Fax: (011) 81-463-96-5577
Prophylactic hepatic arterial infusion chemotherapy for the prevention of liver metastasis in patients with colon carcinoma†
A randomized controlled trial
Article first published online: 12 DEC 2003
Copyright © 2003 American Cancer Society
Volume 100, Issue 3, pages 590–597, 1 February 2004
How to Cite
Sadahiro, S., Suzuki, T., Ishikawa, K., Yasuda, S., Tajima, T., Makuuchi, H., Saitoh, T. and Murayama, C. (2004), Prophylactic hepatic arterial infusion chemotherapy for the prevention of liver metastasis in patients with colon carcinoma. Cancer, 100: 590–597. doi: 10.1002/cncr.11945
See related editorial on pages 437–40, this issue.
- Issue published online: 20 JAN 2004
- Article first published online: 12 DEC 2003
- Manuscript Accepted: 15 SEP 2003
- Manuscript Received: 15 JUL 2003
- colon carcinoma;
- liver metastases;
- 5-fluorouracil (5-FU);
- hepatic arterial infusion (HAI)
The liver is the most frequent site of recurrence after curative resection in patients with colon carcinoma. For liver metastasis, a high response rate can be achieved with hepatic arterial infusion (HAI) chemotherapy. In the current study, the authors administered 5-fluorouracil (5-FU) as adjuvant chemotherapy by HAI to patients with colon carcinoma without liver metastases and studied its effects on recurrence in the liver and survival.
A total of 316 patients with preoperative Stage II or Stage III colon carcinoma (according to the 1997 revision of the International Union Against Cancer TNM staging system) were randomly assigned to receive surgery plus 3-week continuous HAI of 5-FU or surgery alone. There were 305 eligible patients, of whom the 119 patients assigned to the HAI arm actually received 5-FU. The primary endpoint was disease-free survival, whereas the secondary endpoints were overall survival and liver metastasis-free survival. Analysis was by intent to treat.
There were no significant differences noted in morbidity between the two treatment arms. During the follow-up period (median, 59.0 months), the incidence of liver metastasis was significantly decreased in the HAI arm whereas there were no significant differences reported between the 2 arms with regard to the frequency of metastasis at other sites. In the HAI arm, the risk ratio for recurrence was 0.40 (95% confidence interval [95% CI], 0.24–0.64; P = 0.0002), the risk ratio for death was 0.37 (95% CI, 0.21–0.67; P = 0.0009), and the risk ratio for liver metastasis was 0.38 (95% CI, 0.22–0.66; P = 0.0005). These differences were found to be significant only for patients with Stage III disease. Toxicities were mild.
A schedule of 3-week HAI of 5-FU given as adjuvant chemotherapy to patients with Stage III colon carcinoma appeared to contribute to a significant decrease in the frequency of liver metastases and was associated with an improved survival rate. Cancer 2004. © 2003 American Cancer Society.
In the case of advanced colon carcinoma, approximately 45% of the patients die of the malignancy1 and 83% of patients with recurrent disease develop a liver metastasis.2 Therefore, prevention of liver recurrence can be expected to improve the prognosis of patients with advanced colon carcinoma most effectively. Taylor et al. reported that infusion of 5-fluorouracil (5-FU) into the portal vein during the perioperative period resulted in the significant prevention of liver metastasis in patients with colorectal carcinoma without liver metastasis and improved their prognosis.3 Among several follow-up studies, to our knowledge only the report by Wereldsma et al.4 reported a decrease in the frequency of liver metastasis after the administration of 5-FU via the portal vein, and to our knowledge, none of these studies demonstrated an improvement in survival.5, 6
Conversely, several randomized control trials (RCTs) published to date have demonstrated that hepatic arterial infusion (HAI) of 5-FU or fluorodeoxyuridine (FUDR) produces a significantly better response rate in patients with colon carcinoma metastatic to the liver compared with systemic chemotherapy,7 but fails to significantly improve survival rates. In patients who underwent resection of liver metastasis from colorectal carcinoma, postoperative adjuvant chemotherapy with a combination of HAI of FUDR and systemic infusion of 5-FU was not found to demonstrate a significant overall survival benefit over systemic chemotherapy alone, although it was associated with a significantly better progression-free survival.8, 9
Therefore, we administered 5-FU by HAI as prophylactic therapy during the perioperative period to patients with advanced colon carcinoma without apparent liver metastasis based on preoperative evaluation. The study endpoints were disease-free survival, overall survival, and liver metastasis-free survival as evaluated by intent-to-treat analysis.
