Proliferative Activity of Intrahepatic Colorectal Metastases After Preoperative Hemihepatic Portal Vein Embolization NORIHIRO KOKUDO,1 KEIICHIRO TADA,1 MAKOTO SEKI,1 HIROTOSHI OHTA,1 KAORU AZEKURA,1 MASASHI UENO,1 KEIICHIRO OHTA,1 TOSHIHARU YAMAGUCHI,1 TOSHIKI MATSUBARA,1 TAKASHI TAKAHASHI,1 TOSHIFUSA NAKAJIMA,1 TETSUICHIRO MUTO,1 TAKAAKI IKARI,2 AKIO YANAGISAWA,3 AND YO KATO3
Although hemihepatic portal vein embolization (PVE) has been used preoperatively to extend indications for hepatectomy in patients with colorectal metastases, the effects of this procedure on tumor growth and outcome remain controversial. To address this issue, we assessed the proliferative activity of intrahepatic metastases after PVE and the long-term outcome of this procedure. Eighteen patients with colorectal metastases underwent preoperative PVE between 1996 and 2000 (PVE group). Twenty-nine patients who underwent major hepatic resection without PVE served as control (non-PVE group). The hepatic parenchymal fraction of the left lobe had significantly increased from 38.1 ⴞ 3.2% to 45.9 ⴞ 2.9% 3 weeks after PVE (ⴙ20.5%, P < .0001). Tumor volume and percent tumor volume had also significantly increased from 223 ⴞ 89 mL to 270 ⴞ 97 mL (ⴙ20.8%, P ⴝ .016) and from 13.7 ⴞ 4.3% to 16.2 ⴞ 4.9% (ⴙ18.5%, P ⴝ .014), respectively. There was no apparent correlation between the increase in parenchymal volume and that in tumor volume. The Ki-67 labeling index of metastatic lesions was 46.6 ⴞ 7.2% in the PVE group and 35.4 ⴞ 12.6% in the non-PVE group (P ⴝ .013). Long-term survival was similar in the PVE and non-PVE groups, however, disease-free survival was significantly poorer in the PVE group than in the non-PVE group (P ⴝ .004). We conclude that PVE increases tumor growth and probably is associated with enhanced recurrence of disease. Although PVE is effective in extending indications for surgery, patient selection for PVE should be cautious. (HEPATOLOGY 2001;34:267-272.) Hepatic resection provides the only chance for cure in patients with colorectal metastases. To extend indications for hepatectomy, hemihepatic portal vein embolization (PVE) has been performed in selected patients.1-3 PVE induces homolateral atrophy of the portion of the liver scheduled for resection and contralateral compensatory hypertrophy of the Abbreviations: PVE, portal vein embolization; HCC, hepatocellular carcinoma; UFT, uracil and tegafur; CPT-11, Irinotecan; 5⬘-DFUR, 5⬘-deoxy-5-fluorouridine; TLV, total liver volume; RLV, volume of right lobe; LLV, volume of left lobe; TV, tumor volume; THPV, total hepatic parenchymal volume; LPV, parenchymal volume of the left lobe; HGF, hepatocyte growth factor; TGF, transforming growth factor; IL, interleukin. From the 1Departments of Surgery and 2Medicine, Cancer Institute Hospital, and the 3Department of Pathology, Cancer Institute, Tokyo, Japan. Received March 7, 2001; accepted May 24, 2001. Address reprint requests to: Norihiro Kokudo, M.D., Hepato-Biliary-Pancreatic Surgery Division, Department of Surgery, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, Japan 113-8655. E-mail:
[email protected]; fax: (81) 3-5684-3989. Copyright © 2001 by the American Association for the Study of Liver Diseases. 0270-9139/01/3402-0009$35.00/0 doi:10.1053/jhep.2001.26513
remnant liver, thus decreasing the risk of postoperative liver failure. PVE is indicated when the remnant liver is expected to be very small, i.e., about 40% smaller than preoperative liver volume, or when tumor spread requires a right hemihepatectomy with partial resection of the left side of the liver.4,5 The positive effects of PVE on hepatic function must be weighed against recent evidence suggesting that this procedure may promote oncogenesis. Elias et al., have reported that after PVE liver metastases may grow more rapidly than liver parenchyma.6 This assumption was based on a study of only 5 patients who had tumors in the nonembolized lobe of the liver, rather than the embolized lobe. In addition, they focused on tumor growth in the nonembolized lobe. Consequently, their findings were considered too premature to warrant contraindication of PVE in patients with colorectal metastases.7,8 When used in combination with arterial chemoembolization, PVE suppresses growth of hepatocellular carcinoma (HCC).9 Long-term results of HCC resection after PVE have recently been shown to be