Is postoperative adjuvant chemotherapy useful for gallbladder carcinoma?
A Phase III multicenter prospective randomized controlled trial in patients with resected pancreaticobiliary carcinoma
Version of Record online: 3 OCT 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 8, pages 1685–1695, 15 October 2002
How to Cite
Study Group of Surgical Adjuvant Therapy for Carcinomas of the Pancreas and Biliary Tract, Takada, T., Amano, H., Yasuda, H., Nimura, Y., Matsushiro, T., Kato, H., Nagakawa, T. and Nakayama, T. (2002), Is postoperative adjuvant chemotherapy useful for gallbladder carcinoma?. Cancer, 95: 1685–1695. doi: 10.1002/cncr.10831
- Issue online: 3 OCT 2002
- Version of Record online: 3 OCT 2002
- Manuscript Accepted: 23 MAY 2002
- Manuscript Revised: 16 MAY 2002
- Manuscript Received: 14 JAN 2002
- pancreatic carcinoma;
- gallbladder carcinoma;
- bile duct carcinoma;
- carcinoma of the ampulla of Vater;
- adjuvant chemotherapy;
- disease-free survival
To the authors' knowledge, the significance of postoperative adjuvant chemotherapy in pancreaticobiliary carcinoma has not yet been clarified. A randomized controlled study evaluated the effect of postoperative adjuvant therapy with mitomycin C (MMC) and 5-fluorouracil (5-FU) (MF arm) versus surgery alone (control arm) on survival and disease-free survival (DFS) for each specific disease comprising resected pancreaticobiliary carcinoma (pancreatic, gallbladder, bile duct, or ampulla of Vater carcinoma) separately.
Between April 1986 and June 1992, a total of 508 patients with resected pancreatic (n = 173), bile duct (n = 139), gallbladder (n = 140), or ampulla of Vater (n = 56) carcinomas were allocated randomly to either the MF group or the control group. The MF group received MMC (6 mg/m2 intravenously [i.v.]) at the time of surgery and 5-FU (310 mg/m2 i.v.) in 2 courses of treatment for 5 consecutive days during postoperative Weeks 1 and 3, followed by 5-FU (100 mg/m2orally) daily from postoperative Week 5 until disease recurrence. All patients were followed for 5 years.
After ineligible patients were excluded, 158 patients with pancreatic carcinoma (81 in the MF group and 77 in the control group), 118 patients with bile duct carcinoma (58 in the MF group and 60 in the control group), 112 patients with gallbladder carcinoma (69 in the MF group and 43 in the control group), and 48 patients with carcinoma of the ampulla of Vater (24 in the MF group and 24 in the control group) were evaluated. Good compliance (> 80%) was achieved with MF treatment. The 5-year survival rate in gallbladder carcinoma patients was significantly better in the MF group (26.0%) compared with the control group (14.4%) (P = 0.0367). Similarly, the 5-year DFS rate of patients with gallbladder carcinoma was 20.3% in the MF group, which was significantly higher than the 11.6% DFS rate reported in the control group (P = 0.0210). Significant improvement in body weight compared with the control was observed only in patients with gallbladder carcinoma. There were no apparent differences in 5-year survival and 5-year DFS rates between patients with pancreatic, bile duct, or ampulla of Vater carcinomas. Multivariate analyses demonstrated a tendency for the MF group to have a lower risk of mortality (risk ratio of 0.654; P = 0.0825) and recurrence (risk ratio of 0.626; P = 0.0589). The most commonly reported adverse drug reactions were anorexia, nausea/emesis, stomatitis, and leukopenia, none of which were noted to be serious.
The results of the current study indicate that gallbladder carcinoma patients who undergo noncurative resections may derive some benefit from systemic chemotherapy. However, alternative modalities must be developed for patients with carcinomas of the pancreas, bile duct, or ampulla of Vater. Cancer 2002;95:1685–95. © 2002 American Cancer Society.
Extensive surgical resection of pancreaticobiliary carcinoma is increasingly and widely accepted to the extent that it is now being performed even in patients with advanced tumors that formerly were considered nonresectable. However, although the surgical resection rate has increased, this increase has not been accompanied by satisfactory improvements in survival. The 5-year survival rate of resected cases currently is reported to be only 4–21% for pancreatic carcinoma,1–3 38% for carcinoma of the ampulla of Vater,4 and 7.5–27% for biliary tract carcinoma.1, 5–7 These findings indicate the limitations of surgical resection and emphasize the need for an adjuvant treatment.