MATERIALS AND METHODS
Eligibility Criteria, Pretreatment Evaluation, and Follow-Up
The inclusion criteria were age ≤ 75 years with histologically proven adenocarcinoma of the colon, no severe major organ dysfunction, a World Health Organization (WHO) performance status of 0 or 1, no prior cancer therapy, and Stage II (T3-4N0M0) or Stage III (anyTN1-2M0) disease (according to the 1997 revision of the International Union Against Cancer TNM staging system) as determined by a preoperative evaluation that included a barium enema examination, fiberoptic colonoscopy, and an abdominal computed tomography (CT) scan. The sites of the tumor were from the cecum to the rectosigmoid. According to the criteria of the Japanese Society for Cancer of the Colon and Rectum,10 the rectosigmoid is defined as the bowel at the level between the promontorium and the lower margin of the second sacral vertebra on the lateral view of a barium enema examination. The rectosigmoid is included in the colon. Therefore, a tumor located on the anal side below the lower margin of the second sacral vertebra according to barium enema examination was considered to be rectal carcinoma and thus was excluded from the current study.
Patients fulfilling the eligibility criteria were randomly assigned to receive either surgery plus 3-week continuous HAI of 5-FU or surgery alone.
Randomization was performed a week before surgery. To prevent selection bias, random numbers were generated by one of the authors (C.M.), who was not involved in enrollment of the patients. Because of the nature of the treatment, randomization was not masked.
In the current study, catheterization was avoided in patients in whom the right hepatic artery branched from the superior mesenteric artery (SMA), even if they were assigned to the HAI arm. Written informed consent was provided by all patients. This RCT protocol was approved by the Institutional Review Board of our study institution.
In the patients assigned to the HAI arm, the catheter was inserted from the femoral artery because the catheter was to be removed after the completion of prophylactic chemotherapy. Under local anesthesia, the celiac artery and the SMA were imaged using a 4-French, J-shaped catheter (Clinical Supply, Tokyo, Japan) to confirm the anatomy of the abdominal vessels. The tip of the catheter (Gentry catheter; Solution Industries, Yokohama, Japan) then was fixed in the gastroduodenal artery using a coil. The drug flowed out from the side hole of the catheter, which was located within the common hepatic artery. Malperfusion was prevented by angiographic embolization of aberrant vasculature. In the current study, catheterization was avoided in patients in whom the right hepatic artery branched from the SMA. The catheter was connected to a subcutaneously implantable reservoir (P-U Cersite Port; Toray, Tokyo, Japan) and implanted subcutaneously in the vicinity.
5-FU was administered at a dose of 250 mg/day for 3 weeks, continuously from 1 week prior to the surgery to 2 weeks after the surgery. The 7-day dose of 5-FU (1750 mg) was mixed with a liposteroid, dexamethasone palmitate emulsion (Limethason; Mitsubishi Pharma, Osaka, Japan), which contained 4.0 mg of dexamethasone palmitate equivalent to 2.5 mg of dexamethasone, to prevent the occlusion of the hepatic artery; saline then was added to the mixture to obtain a 50-mL emulsion, which was injected into the balloon of the intraarterial pump (SFA-501W; Nipro Medical Corporation, Osaka, Japan).11 The intraarterial pump was connected to the implanted reservoir with a percutaneous needle. The balloon of the pump was changed every week, and the infusion was continued for 3 consecutive weeks. After completion of the infusion, both the catheter and the implanted reservoir were removed under local anesthesia. The total dose of 5-FU was 5250 mg.