better than or comparable with those of liver resection without PVE,4,10 but little information is available on the effect of preoperative PVE in patients with colorectal metastases.1,11,12 This retrospective study was designed to address 3 important questions concerning the effect of PVE on colorectal metastases: (1) Does PVE promote tumor growth after PVE? (2) Is tumor growth after PVE, if any, similar to or different from hypertrophy of the nonembolized lobe? (3) Is long-term outcome or tumor recurrence affected by preoperative PVE? PATIENTS AND METHODS Between June 1996 and November 2000, 96 patients underwent 114 hepatic resections (90 first resections, 20 second, and 4 third) for liver metastases from colorectal cancer. Forty-seven of these patients underwent major hepatic resection (resection of more than 1 segment of Healey13), and 18 underwent hemihepatic portal vein embolization (PVE) preoperatively to enhance the safety of extended liver resection (PVE group). PVE was done for patients in whom the parenchymal resection rate was expected to be more than 60%. Twenty-nine patients who underwent major hepatic resection without preoperative PVE were studied as control (non-PVE group). Of the 18 patients in the PVE group, only 1 received intra-arterial chemotherapy [16 cycles of mitomycin C (4 mg bolus) and 5-fluorouracil (250 mg/2 hrs)]. This patient was initially considered as inoperable because of the multiple lesions involving all of the 4 segments of Healey.13 After hepatic arterial chemotherapy, a small single lesion in the left lateral segment became unclear, and hepatic resection was planned. As an adjuvant chemotherapy after liver resection, 5 patients of the PVE group took oral administration of UFT (uracil and tegafur,
267
268 KOKUDO ET AL. 300-600 mg/day, for 3 to 18 months), and 1 was administered intravenous Irinotecan (CPT-11, 200 mg/body, 4 doses). No adjuvant chemotherapy was done in the rest of the patients in the PVE group. In non-PVE group, UFT and 5⬘-deoxy-5-fluorouridine (5⬘-DFUR, 800-1200 mg/day) were administered orally in 8 and 3 patients, respectively. An intravenous 5-fluorouracil (250-500 mg/day) was administered in 5 patients. Two of them also received CPT-11 (170210 mg/week). No adjuvant chemotherapy was done in 13 of the non-PVE group. Percutaneous Transhepatic Portal Vein Embolization (PVE). Preoperative hemihepatic portal vein embolization was done as described previously.4,5 Briefly, an ultrasound-guided percutaneous puncture of the transverse portion of left portal vein was done under local anesthesia, with the use of a standard 5-F needle catheter. The catheter was advanced to the portal trunk to permit venous portography. A guidewire was then placed into the main right portal branch, and embolization was performed with a mixture of gelatin sponges (Spongel, Yamanouchi Ltd., Tokyo, Japan) and contrast medium. Several steel coils (Embolization Coils, Cook Inc. Bloomington, IN) were used to ensure embolization of the major portal branches. Computed Tomographic Hepatic Volumetry. Transverse computed tomographic scans of the entire liver were taken in 1 breath-hold before and approximately 3 weeks after PVE (average 20.4 days, range 15-29 days), immediately before hepatic resection. Average interval between PVE and operation was 24.3 days (18-37 days). Total liver volume (TLV) was calculated by multiplying the area of each crosssectional enhanced liver image (in late phase with visualized hepatic veins) by the slice thickness (1 cm or 0.5 cm).14 The right and left lobes were traced individually, and the volume of each lobe was calculated (RLV and LLV). The middle hepatic vein trunk and gallbladder were used as landmarks to define the borders between the right and left sides of the liver. Tumor volume (TV) was similarly estimated by manually tracing each tumor border. When there were multiple tumors, TV was the sum of all tumor volumes. Total hepatic parenchymal volume (THPV) was calculated by subtracting TV from TLV. Parenchymal volume of the left lobe (LPV) was calculated by subtracting TV in the left lobe, if any, from LLV. Hepatic parenchymal fraction of the left lobe (%LPV) was calculated as follows: %LPV⫽100⫻LPV/THPV Gain in parenchymal fraction after PVE (⌬%LPV) was calculated as: ⌬%LPV ⫽ %LPVpost ⫺ %LPVpre, where %LPVpre and %LPVpost were %LPVs before and 3 weeks after