Previous reports of postoperative adjuvant chemotherapy, radiotherapy, and chemoradiotherapy have given hope that such treatments would be effective in prolonging life.8–17 For example, Ishikawa et al. also reported good results with hepatic perfusion chemotherapy via both the hepatic artery and portal vein to prevent recurrence after resection in patients with pancreatic carcinoma, with a 3-year survival rate of 54%.9 The Gastrointestinal Tumor Study Group (GITSG) has noted life-prolonging effects of chemoradiotherapy after surgical resection of pancreatic carcinoma.12, 17 Some reports have suggested that postoperative adjuvant radiotherapy might be more effective than chemotherapy for biliary tract carcinoma.14–16
However, to our knowledge, all these results are based on small-scale studies rather than on large-scale, controlled clinical trials. In addition, reports concerning postoperative adjuvant therapy for biliary tract carcinoma are experimental and cannot be considered as reports of true clinical trials. There are several reasons for the paucity of reports regarding postoperative adjuvant therapy for pancreaticobiliary carcinoma. First, there have been a small number of resectable cases. Second, the surgery itself is so invasive that postoperative adjuvant therapy produces considerable concern regarding the possibility of increasing surgical complications. Third, to our knowledge no effective therapy has been established for the treatment of patients with advanced, unresectable tumors. For these reasons, there was an obvious need for a large-scale controlled study to evaluate rigorously the usefulness of postoperative adjuvant therapy after surgical resection of pancreaticobiliary carcinoma.
Therefore, we performed a randomized study to evaluate the long-term effect of postoperative adjuvant chemotherapy compared with surgery alone in the treatment of patients with resected pancreaticobiliary carcinoma. We expected to involve many patients who had undergone extensive surgery and we were concerned about delayed postoperative recovery or an increase in complications.18, 19 Therefore, we made a special effort to ensure that such problems would not occur. To assure that postoperative adjuvant therapy proceeded as scheduled, a combined postoperative adjuvant therapeutic regimen of moderately toxic mitomycin C (MMC) and 5-fluorouracil (5-FU) (MF) was selected. In this randomized controlled study, we could obtain data from > 500 cases of resected pancreaticobiliary carcinoma, making this the largest study of this type performed to date to our knowledge. The 5-year follow-up has been completed for all patients, and we are able to provide our final report.
MATERIALS AND METHODS
The patients for the current study were selected from patients admitted to the 31 participating medical institutions for pancreaticobiliary carcinoma between April 1986 and June 1992 who fulfilled the following eligibility criteria: 1) a histologically confirmed, preoperative diagnosis of carcinoma of the pancreas, gallbladder, bile duct, or ampulla of Vater; 2) Stage II–IV disease;20, 21 3) confirmed resection of the primary lesion (patients who underwent extended right or left hepatic lobectomy were excluded); 4) age < 75 years; 5) an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0–3; 6) no previous surgery, radiotherapy or chemotherapy; 7) no serious concomitant disease; 8) no concurrent or nonconcurrent multicentric tumor or double tumor; and 9) at the start of treatment, a leukocyte count ≥ 4000/mm3, a platelet count ≥ 100,000/mm3, liver enzymes (aspartate aminotransferase [AST] and alanine aminotransferase [ALT]) ≤ 100 U, and negative urinary protein. All patients and their families received a full explanation from their medical institution with regard to the objectives and methods of the current study, the possible harmful effects of chemotherapy, and matters related to the protection of individual privacy; informed consent to participate in the current study was obtained from all patients. Approval was obtained from the ethics committee of each participating institution.
The surgical and histopathologic findings were recorded in accordance with the Classification of Pancreatic Carcinoma20 and the General Rules for Surgical and Pathological Study on the Cancer of the Biliary Tract,21 as were the findings regarding curability. In the current study, the term “curative resection” indicates complete removal of the tumor with a histologically clear surgical margin and removal of metastasis from the primary and secondary lymph nodes; all other cases are designated as a “noncurative resection.”20, 21
Patients who met the eligibility criteria randomly were allocated on the day of surgery into the postoperative adjuvant chemotherapy group (MF regimen arm) or the surgery alone group (control arm). A randomized design, stratified according to institution and disease, was used. The MF group was given MMC (6 mg/m2) by rapid intravenous infusion on the day of surgery and a slow intravenous infusion of 5-FU (310 mg/m2) in 2 courses of treatment for 5 consecutive days (each during postoperative Weeks 1 and 3), followed by daily oral 5-FU (100 mg/m2) from postoperative Week 5 until disease recurrence. The control group was not treated with any drugs including placebo. In the event of serious adverse drug reactions or abnormal laboratory findings such as a leukocyte count < 3000/mm3, a platelet count < 70,000/mm3, AST and ALT > 200 U, positive findings for urinary protein, or the development of surgical complications, treatment was suspended or discontinued according to the judgment of the attending physician. With the exception of drugs for symptomatic treatment, the use of any concomitant therapy that might interfere with the evaluation of study results was prohibited, such as the use of other anticancer drugs, immunotherapy, or radiotherapy. Treatment after disease recurrence was left to the discretion of the attending physician at each medical institution.