The control arm received neither chemotherapy nor the administration of dexamethasone-palmitate emulsion during the perioperative period. For both arms, adjuvant chemotherapy beyond 3 weeks after surgery was not dictated by the study and was left to the discretion of the physician.
Surgery and Pathology
Patients underwent surgery approximately 1 week after the initiation of HAI (range, 2–10 days). All patients were prepared mechanically and chemically. Antibiotics were administered for 1–4 days postoperatively. Information regarding the intraabdominal findings with special reference to regional and distant metastases was obtained from the surgical reports.
Data regarding tumor size, histologic type, tumor penetration, lymph node metastases, and pathological TNM disease stage were obtained from the pathologic records.
The follow-up schedule after surgery for asymptomatic patients was a chest X-ray, abdominal ultrasound or abdominal CT scan, and blood tests including carcinoembryonic antigen every 3–4 months for 2 years after surgery. These examinations and tests were continued every 4–6 months for 3–4 years after surgery, and every 6 months thereafter. As a rule, fiberoptic colonoscopy was conducted every 2 years. For symptomatic patients, the above examinations and tests were conducted more frequently, as needed, without a specified schedule.
The primary endpoint of the current study was disease-free survival, and the secondary endpoints were overall survival and liver metastasis-free survival by intent-to-treat analysis. It is estimated that approximately 40% of patients with advanced colon carcinoma develop recurrent disease; of this 40%, approximately 60% (or 25% of all patients) develop a liver metastasis. Based on the hypothesis that prophylactic HAI would reduce liver recurrence from 25% to 10% of all patients, the sample size required was calculated to be 100 patients per arm (significance level of 5% and power of 80%). It has been reported that the right hepatic artery branches from the SMA in approximately 20% of the Japanese population,12 and because these patients were considered not eligible for catheter insertion, the sample size was established at 150 patients per arm.
Overall survival curves and disease-free survival curves were generated by the Kaplan–Meier method13 and compared by the log-rank test. The Student t test and chi-square test also were used for the analysis.
Characterization of the Patients
A total of 316 patients admitted to Tokai University Hospital were enrolled in the study during a 7-year period (June 1994 to June 2001) and randomized to 1 of the 2 treatment arms prior to surgery. Eleven patients were excluded because of liver or peritoneal metastasis observed at the time of surgery but not detected in the preoperative evaluation (Table 1). Among those assigned to the HAI arm, 119 patients received 5-FU by HAI. These patients comprised the HAI-1 group. In the remaining 31 patients, the catheter was not inserted because the right hepatic artery branched from the SMA or because of technical problems. These 31 patients comprised the HAI-2 group.
|Lost to follow-up||0||0|
There were no significant differences noted between the HAI arm and the control arm with regard to gender, age, performance status (WHO), location of the primary tumor, tumor size, histopathologic stage of disease, prior cancer therapy, postoperative adjuvant chemotherapy, or duration of follow-up (Table 2). Forty-five patients with preoperative Stage II or Stage III disease were determined to have histopathologic Stage I disease. As of the time of last follow-up (March 2003), no patient had been lost to follow-up, and the follow-up period was 61.3 ± 23.8 months (range, 19.7–107.5 months; median, 59.0 months).