PVE, respectively. Percent tumor volume was calculated as follows: %TV⫽100⫻TV/TLV The change in %TV between the 2 computed tomographic scans (⌬%TV) was calculated as: ⌬%TV ⫽ %TVpost ⫺ %Tvpre, where %TVpre and %TVpost were %TV before and 3 weeks after PVE, respectively. Ki-67 Labeling Index. The proliferative activity of intrahepatic tumors was assessed on the basis of nuclear Ki-67 labeling index, as described previously.15-17 In brief, after deparaffinization and inhibition of endogenous peroxidase, the sections were processed for avidin-biotin staining. The primary antibody used was monoclonal rabbit antihuman Ki-67 antigen (A0047, DAKO A/S, Glostrup, Denmark). The labeling index was determined by observing approximately 1000 nuclei in randomly selected areas from the region with the greatest antigen expression and calculating the percentage of Ki-67–labeled nuclei. Necrotic areas were disregarded, so that only viable tumor tissue was evaluated. This was done by an experienced pathologist who was blinded to all related data. Statistics. All values are expressed as means ⫾ SEM. Mann-Whitney U test was used to compare group data. Fisher’s exact test was used for 2⫻2 categorical data. The level of significance was established at P ⬍ .05. Survival was calculated from the date of hepatic
HEPATOLOGY August 2001
resection until death. Survival curves were generated by the KaplanMeier method and were compared with log-rank test. RESULTS Demographics and Operative Data. Patient characteristics in the PVE group and non-PVE group are summarized in Table 1. There was no significant difference in age, sex, primary site (colon vs. rectum), histological type of primary lesion, temporal relationship (metachronous vs. synchronous), number of tumors, or surgical margin. Only the maximal tumor diameter differed significantly (6.80 ⫾ 1.00 cm in the PVE group vs. 4.86 ⫾ 0.66 cm in the non-PVE group, P ⫽ .049). Synchronous metastases were defined as concurrent metastases or metastases occurring within 1 year after operation for the primary lesion. One patient in the PVE group was hepatitis B (HBs) antigen positive, and one in the non-PVE group was hepatitis C antibody positive. Liver function tests were normal in the above 2 patients and histology of liver parenchyma in the resected specimens was normal. Portal Vein Embolization (PVE). PVE was technically possible in all attempted cases. Tolerance to PVE was excellent, and the only discernible side effect was moderate, transient fever, approaching 38°C to 38.5°C. There were no other complications related to PVE. Operative Procedures. Operative procedures in the PVE group were right trisegmentectomy in 1 patient, extended right hepatectomy in 12 (including 2 who concurrently underwent wedge resection), and right hepatectomy in 5. In the non-PVE group, we performed right trisegmentectomy in 1 patient, right hepatectomy in 6, extended left hepatectomy in 1, left hepatectomy in 7, central bisegmentectomy in 1, and segmentectomy with wedge resection in 13 (Table 2). The resected specimen weight and resected parenchymal fraction in the PVE group were 776 ⫾ 101 g and 48.3 ⫾ 4.5%, respectively. These values were significantly greater than those in the non-PVE group: 376 ⫾ 42 g (P ⫽ .001) and 30.5 ⫾ 2.8% (P ⫽ .003), respectively. There were not significant differences in mean operating time and intraoperative blood loss between PVE and non-PVE group, 397 ⫾ 18 minutes versus 397 ⫾ 17 minutes and 1,504 ⫾ 224 g versus 1,301 ⫾ 170 g, respectively. In the PVE group, there was no increase in surgical difficulty in controlling the vessels in the porta hepatis. Ligation of the right portal trunk was possible in all of the patients. There was no operative mortality either in the PVE or nonPVE group. Transient signs of postoperative liver failure developed in 1 patient in the PVE group. This patient had jaun-
TABLE 1. Patient Characteristics
Age (range) Sex (M/F) Colon/rectum Histologic type (well/mod 䡠 por)* Synchronous/metachronous No. of metastases Maximal diameter (cm) Surgical margin (mm)
PVE Group (n ⴝ 18)
Non-PVE Group (n ⴝ 29)
P
59.7 (40⬃73) 12/6 12/6 11/7 14/4 2.50 ⫾ 0.47 6.80 ⫾ 1.00 4.47 ⫾ 0.98
60.4 (40⬃73) 18/11 23/6 20/9 17/12 2.86 ⫾ 0.62 4.86 ⫾ 0.66 5.26 ⫾ 0.93
.800 ⬎.999 .493 .753 .219 .313 .049 .800
*Histological type of primary lesion (well, mod, por: well, moderately, poorly differentiated adenocarcinoma).