The primary endpoint was survival. Survival time was calculated from the day of surgery, with deaths from all causes treated as events. Follow-up for a period of 5 years was conducted in all patients enrolled in the study. The secondary endpoints were disease-free survival (DFS), ECOG PS, improvement in body weight, and adverse effects. As a rule, patients were monitored for disease recurrence (or obvious progression of residual lesions in patients who underwent noncurative resection) monthly when an overall assessment was made based on the results of a physical examination, X-rays, ultrasound, computed tomography scan, and laboratory findings. Improvements in the ECOG PS were assessed in patients having a preoperative PS of 1–3. A postoperative PS improvement of ≥ 1 grade was considered to be evidence of improvement, as was a postoperative weight gain of at least 2 kg. Adverse drug reactions were assessed according to the criteria of the Japan Society for Cancer Therapy.22
The current trial was scheduled to be analyzed per protocol for each disease of resected pancreaticobiliary carcinoma (such as carcinoma of the pancreas, gallbladder, bile duct, or ampulla of Vater) separately. However, the intent-to-treat analysis of all gallbladder carcinomas or subset analysis also was performed as reference data. Sample size was selected based on an assumed 5-year survival rate of 15% in the control group for pancreatic, bile duct, and gallbladder carcinomas. To detect a difference of 20% under conditions of an alpha of 0.05 (two-sided) and 1-beta of 0.8 (power), a minimum of 60 patients per group was required in each disease category. Because carcinoma of the ampulla of Vater was not frequent, we endeavored to enroll as many patients as possible, but the target number was not reached for that disease.
The duration of survival and DFS were calculated using the Kaplan–Meier method,23 with the log-rank test used to compare the two groups. Patient characteristics, surgical methods, disease recurrence, and improvements in ECOG PS and body weight were compared between the groups using the chi-square test, Fisher exact test, Student t test, or Mann–Whitney U test. In the event a significant difference in survival was noted between the groups, the Cox proportional hazards model (multivariate analysis)24 was used to define whether the postoperative chemotherapy was a true prognostic factor. Data were analyzed using the SAS System for Windows (Release 6.12; SAS Institute Inc., Cary, NC), with the level of significance set at < 0.05.
A total of 508 patients were enrolled and randomly allocated into 2 groups with stratification according to institution and disease. The breakdown of patients was as follows: 173 pancreatic carcinoma patients (89 in the MF arm and 84 in the control arm), 139 bile duct carcinoma patients (67 in the MF arm and 72 in the control arm), 140 gallbladder carcinoma patients (73 in the MF arm and 67 in the control arm), and 56 patients with carcinoma of the ampulla of Vater (26 in the MF arm and 30 in the control arm). Of these, a total of 63 ineligible patients (18 in the MF arm and 45 in the control arm) were excluded from the current study. The excluded patients were comprised of 9 patients who were age ≥ 75 years, 3 patients with other tumors, 34 patients with Stage I disease, 3 patients with a multicentric tumor or double tumor, 11 patients with an ECOG PS of 4, and 3 patients with benign tumors. There also were nine incomplete cases (five patients in the MF group and four patients in the control group), comprised of seven cases of protocol violations and two cases that could not be assessed due to cancellation of informed consent. Table 1 shows the background characteristics of the evaluable patients for each disease. Although gallbladder carcinoma patients randomly were allocated into 2 well balanced groups (73 in the MF arm and 67 in the control arm), the number of eligible cases was unbalanced between the 2 groups (69 in the MF arm and 43 in the control arm) when the ineligible patients, the majority of whom had Stage I disease (3 in the MF arm and 20 in the control arm), were excluded. A statistical analysis showed no imbalance between the two groups with regard to gender, age, ECOG PS, incidence or site of metastasis, disease stage, or curability.