|Age (yrs)||60.0 ± 10.0a||60.0 ± 8.9||0.96|
|Performance status (WHO)|
|Location of primary tumor|
|Tumor size (cm)||4.6 ± 1.7a||4.7 ± 2.0||0.47|
|No adjuvant chemotherapy||28||34|
|Follow-up period (mos) (range)||61.9 ± 24.5a (19.8–107.0)||60.7 ± 23.3 (19.7–107.5)||0.66|
Overall Disease-Free Survival
The perioperative mortality rate was 0%. Using intent-to-treat analysis, which included the 31 patients who did not receive 5-FU by HAI (the HAI-2 group), the HAI arm had a significantly better disease-free survival compared with the control arm (P = 0.0002) (Fig. 1). The overall disease-free survival rate was 88% in the HAI arm and 72% in the control arm at 3 years and was 86% and 68%, respectively, at 5 years. When those patients with Stage II and Stage III disease were analyzed and compared separately, there were no significant differences noted between the two treatment arms in the Stage II patients (P = 0.07), but there was a significant difference noted in the Stage III patients (P = 0.0009) (Figs. 2, 3). and The risk ratio for disease-free survival was 0.40 (95% confidence interval [95% CI], 0.24–0.64; P = 0.0002) in the HAI arm, indicating that the prophylactic therapy reduced the risk of disease recurrence and death by 60%.
Using intent-to-treat analysis, which included the 31 patients who did not receive 5-FU by HAI (the HAI-2 group), the HAI arm demonstrated a significantly better survival compared with the control arm (P = 0.0005) (Fig. 4). The overall survival was 94% in the HAI arm and 82% in the control arm at 3 years, and was 89% and 76%, respectively, at 5 years. When the Stage II and Stage III patients were analyzed and compared separately, there were no significant differences noted between the two treatment arms in the Stage II patients (P = 0.15), but there was a significant difference in the Stage III patients (P = 0.002) (Figs. 5 and 6). and In the HAI arm, the risk ratio of overall survival was 0.37 (95%CI, 0.21–0.67; P = 0.0009), indicating that the prophylactic therapy reduced the risk of death by 63%.
Liver Metastasis-Free Survival
Using intent-to-treat analysis, the HAI arm demonstrated a significantly better liver metastasis-free survival compared with the control arm (P = 0.0002) (Fig. 7). The liver metastasis-free survival rate was 94% in the HAI arm and 79% in the control arm at 3 years, and was 87% and 73%, respectively, at 5 years. When the Stage II and Stage III patients were analyzed and compared separately, there were no significant differences noted between the two treatment arms in the Stage II patients (P = 0.06), but there was a significant difference noted in the Stage III patients (P = 0.002). In the HAI arm, the risk ratio for liver metastasis-free survival was 0.38 (95% CI, 0.22–0.66; p = 0.0005), indicating that the prophylactic therapy reduced the risk of liver metastasis and death by 62%.
Toxicity and Complications
Eight of the 119 subjects in the HAI-1 group experienced Grade 1 nausea (according to the National Cancer Institute Common Toxicity Criteria). Although elevation of serum aspartate aminotransferase or alanine aminotransferase was observed in 50% of patients in the HAI arm and in 49% of patients in the control arm during the perioperative period, Grade 3 hepatic toxicity was observed in only 2% of patients in the HAI arm and in 3% of patients in the control arm. No other toxicities were observed. There were no differences noted between the two treatment arms with respect to postoperative complications. No serious complications were reported. There were no femoral artery complications noted in relation to the catheterization. In three patients in the HAI arm, the 5-FU solution extravasated around the implanted reservoir and caused a skin ulcer that improved with conservative treatment (Table 3).