HEPATOLOGY Vol. 34, No. 2, 2001
KOKUDO ET AL.
TABLE 2. Type of Surgical Procedures
Trisegmentectomy Right trisegmentectomy Bisegmentectomy Extended right hepatectomy Right hepatectomy Extended left hepatectomy Left hemihepatectomy Central bisegmentectomy Segmentectomy Left lateral segmentectomy Left medial segmentectomy Right posterior segmentectomy
PVE Group (n ⴝ 18)
Nonv-PVE Group (n ⴝ 29)
1
1
12(2*) 5 0 0 0
0 6 1 7(4) 1
0 0 0
7(7) 1(1) 5(5)
*No. of patients who underwent associated wedge resections.
dice (serum total bilirubin level, ⬎6 mg/dL) and massive ascites after extended right hemihepatectomy with a hepatic parenchymal resection rate of 76.4%. Effect of PVE on Liver Volume. Hypertrophy of the portion of the liver scheduled to be preserved, i.e., the left lobe, was evident in all patients in the PVE group. The parenchymal volume of the left lobe (LPV) had significantly increased from 385 ⫾ 34 mL to 464 ⫾ 28 mL 3 weeks after PVE (⫹20.6%, P ⫽ .0029). The hepatic parenchymal fraction of the left lobe (%LPV) had also significantly increased from 38.1 ⫾ 3.2% to 45.9 ⫾ 2.9% after PVE (P ⬍ .0001). One patient with a large left lobe (%LPVpre ⫽ 72.2%) apparently did not meet the criteria for PVE. This patient also had a tumor in the left lobe, and PVE was attempted to enhance safety and increase the surgical margin for wedge resection, scheduled to be done concomitantly with extended right hepatectomy. The average gain in the parenchymal fraction after PVE (⌬%LPV) was ⫹20.5%. There was a significant negative correlation between %LPVpre and ⌬%LPV (r ⫽ ⫺.716, P ⫽ .0012, Fig. 1). Change in Tumor Volume During Waiting Period After PVE. Of the 18 patients in the PVE group, 3 had tumors not only in the
FIG. 1. Correlation between preoperative hepatic parenchymal fraction of the left lobe (%LPVpre) and gain in parenchymal fraction after PVE (⌬%LPV). There was a significant negative correlation between %LPVpre and ⌬%LPV (r ⫽ ⫺.716, P ⫽ .0012).
269
TABLE 3. Changes in Tumor volumes in Embolized and Nonembolized Lobes
Case 1 Case 2 Case 3
Tumors in Nonembolized Lobe
Tumors in Embolized Lobe
Before PVE
After PVE
% Change
Before PVE
After PVE
% Change
19.9* 8.0 2.3
20.0 11.4 2.5
⫹0.5% ⫹42.1% ⫹9.7%
1183 12.1 111
1218 12.4 118
⫹3.0% ⫹2.5% ⫹6.3%
*cm3.
embolized lobe but also in the nonembolized lobe. The changes in the tumor volumes in the embolized and nonembolized lobes are shown in Table 3. The following variables were assessed on the bases of data derived from tumors in the embolized lobe only. Tumor volume (TV) had significantly increased from 223 ⫾ 89 mL to 270 ⫾ 97 mL 3 weeks after PVE (⫹20.8%, P ⫽ .016, Fig. 2). Percent tumor volume (%TV) also significantly
FIG. 2. Change in tumor size after PVE as shown by serial computed tomography. A 51-year-old woman underwent PVE before hepatic resection for colorectal metastases. The volume of her future remnant liver increased from 315 cm3 (32.2%) to 485 cm3 (40.6%) 3 weeks after PVE, thus warranting a safe right hepatectomy. However, tumor volume in the embolized lobe also substantially increased from 135 cm3 to 260 cm3. Sets of computed tomographic scans taken at the same level before (above) and 3 weeks after PVE (below).
270 KOKUDO ET AL.
HEPATOLOGY August 2001
FIG. 5. Kaplan-Meier estimates of overall survival after hepatic resection for metastatic colorectal cancer in patients with or without preoperative PVE. Survival was similar in the PVE (solid line, n ⫽ 18) and non-PVE (dashed line, n ⫽ 29) groups. FIG. 3. Relation between gain in left hepatic parenchymal fraction (⌬%LPV) and tumor growth (⌬%TV) during the waiting period after PVE. There was no apparent correlation between ⌬%LPV and ⌬%TV (r ⫽ ⫺.126, P ⫽ .641).
