LA, Margolin KA, Terz JJ (1985) Sclerosing cholangitis after continuous hepatic artery infusion of FUDR, Ann Surg 202:176-181 Ku Y, Samizo M, Fukumoto T, Shiki H, Maekawa Y, Iwasaki T, Tominaga M, Kuroda Y, Saitoh Y (1993) A comparative study of tumor responses of hepatocellular carcinoma and colorectal hepatic metastasis to cytotoxic anticancer drug. Jpn J Cancer Chemother 20:1355-1361 Niederhuber JE, Ensminger WG, Thrall J, Walker S, Cozzi E (1984) Regional chemotherapy of colorectal cancer metastatic to the liver. Cancer 53:1336-1343 Chang AE, Schneider PD, Sugarbaker PH, Simpson C, Culnane M, Steinberg SM (1987) A prospective randomized trial of regional versus systemic continuous 5-fluorodeoxyuridine chemotherapy in the treatment of colorectal liver metastases. Ann Surg 206:685-693 Cohen AM, Greenfield A, Wood WC, Waltman A, Novelline R, Athanasoulis C, Scheeffer NJ (1983) Treatment of hepatic metastases by transaxillary hepatic artery chemotherapy using an implanted drug pump. Cancer 51:2013-2019