|Pancreas||Bile duct||Gallbladder||Ampulla of Vater|
|No. of patients evaluated||81||77||58||60||69||43||24||24|
|Male||47 (58)||46 (59)||46 (79)||41 (68)||30 (43)||11 (26)||13 (54)||11 (46)|
|Female||34 (42)||31 (40)||12 (21)||19 (32)||39 (57)||32 (74)||11 (46)||13 (54)|
|0–1||52 (64)||55 (71)||38 (66)||35 (58)||56 (81)||33 (77)||14 (58)||17 (71)|
|2–3||29 (36)||22 (29)||20 (34)||25 (42)||13 (19)||10 (23)||10 (42)||7 (29)|
|II||25 (31)||25 (32)||15 (26)||12 (20)||14 (20)||3 (7)||11 (46)||9 (38)|
|III||17 (21)||21 (27)||28 (48)||22 (37)||23 (33)||14 (33)||9 (36)||15 (63)|
|IV||39 (48)||31 (40)||15 (26)||26 (43)||32 (46)||26 (60)||4 (17)||0 (0)|
|Yes||72 (89)||62 (81)||49 (84)||54 (90)||65 (94)||43 (100)||12 (50)||8 (33)|
|No||9 (11)||15 (19)||9 (16)||6 (10)||4 (6)||0||12 (50)||16 (67)|
|Localization of metastasis|
|Lymph nodes||68 (84)||58 (75)||49 (84)||53 (88)||63 (91)||42 (98)||12 (100)||8 (100)|
|Peritoneal||5 (6)||5 (6)||1 (2)||0||6 (9)||3 (7)||1 (4)||0|
|Liver||9 (11)||7 (9)||0||0||8 (12)||5 (12)||0||0|
|Others||0||0||1 (2)||1 (2)||0||0||0||0|
|Curability of resection|
|Curative||45 (56)||47 (61)||34 (59)||38 (63)||31 (45)||20 (47)||21 (88)||20 (83)|
|Noncurative||36 (44)||30 (39)||24 (41)||22 (37)||38 (55)||23 (53)||3 (12)||4 (17)|
With regard to major surgical procedures for pancreatic carcinoma, total pancreatectomy was performed in 5.7% of cases (9 of 158 cases), pancreatoduodenectomy was performed in 64.6% of cases (102 of 158 cases), distal pancreatectomy was performed in 24.7% of cases (39 of 158 cases), and other procedures were performed in 5.1% of cases (8 of 158 cases). For bile duct carcinoma, pancreatoduodenectomy was performed in 61.0% of cases (72 of 118 cases), extrahepatic bile duct resection was performed in 19.5% of cases (23 of 118 cases), extended right lobectomy was performed in 11.9% of cases (14 of 118 cases), extended left lobectomy was performed in 5.9% of cases (7 of 118 cases) and other procedures were performed in 1.7% of cases (2 of 118 cases). For gallbladder carcinoma, cholecystectomy with liver bed resection was performed in 31.3% of cases (35 of 112 cases; 30.4% in the MF arm and 32.6% in the control arm), extended right lobectomy was performed in 27.7% of cases (31 of 112 cases; 30.4% in the MF arm and 23.3% in the control arm), cholecystectomy with bile duct resection was performed in 28.6% of cases (32 of 112 cases; 27.5% in the MF arm and 30.2% in the control arm), and pancreatoduodenectomy was performed in 12.5% of cases (14 of 112 cases; 11.6% in the MF arm and 14.0% in the control arm). With regard to carcinoma of the ampulla of Vater, pancreatoduodenectomy was performed in 97.9% cases (47 of 48 cases) and other procedures were performed in 2.1% of cases (1 of 48 cases). The majority of patients with pancreatic, bile duct, and ampulla of Vater carcinomas were treated with a pancreatoduodenectomy, whereas the majority of patients with gallbladder carcinoma were treated with pancreatoduodenectomy with bile duct resection, liver bed resection, or hepatic resection of fewer than three hepatic segments (extended right or left lobectomy). Our study group had attempted to remove more advanced biliary tract tumors, often in combination with hepatectomy or pancreatoduodenectomy. A statistical analysis demonstrated no imbalance between the two groups with regard to the distribution of surgical procedures.
The rate of completion of MMC treatment was high (94–97%), with no major differences noted among disease categories. The completion rates for intravenous 5-FU were 96.3% (78 of 81 cases) in pancreatic carcinoma patients, 79.3% (46 of 58 cases) in bile duct carcinoma patients, 94.2% (65 of 69 cases) in gallbladder carcinoma patients, and 83.3% (20 of 24 cases) in patients with carcinoma of the ampulla of Vater, with minor differences noted between diseases. The mean total dose of oral 5-FU was 56.9–81.0 g, and it was administered for 12.4–17.7 months.
Five-Year Survival Rate
The survival of patients with gallbladder carcinoma is shown in Figure 1 and that of the other patients is shown in Figure 2. The 5-year survival rate of gallbladder carcinoma patients was significantly better in the MF group (26.0%) compared with the control group (14.4%) (P = 0.0367, two-sided log-rank test). The 5-year survival rate in patients with pancreatic carcinoma was 11.5% in the MF group and 18.0% in the control group, the 5-year survival of patients with bile duct carcinoma was 26.7% in the MF group and 24.1% in the control group, and the 5-year survival of patients with carcinoma of the ampulla of Vater was 28.1% in the MF group and 34.3% in the control group, with no significant differences noted between the groups.