increased from 13.7 ⫾ 4.3 % to 16.2 ⫾ 4.9% after PVE (P ⫽ .0139). The change in %TV between the 2 computed tomographic scans (⌬%TV) was ⫹18.5%. There was no apparent correlation between ⌬%LPV and ⌬%TV (r ⫽ ⫺.126, P ⫽ .641, Fig. 3) or between %LPVpre and ⌬%TV (r ⫽ ⫺.083, P ⫽ .783). Ki-67 Labeling Index of the Metastatic Lesion. The Ki-67 labeling index of metastatic lesions was 46.6 ⫾ 7.2% in the PVE group and 35.4 ⫾ 12.6% in the non-PVE group (P ⫽ .013, Fig. 4). This index was determined for the most recent 30 cases studied (12 in the PVE group and 18 in the non-PVE group). Long-Term Outcome After Liver Resection. The patients were followed up for 4 to 55 months (average, 23.5 months) after liver resection. The overall actuarial survival rate of the 18
FIG. 4. Ki-67 labeling index of metastatic lesions. The Ki-67 labeling index of metastatic lesions was 46.6 ⫾ 7.2% in the PVE group and 35.4 ⫾ 12.6% in the non-PVE group (P ⫽ .013).
patients in the PVE group was 59.7% and 47.8% at 2 and 4 years, respectively. These rates were comparable with those in the non-PVE group (67.8% and 50.2%, respectively, P ⫽ .421), as shown in Fig. 5. The disease-free survival rate in the PVE group was 15.2% and 0% at 2 and 4 years, respectively. These rates were significantly poorer than those in the nonPVE group (45.8% and 34.4%, respectively, P ⫽ .004, Fig. 6). The most common site of first recurrence was the remnant liver in both the PVE and non-PVE groups (7 and 8 cases, respectively, Table 4). Three patients with recurrence underwent a second hepatectomy. The lung was the second most common site of recurrence in both groups. There was no apparent difference in pattern of recurrence between the 2 groups. DISCUSSION
In our study, preoperative PVE induced sufficient hypertrophy of the nonembolized lobe of the liver and thus enhanced the safety of major hepatic resection in patients with advanced colorectal metastases. However, tumor growth during the waiting period after PVE was appreciable and perhaps was slightly promoted. Tumor volume significantly increased by 20.8% during the 3 weeks after PVE. On the basis of average values, we estimated that tumor-doubling time after PVE was
FIG. 6. Kaplan-Meier estimates of disease-free survival after hepatic resection in patients with metastatic colorectal cancer, with or without preoperative PVE. Patients in the PVE group (solid line, n ⫽ 18) had a significantly higher rate of recurrence than those in the non-PVE group (dashed line, n ⫽ 29, P ⫽ .004).
HEPATOLOGY Vol. 34, No. 2, 2001
KOKUDO ET AL.
TABLE 4. Site of First Recurrence After Liver Resection Site
PVE Group (n ⴝ 18)
Non-PVE Group (n ⴝ 29)
Remnant liver Lung Peritoneum Lymph node Local (pelvic) Brain Total
7(2*) 2(2) 1 1 1 1(1) 12†
8(1) 4(1) 2 1 1 0 14
*No. of patients who underwent resection. †Patients with multiple recurrence sites are included.
76.1 days. This value is slightly less than the reported average doubling time of 92.4 days for resectable liver metastases from colorectal carcinoma.18 The significantly increased tumor Ki-67 labeling index in the PVE group (Fig. 4) also provides indirect evidence of increased tumor growth after PVE. Because tumor biopsy before PVE is not ethically acceptable, Ki-67 labeling index was not compared in the same patient before and after PVE, but was compared between the 2 groups of patients, the PVE group and non-PVE group. Although the average volume gain of the nonembolized left lobe (⫹20.6%) was similar to that of liver tumors (⫹20.8%), these gains were not parallel in individual patients (Fig. 3). Apparently, tumor growth after PVE is not controlled by the same mechanisms as hypertrophy of the preserved portion of the liver. Tumor growth after PVE may be controlled by 3 factors: malignant potential of the tumors, changes in cytokines or growth factors induced by PVE, and changes in blood supply after PVE. In patients who undergo hepatectomy, the regeneration rate of the remnant liver during the first several weeks is directly proportional to resection rate.19,20 The significant correlation between the hepatic parenchymal fraction of the left lobe (%LPV) and the gain in parenchymal fraction after PVE (⌬%LPV, Fig. 1) is consistent with the direct relation between the regeneration rate and resection