|Total no. of patients||150||155||119||31|
|Ulcer around the port||3||0||3||0|
Sites of First Recurrence
The sites of first recurrence in the HAI arm were the liver in 6 patients (4%), the lungs in 5 patients (3%), and local recurrence in 5 patients (3%). In the control arm, the most frequent site of recurrence was the liver (24 patients; 15%), followed by the lungs (9 patients; 6%), peritoneum (8 patients; 5%), and local recurrences (6 patients, 4%). Recurrence in the liver was reported to be reduced significantly in the HAI arm (P = 0.002), but there were no significant differences noted in the frequency of recurrence at other sites between the 2 treatment arms (Tables 4, 5). Among the 119 patients in the HAI-1 group, 1 patient developed a liver metastasis (1%). The liver recurrence rate in the HAI-2 group was 16%, which was similar to that noted in the control arm (15%). The liver recurrence rates by stage of disease were none of the 18 patients with Stage I disease (0%) none of the 56 patients with Stage II disease (0%), and 1 of the 45 patients with Stage III disease (3%) in the HAI-1 group, whereas the corresponding rates in the HAI-2 group were 1 of 8 patients with Stage I disease (13%), 2 of 12 patients with Stage II disease (17%), and 2 of 11 patients with Stage III disease (18%). The liver recurrence rates in the control arm were none of 19 patients with Stage I disease (0%), 8 of 73 patients with Stage II disease (11%), and 16 of 63 patients with Stage III disease (25%).
|Total no. of patients||150||155||119||31|
|Liver||6 (4)||24 (15)||1 (1)||5 (16)|
|Lung||5 (3)||9 (6)||3 (3)||2 (6)|
|Peritoneum||3 (2)||8 (5)||3 (3)||0 (0)|
|Lymph node||2 (1)||4 (3)||1 (1)||1 (3)|
|Locoregional||5 (3)||6 (4)||4 (3)||1 (3)|
|No. of patients||Liver||Lung||Peritoneum||Lymph node||Locoregional|
|Stage I||26||1 (4)||0 (0)||0 (0)||0 (0)||0 (0)|
|Stage II||68||2 (3)||1 (1)||1 (1)||0 (0)||0 (0)|
|Stage III||56||3 (5)||4 (7)||2 (4)||2 (4)||5 (9)|
|Stage I||19||0 (0)||1 (5)||0 (0)||1 (5)||0 (0)|
|Stage II||73||8 (11)||2 (3)||2 (3)||0 (0)||2 (3)|
|Stage III||63||16 (25)||6 (10)||6 (10)||3 (5)||4 (6)|
Status of the Patients
Table 6 shows the status of the patients at the time of last follow-up (March 2003).
|Total no. of patients||150||155||119||31|
|Alive, no TF||124||100||101||23|
|Dead, no TF||5||5||5||0|
|Secondary primary tumor|
It has been estimated that in patients with colon carcinoma but no apparent liver metastasis at the time of surgery who subsequently developed liver metastases, the primary liver metastasis is considered to have developed > 2 years prior to the surgery, based on calculations from the tumor doubling time.14 Thomas et al. studied liver tissues removed from cancer patients and found that ultrasonography performed ex vivo was able to detect 95% of tumors with a greatest dimension of ≥ 10 mm, but detected none of the tumors measuring < 5 mm in greatest dimension.15 Therefore, in patients with no liver metastasis at the time of surgery and judged by the surgeon to have been curatively resected, the metachronous liver metastasis occurring postoperatively is believed to originate from micrometastases present at the time of surgery. The liver is the most frequent site of disease recurrence in patients with colon carcinoma, so that a reduction in the incidence of liver metastases may be the most effective way to reduce the likelihood of recurrence and to improve the prognosis. Taylor et al. reported that in patients with curatively resected Dukes Stage A, B, or C colorectal carcinoma, a 7-day infusion of 5-FU at a dose of 1000 mg/body/day into the portal vein during the perioperative period resulted in a significant reduction in liver metastases and a significant improvement in the overall survival.3 However, subsequent RCTs of portal vein infusion of 5-FU in the perioperative period did not demonstrate any reduction in liver metastases or improvement in survival.5, 6 A meta-analysis of RCTs of portal vein chemotherapy involving 4000 patients, including the aforementioned studies, indicated that portal vein infusion of 5-FU with or without mitomycin C demonstrated no clear specific reduction in the frequency of liver metastases, and the improvement in the survival rate demonstrated by a small percentage was attributed to the systemic effects of these drugs.16
With regard to liver metastases from colon carcinoma, a meta-analysis of seven RCTs has shown that HAI using FUDR has a significantly better response rate compared with systemic chemotherapy with FUDR or 5-FU.7 This finding is explained by the observation that the HAI of FUDR or 5-FU can achieve a much higher drug concentration in the liver tumor compared with portal vein infusion or intravenous infusion.17 It also has been shown experimentally that liver tumors measuring 0.5 to ≥ 3 mm in greatest dimension receive abundant blood supply from the hepatic artery.18–20
These studies have led physicians to speculate that HAI may be extremely effective in the treatment of occult liver metastasis, but to our knowledge there have been no previous RCTs that support this idea.