Table 2 shows the 5-year survival rate stratified for curability. The 5-year survival rate after a noncurative resection of gallbladder carcinoma was 8.9% in the MF group, which was significantly better than that in the control group (0%) (P = 0.0226). No differences were noted between the 2 groups with regard to the 5-year survival after either a curative or noncurative resection of pancreatic, bile duct or ampulla of Vater carcinomas or after a curative resection of gallbladder carcinoma.
|Primary site||Curability||No.||MF (%)||Control (%)||P value (log-rank test)|
|Ampulla of Vater||Curative||41||32.1||38.8||0.4953|
The results of gallbladder carcinoma patients were subjected to intent-to-treat analysis. The median survival time was 16.4 months in the MF group and 14.1 months in the control group, with no significant differences noted between the groups (P = 0.2819; data not shown). This finding may have been caused by the unbalanced number of ineligible cases, which mainly were of Stage 1 (3 in the MF arm and 20 in the control arm) and were excluded from the per-protocol analysis.
DFS and Recurrence
As Table 3 shows, the 5-year DFS rate of patients with gallbladder carcinoma was 20.3% in the MF group, a finding that is significantly higher than the 11.6% DFS rate reported in the control group (P = 0.0210, two-sided log-rank test). The 5-year DFS rate for gallbladder carcinoma patients who underwent noncurative resection also was significantly higher in the MF group (7.9%) compared with the control group (0%) (P = 0.0254). There was no significant difference in the 5-year DFS rate of patients with carcinomas of the pancreas, bile duct, or ampulla of Vater between the 2 groups.
|Primary site||Curability||No.||MF (%)||Control (%)||P value (log-rank test)|
|Ampulla of Vater||Overall||48||25.0||20.8||0.9004|
The results for gallbladder carcinoma patients were subjected to intent-to-treat analysis. The median DFS time was 11.9 months in the MF group and 12.3 months in the control group, with no significant differences noted between the groups (P = 0.2994; data not shown), which was similar to the 5-year survival rate.
During the 5-year follow-up period, the carcinoma recurred in 91.0% of patients with pancreatic carcinoma (recurrence/evaluable, 132/145), in 80.0% of patients with bile duct carcinoma (84/105), in 81.4% of patients with gallbladder carcinoma (83/102), and in 76.1% of patients with carcinoma of the ampulla of Vater (35/46). Hepatic recurrence accounted for approximately 50% of cases, followed by disease recurrence in the peritoneal or lymph nodes. There was no statistically significant difference between the groups for any of the diseases.
Improvement in ECOG PS and Body Weight
As Table 4 shows, improvements in body weight were noted in 13.0% (9 of 69 patients) of gallbladder carcinoma patients in the MF group and in no patients in the control group (P = 0.0122, two-sided Fisher exact test). No other significant differences between the two groups were noted with regard to improvements in ECOG PS or body weight.
|Primary site||MF||Control||P value|
|Evaluable no.||No.||%||Evaluable no.||No.||%|
|Ampulla of Vater||20||15||75.0||14||14||100||0.0629|
|Ampulla of Vater||24||2||8.3||24||0||0||0.4893|
Further analysis to identify prognostic factors was performed for the gallbladder carcinoma patients and showed significant differences with regard to the duration of survival and DFS. Based on the results of univariate analysis, which was performed for 14 factors (age, gender, ECOG PS, disease stage, serosal invasion, venous invasion, lymph node metastasis, peritoneal metastasis, hepatic metastasis, histologic lymph node metastasis, histologic depth of tumor invasion, lymph node dissection, curability, and postoperative chemotherapy), multivariate analysis then was performed on the factors for which univariate analysis demonstrated a P value ≤ 0.2. As the results of multivariate analysis in Table 5 show, curability, histologic lymph node metastasis, and the histologic depth of tumor invasion were found to be the factors affecting survival most significantly. Curability also was identified as a factor affecting DFS. Postoperative adjuvant chemotherapy tended to reduce both the risk of death (hazard ratio of 0.654; P = 0.0825) and the risk of disease recurrence (hazard ratio of 0.626; P = 0.0589), although these reductions were not statistically significant. When the same analytical procedures were performed on the patients (with the exception of those with hepatic metastasis or peritoneal dissemination who were not candidates for traditional surgery), postoperative adjuvant chemotherapy was identified as one of the factors that reduced the risk of death (hazard ratio of 0.551; P = 0.0284) and the risk of disease recurrence (hazard ratio of 0.569; P = 0.0497).
|Factors||Risk ratio||95% CI||P value|
|Histologic lymph node metastasis||1.788||1.065–3.001||0.0280b|
|Histologic depth of tumor invasion||2.050||1.262–3.330||0.0037c|
|Invasion of the serosa||2.114||0.733–6.099||0.1661|
|Histologic depth of tumor invasion||1.312||0.793–2.171||0.2899|
A total of 432 patients (232 in the MF group and 202 in the control group) were evaluated with regard to adverse effects. The most frequent surgical complication was intraperitoneal infection, which was observed in 17.7% of patients in the MF group and 13.3% of patients in the control group. Although an ileus (4.3% in the MF group and 4.9% in the control group), biliary leakage (6.0% in the MF group and 8.4% in the control group), or pancreatic leakage (6.5% in the MF group and 6.4% in the control group) developed in approximately 4–8% of patients, there was no apparent difference in the incidence of these complications between the 2 groups.