rate. This finding suggests that hypertrophy of the nonembolized lobe after PVE is controlled by the same factors as liver regeneration after hepatectomy. Our results are in accordance with those of other studies, in which PVE was used primarily in patients with HCC or bile duct cancer.21,22 Animal models of portal vein branch ligation have been recently used to assess molecular events after PVE. Several cytokines known to play important roles in liver regeneration are also expected to be key players after PVE. Expression of hepatocyte growth factor (HGF)-mRNA markedly increases in the nonligated growing lobe between 6 and 24 hours after ligation, but is only slightly elevated in the ligated shrinking lobe.23,24 Intraportal injection of exogenous HGF is known to stimulate liver regeneration after portal branch ligation in
271
dogs.25 Because HGF stimulates growth and motility of colorectal carcinoma cells in vitro,26,27 increased tissue levels of HGF may increase the level in plasma,28 thus stimulating the growth of hepatic tumors not only in the nonembolized lobe but also in the embolized lobe. Negative regulators of hepatocyte proliferation, such as transforming growth factor (TGF)-1 and interleukin (IL)-1 are strongly expressed in the ligated shrinking lobe in rats.24 Because TGF--signal transduction is abnormal in approximately one-third of all colorectal carcinomas with distant metastases,29 intrahepatic tumors might be resistant to growth suppression induced by TGF-. Another factor potentially stimulating tumor growth after PVE is increased hepatic arterial blood flow after embolization of the portal branch.30 This arterial compensation has long been known to occur after reduction of segmental portal blood flow.31,32 Computed tomographic studies of the hepatic parenchyma estimate that arterial blood flow in embolized segments is 1.7 times that in nonembolized segments.33 Portal branch occlusion also decreases hepatic artery resistance index in the embolized lobe, without modifying total portal flow.34 Because intrahepatic metastases depend solely on arterial blood supply,35 increased hepatic arterial flow may provide nutritional advantages for tumor growth. Table 5 summarizes events known to take place after PVE. In their first article describing the possibility of rapid intrahepatic tumor growth after PVE, Elias et al.6 focused only on tumors in the nonembolized lobe. Given that many events that occur after PVE (Table 5), tumor growth should be estimated separately in the embolized and nonembolized lobes. Because only 3 patients had tumors in the nonembolized lobe in our series, we primarily assessed tumor growth in the embolized lobe. A direct comparison of long-term outcome after liver resection between the PVE group and non-PVE group may lead to erroneous conclusions, because the 2 groups differed with respect to maximum tumor diameter (Table 1). However, similar survival in the PVE group and non-PVE group despite the larger tumor size in the former indicates that major hepatic resection after PVE is a useful treatment strategy in patients with advanced colorectal metastases. Randomized controlled trials are not technically or ethically feasible because most candidates for PVE have advanced disease, unresectable without preoperative PVE. Only a few studies have evaluated long-term outcome in patients with colorectal metastases who underwent PVE followed by liver resection. Kawasaki et al. reported on 5 patients, 4 surviving for more than 3 years and 1 surviving for more than 5 years.1 Azoulay et al. have recently reported a 5-year survival rate of 38% in a series of 19 patients who underwent liver resection after PVE, with an operative mortality of 4% (1/19).12 Of the 30 patients who underwent PVE in their series, 11 (37%) did not receive liver resection, mainly because of extensive tumor spread at laparotomy. In
TABLE 5. Summary of Events Known to Take Place After PVE or Portal Vein Ligation
DNA synthesis in hepatocytes (rat)24 Expression of HGF-mRNA (rat)23,24 Expression of TGF-, IL-1-mRNA (rat)24 Hepatic arterial blood flow (human)33 Hepatic artery resistance (human)34 Portal flow (human)34
In Embolized Lobe
In Nonembolized Lobe
No increase Slightly elevated Strongly expressed Increased by 1.7 times Decreased