Because the first-pass extraction of FUDR is greater than 5-FU, FUDR is more widely used for HAI in Western countries.21 Because of National Health Insurance restrictions in Japan, FUDR cannot be used; therefore, 5-FU was used in the current study. There has been a previous report on the administration of 5-FU concurrently with folinic acid using HAI,22 but at the time the study was initiated, the use of folinic acid in combination with 5-FU had not been approved in Japan. Thus, only 5-FU was administered by HAI in the current study.
With regard to the method of 5-FU administration, continuous infusion was chosen in view of the report that 5-FU has a higher response rate when given as a continuous infusion rather than as a bolus.23 Warren et al. reported that, in combination with folinic acid, a dose of 1500 mg/m2 of 5-FU given as a 24-hour continuous infusion once a week for 6 weeks for liver metastases resulted in a 48% response rate.22 In the current study, in view of the fact that the target metastatic lesions were occult or small and that the treatment was 3-week continuous infusion, the dose chosen was 250 mg/day (1750 mg/week) for a total dose of 5250 mg. A liposteroid, dexamethasone palmitate emulsion, was used concurrently with 5-FU to prevent hepatic artery occlusion.11
In three patients in the HAI arm, drug extravasation resulted in skin ulcers, but no severe complications were observed. There were no significant differences noted between the two arms with regard to the occurrence of other complications.
Compared with the control group, patients in the HAI arm were found to have a significantly better disease-free survival, overall survival, and liver metastasis-free survival. The frequency of first recurrence in the liver was significantly decreased in the HAI arm; however, the rate of recurrence at other sites did not appear to differ, indicating that the survival benefit of 5-FU administered by HAI could be attributable to a decreased recurrence in the liver.
At the time of the initiation of the current study, the standard adjuvant chemotherapy used in Western countries, namely leucovorin combined with 5-FU, had not been approved by the Japanese National Health Insurance System. With regard to adjuvant chemotherapy for Stage II or Stage III colorectal carcinoma, oral fluoropyrimidines generally are used in Japan because they are very convenient.24, 25 Although the postoperative adjuvant chemotherapy was not specified in the protocol, all patients who did receive adjuvant chemotherapy received an oral fluoropyrimidine including oral 5-FU. Adjuvant chemotherapy was given to 81% of the patients in the HAI arm and 78% of the patients in the control arm, with no significant differences reported between the 2 arms. Therefore, we concluded that the differences in the disease-free survival observed between the HAI arm and the control arm were independent of any effects of the postoperative adjuvant chemotherapy.
Although further studies are required to confirm the results of the current study, the findings herein raise the possibility that the prophylactic administration of 5-FU by HAI may become the standard adjuvant chemotherapy for patients with Stage III colon carcinoma.
- 10Japanese Society for Cancer of the Colon and Rectum. Japanese classification of colorectal carcinoma. 1st edition. Tokyo, Kanehara, 1997.
- 12Manual of abdominal angiography [in Japanese]. 2nd edition. Tokyo: Nankodo, 1998..
- 23Meta-Analysis Group In Cancer. Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. J Clin Oncol. 1998; 16: 301–308.