Adverse drug reactions of ≥ Grade 2 that were found to occur significantly more frequently in the MF group compared with the control group were leukopenia (12.9% in the MF group and 3.0% in the control group), anorexia (22.4% in the MF group and 13.9% in the control group), and nausea/emesis (12.9% in the MF group and 6.9% in the control group) (P < 0.05, chi-square test). Leukopenia developed after intravenous injections of MMC and 5-FU in > 50% of the patients treated. There were no serious adverse drug reactions that were attributed to the chemotherapy.
The current trial, a randomized controlled study of > 500 cases of resected pancreaticobiliary carcinomas, is to our knowledge the largest study of this type reported to date as well as the first randomized controlled study of resected biliary tract carcinoma. The results of the current study demonstrated a significantly better 5-year survival rate and 5-year DFS rate in patients with gallbladder carcinoma in the MF group compared with the control group, in which patients did not receive chemotherapy (overall and noncurative patients). In the current study, there was an imbalance in the number of eligible patients (69 vs. 43) between the 2 gallbladder carcinoma groups as a result of the exclusion of ineligible patients, the majority of whom mainly had Stage I disease (3 in the MF group and 20 in the control group) and were ineligible for the per-protocol analysis. Nevertheless, the results of the multivariate analysis demonstrated that postoperative adjuvant chemotherapy tended to reduce both the risk of death and the risk of disease recurrence. The multivariate analysis of the patients, in which those with hepatic metastasis or peritoneal dissemination were excluded because they were not candidates for traditional surgery, identified postoperative adjuvant chemotherapy to be a factor that reduced both the risk of death and the risk of disease recurrence. Because large-scale studies of this type are so rare, we believe that these findings, which suggest the potential usefulness of postoperative adjuvant chemotherapy, are of considerable importance in the treatment of patients with gallbladder carcinoma.
In the current study, we selected chemotherapy alone as the postoperative adjuvant modality rather than chemoradiotherapy, which commonly is used for the postoperative adjuvant treatment of pancreaticobiliary carcinoma in both the U.S. and Europe. Although there are advantages and disadvantages with both modalities, we selected chemotherapy alone for the following reasons. Extensive surgery is one of the favorable surgical procedures for the treatment of pancreaticobiliary carcinoma in Japan. Especially aggressive Japanese surgeons have attempted to remove more advanced tumors of the biliary tract in combination with hepatectomy and/or pancreatoduodenectomy, although to our knowledge no scientific evaluation of this procedure has been undertaken by Japanese surgeons.25–28 This extensive surgery is highly invasive and associated with a high incidence of surgical complications. Therefore, we chose chemotherapy as the postoperative adjuvant method that would be most likely to help us avoid such complications. We hoped that through these precautions we would be able to maintain patient safety and prevent the development of a higher rate of complications in the postoperative adjuvant therapy group compared with patients treated with surgery alone. The highly invasive magnitude of this surgery can be a disadvantage, and discontinuation of postoperative adjuvant therapy frequently may be required. Discontinuation of postoperative chemoradiotherapy because of delayed recovery has been reported in 24% of patients in a study by Spitz et al.19 and in 28% of patients in a study by Yeo et al.2 In contrast, treatment compliance in the current study was high, with approximately 96% of patients receiving MMC and 79–96% receiving intravenous 5-FU. Treatment had to be discontinued as a result of delayed recovery in only 1.1% of patients.
Earlier reports of postoperative adjuvant therapy for the treatment of carcinoma of the biliary tract have indicated the potential usefulness of radiotherapy and chemotherapy.10, 11, 14–16 Nearly all these studies were based on cases of noncurative resection, and were experimental reports rather than clinical trial reports. One group of authors, Mahe et al., reported an MST of 22 months,14 which is similar to the MST reported in the adjuvant therapy group in the current study (20.7 months in the patients with bile duct carcinoma and 16.3 months in the patients with gallbladder carcinoma). Previous results were obtained in a mixed population (i.e., patients with curative resection and palliative resection). Thus, the benefit of adjuvant therapy for carcinoma of the biliary tract have not yet been clarified.