Increased (peak at 36 hrs) Markedly increased
Unchanged Increased
272 KOKUDO ET AL.
HEPATOLOGY August 2001
our study, no patient had disease progression severe enough to preclude surgery after PVE. It is unclear whether the significant difference in diseasefree survival between the PVE group and non-PVE group was due to tumor growth induced after PVE or to other demographic differences between the groups. Further study is needed to define the effect of neo-adjuvant or adjuvant chemotherapy in patients who undergo liver resection after PVE. Segmental arterial compensation after PVE33 may facilitate transarterial infusion chemotherapy for liver tumors in embolized hepatic segments. In conclusion, preoperative PVE induced sufficient hypertrophy of the preserved portion of the liver and thus decreased the risk of major hepatic resection in patients with advanced colorectal metastases. There was no operative mortality in our series, and long-term survival was similar to that of patients who underwent major hepatic resection without PVE. However, PVE increases tumor growth and probably is associated with enhanced recurrence of disease. Although PVE is effective in extending indications for surgery in patients with advanced colorectal metastases against which no other potentially curative treatment is available,36 patient selection for PVE should be done cautiously. REFERENCES 1. Kawasaki S, Makuuchi M, Kakazu T, Miyagawa S, Takayama T, Kosuge T, Sugihara K, et al. Resection for multiple metastatic liver tumors after portal embolization. Surgery 1994;115:674-677. 2. Makuuchi M, Thai BL, Takayasu K, Takayama T, Kosuge T, Gunven P, Yamazaki S, et al. Preoperative portal embolization to increase safety of major hepatectomy for hilar bile duct carcinoma: a preliminary report. Surgery 1990;107:521-527. 3. Minagawa M, Makuuchi M, Torzilli G, Takayama T, Kawasaki S, Kosuge T, Yamamoto J, et al. Extension of the frontiers of surgical indications in the treatment of liver metastases from colorectal cancer: long-term results. Ann Surg 2000;231:487-499. 4. Azoulay D, Castaing D, Krissat J, Smail A, Hargreaves GM, Lemoine A, Emile J-F, et al. Percutaneous portal vein embolization increases the feasibility and safety of major liver resection for hepatocellular carcinoma in injured liver. Ann Surg 2000;232:665-672. 5. Elias D, Debaere T, Roche A, Bonvallot S, Lasser P. Preoperative selective portal vein embolizations are an effective means of extending the indications of major hepatectomy in the normal and injured liver. HepatoGastroenterology 1998;45:170-177. 6. Elias D, de Baere T, Roche A, Ducreux M, Leclere J, Lasser P. During liver regeneration following right portal embolization the growth rate of liver metastases is more rapid than that of the liver parenchyma. Br J Surg 1999;86:784-788. 7. Seymour K, Manas D, Charnley RM. During liver regeneration following right portal vein embolization the growth rate of liver metastases is more rapid than that of the liver parenchyma [Letter]. Br J Surg 1999;86:1482-1483. 8. Elias D. During liver regeneration following right portal vein embolization the growth rate of liver metastases is more rapid than that of the liver parenchyma [Letter, author’s reply]. Br J Surg 1999;86:1483. 9. Yamakado K, Hirano T, Kato N, Takeda K, Nakagawa T, Takase K, Nakano T, et al. Hepatocellular carcinoma: treatment with a combination of transcatheter arterial chemoembolization and transportal ethanol injection. Radiology 1994;193:75-80. 10. Tanaka H, Hirohashi K, Kubo S, Shuto T, Higaki I, Kinoshita H. Preoperative portal vein embolization improves prognosis after right hepatectomy for hepatocellular carcinoma in patients with impaired hepatic function. Br J Surg 2000;87:879-882. 11. Takayama T, Makuuchi M, Kosuge T, Yamamoto J, Shimada K, Inoue K. Preoperative portal embolization. Ann Ital Chir 1997;68:745-750. 12. Azouley D, Castaing D, Smail A, Adam R, Cailliez V, Laurent A, Lemoine A, et al. Resection of nonresectable liver metastases from colorectal cancer after percutaneous portal vein embolization. Ann Surg 2000;231:480-486. 13. Healey JE, Schroy PC. Anatomy of the biliary ducts within the human liver. Analysis of the prevailing pattern of branchings and the major variations of the biliary ducts. Arch Surg 1953;66:599-616.