We have no explanation why postoperative adjuvant chemotherapy was useful only in patients with gallbladder carcinoma. Results of in vitro drug sensitivity testing in the study laboratories, using the collagen gel droplet-embedded culture drug sensitivity test, have shown the efficacy of 5-FU and MMC to be 18.8% and 25.0%, respectively, in patients with pancreatic carcinoma and 50.0% and 50.0%, respectively, in patients with biliary tract carcinoma, indicating the particularly good sensitivity of biliary tract carcinoma to these agents.29 In in vitro experiments, biliary tract carcinoma demonstrated a higher sensitivity to these agents compared with pancreatic carcinoma. Our group also has performed two randomized controlled studies of nonresected pancreaticobiliary carcinomas. In both studies, we found that combined chemotherapy with 5-FU, doxorubicin, and MMC (FAM) was effective in gallbladder carcinoma only, a finding that is consistent with the findings of the current study. In the what to our knowledge is the first study of 5-FU monotherapy compared with modified FAM therapy, patients with gallbladder carcinoma who received modified FAM therapy demonstrated improvements in PS and symptoms, and had a tendency toward longer survival.30 In the second study of FAM therapy, in which a half-dose level (1/2 FAM) was used in comparison with best supportive care, 1/2 FAM therapy demonstrated a significantly longer time to disease progression, but only in gallbladder carcinoma patients.31 These findings suggest that the results of drug sensitivity testing described earlier, as well as the results from our previous two clinical trials of patients with nonresected disease, support to some extent the results of the current study, which indicate a significantly prolonged survival time in gallbladder carcinoma patients only.
In this trial, 5-FU (100 mg/m2) was administered orally each day from postoperative Week 5 until disease recurrence. Although the peak plasma concentrations after the oral administration of 5-FU (100 or 200 mg/body) had been reported to vary between 30–750 ng/mL,32 a clinical benefit for postoperative adjuvant chemotherapy was expected in this trial and therefore this treatment was included in the protocol.
Molecular biologic approaches to biliary tract carcinoma have demonstrated immunohistochemical p53 and ras p21 protein positivity and the staining patterns appear to be different between gallbladder carcinoma and carcinomas of the bile duct or ampulla of Vater,33–37 suggesting that these tumors may differ in etiology, pathogenesis, and molecular origin.33, 34 Such differences at the molecular level are known to affect drug sensitivity. For example, the expression of p53, bcl-2, and bax38, 39 and, more recently, the mRNA expression level of the 5-FU target enzyme thymidylate synthase38–41 and its catabolic enzyme dihydropyrimidine dehydrogenase41 in tumor tissue, have been found to influence the effects of 5-FU. Further research is necessary to clarify these points, but these findings may indicate specific microbiologic effects of 5-FU in gallbladder carcinoma.
The postoperative adjuvant chemotherapy regimen used in the current study was not found to be useful in the treatment of carcinomas of the pancreas, bile duct, or ampulla of Vater. Bakkevold et al. evaluated the effects of systemic chemotherapy in a randomized controlled study of adjuvant combination chemotherapy comprised of FAM compared with surgery alone after radical resection of carcinoma of the pancreas or ampulla of Vater.8 To our knowledge this was the only randomized controlled study published to date that evaluated the usefulness of systemic chemotherapy as a postoperative adjuvant modality. The authors reported an MST of 23 months in the FAM group, which was significantly longer than the 11 months noted in the group treated with surgery alone, but Bakkevold et al. found no significant difference in long-term survival between the 2 groups after 2 years.8 The GITSG study, which was representative of randomized controlled trials comparing a combination of radiotherapy and chemotherapy with surgery alone, reported the results of curative pancreatic resection followed by external radiotherapy at a dose of 40 grays (Gy) combined with systemic 5-FU at a dose of 500 mg/m2/day.17 The MST was 20 months and the 2-year survival rate was 43% in the combined postoperative adjuvant therapy group in the GITSG study. These figures were significantly better than the MST of 11 months and 2-year survival rate of 18% reported in the group of patients treated with surgery alone.17 The additional GITSG study of 30 patients using the same treatment demonstrated an MST of 18 months and a 2-year survival rate of 46%, confirming the reproducibility of their previous findings.12 Results from previous trials, including the GITSG study, suggest an MST of 18–23 months after curative resection of pancreatic carcinoma followed by external radiotherapy in a dose range of 35–60 Gy combined with the systemic administration of 5-FU,2, 12, 13, 17–19 whereas the MST after surgery alone has been reported to be between 11–12 months.2, 12 Although these findings are not directly comparable to our study because of differences with regard to patient characteristics, results from the current study with regard to MTS (13.2 months in MF group and 14.6 months in the control group; data not shown) in patients with curatively resected pancreatic carcinoma are somewhat poorer than those described earlier.