14. Heymsfield SB, Fulenwider T, Nordlinger B, Barlow R, Sones P, Kutner M. Accurate measurement of liver, kidney, and spleen volume and mass by computerized axial tomography. Ann Intern Med 1979;90:185-187. 15. Kinoshita T, Ito H, Miki C. Serum interleukin-6 level reflects the tumor proliferative activity in patients with colorectal carcinoma. Cancer 1999; 85:2526-2531. 16. De Jong KP, Stellema R, Karrenbeld A, Koudstaal J, Gouw ASH, Sluiter WJ, Peeters PMJG, et al. Clinical relevance of transforming growth factor ␣, epidermal growth factor receptor, p53, and Ki67 in colorectal liver metastases and corresponding primary tumors. HEPATOLOGY 1998;28:971-979. 17. Petrowski H, Sturm I, Graubitz, Kooby DA, Staib-Selber E, Gog C, Kohne C-H, et al. Relevance of Ki-67 antigen expression and K-ras mutation in colorectal liver metastases. Eur J Surg Oncol 2001;27:80-87. 18. Nomura K, Miyagawa S, Harada H, Kitamura H, Seki H, Shimada R, Kobayashi A, et al. Relationship between doubling time of liver metastases from colorectal carcinoma and residual primary cancer. Dig Surg 1998;15:21-24. 19. Yamanaka N, Okamoto E, Kawamura E, Kato T, Oriyama T, Fujimoto J. Furukawa K, et al. Dynamics of normal and injured human liver regeneration after hepatectomy as assessed on the basis of computed tomography and liver function. HEPATOLOGY 1993;18:79-85. 20. Kokudo N, Vera DR, Koizumi M, Seki M, Sato T, Stadalnik RC, Takahashi T. Recovery of hepatic asialoglycoprotein receptors after major hepatic resection. J Nucl Med 1999;40:137-141. 21. Yamakado K, Takeda K, Matsumura K, Nakatsuka A, Hirano T, Kato N, Sakuma H, et al. Regeneration of the un-embolized liver parenchyma following portal vein embolization. J Hepatol 1997;27:871-880. 22. Imamura H, Shimada R, Kubota M, Matsuyama Y, Nakayama A, Minagawa S, Makuuchi M, et al. Preoperative portal vein embolization: an audit of 84 patients. HEPATOLOGY 1999;29:1099-1105. 23. Shimizu Y, Suzuki H, Nimura Y, Onoue S, Nagino M, Tanaka M, Ozawa T. Elevated mitochondrial gene expression during rat liver regeneration after portal vein ligation. HEPATOLOGY 1995;22:1222-1229. 24. Uemura T, Miyazaki M, Hirai R, Matsumoto H, Ota T, Ohashi R, Shimizu N, et al. Different expression of positive and negative regulators of hepatocyte growth in growing and shrinking hepatic lobes after portal vein branch ligation in rats. Int J Mol Med 2000;5:173-179. 25. Ueno S, Aikou T, Tanabe G, Kobayashi Y, Hamanoue M, Mitsuse S, Kawaida K, et al. Exogenous hepatocyte growth factor markedly stimulates liver regeneration following portal branch ligation in dogs. Cancer Chemother Pharmacol 1996;38:233-237. 26. Dignass AU, Lynch-Devancy K, Podolsky DK. Hepatocyte growth factor/ scatter factor modulates intestinal epithelial cell proliferation and migration. Biochem Biophys Res Commun 1994;202:701-709. 27. Nabeshima K, Shimano Y, Inoue T, Itoh H, Kataoka H, Koono M. Hepatocyte growth factor/scatter factor induces not only scattering but also cohort migration of human colorectal-adenocarcinoma cells. Int J Cancer 1998;78:750-759. 28. Michalopoulos GK, DeFrances MC. Liver regeneration. Science 1997; 276:60-66. 29. Miyaki M, Iijima T, Konishi M, Sakai K, Ishii A, Yasuno M, Hishima T, et al. Higher frequency of Smad4 gene mutation in human colorectal cancer with distant metastasis. Oncogene 1999;18:3098-3103. 30. Price JB Jr, Voorhees AB JR, Britton RC. Operative hemodynamic studies in portal hypertension: significance and limitations. Arch Surg 1967;95: 843-851. 31. Itai Y, Moss AA, Goldberg HI. Transient hepatic attenuation difference of lobar or segmental distribution detected by dynamic computed tomography. Radiology 1982;144:835-839. 32. Matsui O, Takashima T, Kadoya K, Kitagawa K, Kamimura R, Itoh H, Suzuki M, et al. Segmental staining on hepatic arteriography as a sign of intrahepatic portal vein obstruction. Radiology 1984;152:601-606. 33. Nagino M, Nimura Y, Kamiya J, Kanai M, Hayakawa N, Yamamoto H. Immediate increase in arterial blood flow in embolized hepatic segments after portal vein embolization: CT demonstration. Am J Roentgenol 1998;171:1037-1039. 34. Denys AL, Abehsera M, Leloutre B, Sauvanet A, Vilgrain V, O’Toole O, Belghiti J, et al. Intrahepatic hemodynamic changes following portal vein embolization: a prospective Doppler study. Eur Radiol 2000;10:1703-1707. 35. Archer SG, Gray BN. Vascularization of small liver metastases. Br J Surg 1989;76:545-548. 36. Seymour K, Chsrnley RM, Rose J, Baudouin CJ, Manas DM. Extending the indications for curative liver resection by portal vein embolization. Br J Surg 2000;87:362-373.