The results of the current study indicate that moderately toxic MF therapy was not efficacious as a postoperative adjuvant treatment modality in patients with pancreatic carcinoma. It is evident from the pattern of disease recurrence noted in the current study that future research should focus on local treatment to inhibit hepatic recurrence. It also will be necessary to conduct large-scale, randomized controlled studies that focus on postoperative adjuvant therapy with a combination of radiotherapy and chemotherapy, and to evaluate the usefulness of gemcitabine ± radiotherapy and preoperative therapy. At the current time, a clinical study to evaluate treatment with preoperative and postoperative intravenous 5-FU combined with perioperative and postoperative radiotherapy is underway. We also are evaluating the usefulness of postoperative prophylactic intrahepatic arterial infusion.
The results of the current randomized controlled clinical trial indicate that patients who received postoperative therapy after resection of their gallbladder carcinoma had an overall improved survival; however, the improvement in survival appeared to be limited to the subgroup of patients who underwent noncurative resections. We believe that these results will be encouraging to many practicing clinicians.
The results of the current study were from a multicenter trial performed at the following medical institutions. The authors would like to express their gratitude to all the cooperating surgical oncologists who were at the institutions at the time the study was performed: Hiroshi Kato, Hokkaido University School of Medicine, Hokkaido; Koichi Hirata, Sapporo Medical University, Hokkaido; Mutsuo Sasaki, Hirosaki University School of Medicine, Aomori; Kazuyoshi Saito, Iwate Medical University School of Medicine, Iwate; Takashi Matsushiro, Tohoku Rosai Hospital, Miyagi; Hidemi Yamauchi, Sendai National Hospital, Miyagi; Tadashi Yamazaki, Sendai Open Hospital, Miyagi; Tadahiro Takada and Hideki Yasuda, Teikyo University School of Medicine, Tokyo; Goro Watanabe, Toranomon Hospital, Tokyo; Tetsuro Kajiwara, Tokyo Women's Medical College Daini Hospital, Tokyo; Munemasa Ryu, National Cancer Center Hospital East, Chiba; Shuichi Okada, National Cancer Center Hospital, Tokyo; Hitoshi Hanaue, Tokai University Oiso Hospital, Kanagawa; Masayuki Uchimura, Hamamatsu Clinic Center, Shizuoka; Naokazu Hayakawa, Tokai Hospital, Aichi; Yuji Nimura, Nagoya University School of Medicine, Aichi; Shuichi Miyakawa, The Department of Miura Surgery, Fujita Health University School of Medicine, Aichi; Takahiko Funahiki, The Department of Funahiki Surgery, Fujita Health University School of Medicine, Aichi; Akihiro Yamaguchi, Ogaki Municipal Hospital, Gifu; Takeaki Shimizu, Shinrakuen Hospital, Niigata; Keisuke Yoshida, Niigata University School of Medicine, Niigata; Takukazu Nagakawa, Kanazawa University School of Medicine, Ishikawa; Shigeki Takashima, Kanazawa Medical University, Ishikawa; Hiroshi Shimada, Fukui Medical School, Toyama; Hiroshige Nakano, Nara Medical University, Nara; Yuji Nishino, The 1st Department of Surgery, Osaka City University Medical School, Osaka; Hiroaki Kinoshita, The 2nd Department of Surgery, Osaka City University Medical School, Osaka; Hiroshi Ashida, Hyogo College of Medicine, Hyogo; Kazutoyo Shirakawa, Mitoyo General Hospital, Kagawa; Seiyo Ikeda, Fukuoka University School of Medicine, Fukuoka; Toshimichi Nakayama, Kurume University School of Medicine, Fukuoka; Seiki Tashiro, Kumamoto University School of Medicine, Kumamoto; and Koro Sakota, Kagoshima Medical Association Hospital, Kagoshima.
- 20Japan Pancreatic Society. Classification of pancreatic carcinoma, 1st English ed., Tokyo: Kanehara, 1996.
- 21Japanese Society of Biliary Surgery. General rules for surgical and pathological study on cancer of the biliary tract, 4th ed. Tokyo: Kanehara, 1997.
- 22Japan Society for Cancer Therapy. Japanese criteria for the evaluation of the direct effects of cancer chemotherapy for solid tumors. J Jpn Soc Cancer Ther. 1986; 21: 943–968.
- 24Regression models and life tables. J R Stat Soc B. 1972; 34: 187–220..
- 27Hepatopancreatoduodenectomy for carcinoma of the gallbladder invasive to both the liver and the hepatoduodenal ligament [in Japanese with English abstract]. Nippon Geka Gakkai Zasshi. 1987; 88: 1332–1335., , , et al.
- 28Hepato-ligament pancreatoduodenectomy [in Japanese]. Geka Chiryo. 1988; 59: 12–21., , .
- 32Clinical trial of the oral carcinostatica, 5-fluorouracil dry syrup, with special reference to plasma and tissue concentrations of the drug [in Japanese]. Jpn J Cancer Chemother. 1975; 2: 297–303., , .