| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
NEW APPROACHES TO THE TREATMENT OF HEPATIC MALIGNANCIES |
From the Departments of Surgical Oncology (QDC, HOD, JFG) and Medical Oncology (NK, MMJ), Roswell Park Cancer Institute, State University of New York at Buffalo, Buffalo, New York.
Correspondence: Address correspondence and reprint requests to: John F. Gibbs, MD, Department of Surgical Oncology, Roswell Park Cancer Institute, Elm and Carlton Street, Buffalo, NY 14263; Fax: 716-845-2320; E-mail: john.gibbs{at}roswellpark.org
ABSTRACT
Adenocarcinoma of the pancreas continues to be a formidable disease. In the United States, patients who have had resected disease are generally offered adjuvant chemoradiation. The current National Comprehensive Cancer Network practice guidelines uniformly support this practice. We reviewed seven selected series to evaluate the efficacy of adjuvant therapy for patients who had resected adenocarcinoma of the pancreas. Current evidence-based analysis demonstrates that an adjuvant therapy regimen as a standard of care is lacking. We, therefore, believe that it should be used judiciously because its benefit is confined to only a fraction of patients treated by complete resection (R0); patients with residual microscopic disease (R1) derived negligible benefits. Given the financial constraints and the small effect that current therapies have on this fatal disease, clinicians should concentrate on developing novel therapies and new paradigms to address this age-old problem.
Key Words: Pancreatic adenocarcinoma • Adjuvant therapy • GITSG trials • EORTC trial • ESPAC-1 trial • Norwegian trial
Adenocarcinoma of the pancreas continues to be a formidable disease. It is the fourth leading cause of cancer-related death for both men and women in the United States.1 It is estimated that in 2003, 30,700 new cases of adenocarcinoma of the pancreas will be diagnosed and that approximately 30,000 patients will die of the disease.1 Unlike with other cancers of epithelial origin (e.g., breast and colon), the prognosis of patients with pancreatic cancer is dismal, with fewer than 5% of patients surviving beyond 5 years.2 Most patients present with advanced disease, with only 10% to 15% of patients having disease confined to the pancreas (T1/2N0).2 It is this group of patients in whom there is the greatest potential for surgical cure. However, long-term survival after curative resection is only between 10% and 30%.2 The logical approach to managing patients with resected pancreatic cancer, therefore, is to evaluate whether adjuvant therapy might further improve survival.
The role of adjuvant therapy for solid tumors, such as breast and colorectal carcinomas, has been well established by well-controlled clinical trials and, therefore, has been accepted as the standard of care in these diseases. Given the current enthusiasm about adjuvant therapy for these tumors, should the same paradigm also be applicable in patients with resected adenocarcinoma of the pancreas? The current National Comprehensive Cancer Network (NCCN) guidelines classified the use of adjuvant therapy under category 2A (uniform NCCN consensus, based on lower-level evidence—including clinical experience—that the recommendation is appropriate).3
We reviewed seven selected studies to evaluate the merits of adjuvant therapy for resected adenocarcinoma of the pancreas. The role of neoadjuvant therapy will not be reviewed in this article.
EARLIER WORKS
Historical data in the late 1960s and early 1970s from Haslam et al.4 and Moertel et al.5 demonstrated that a good tumor response occurred when radiation alone or combined chemoradiation was used to treat patients with locally unresectable adenocarcinoma of the pancreas. On the basis of such seminal works, the Gastrointestinal Tumor Study Group (GITSG) began evaluating the efficacy of high dose radiation only (6000 cGy), high dose radiation with 5-fluorouracil (5-FU), and moderate dose radiation (4000 cGy) with 5-FU for regional adenocarcinoma of the pancreas.6 They were able to demonstrate that high dose radiation alone was inferior to the combined-modality therapy. Additionally, moderate dose radiation with 5-FU was just as effective as high dose radiation with 5-FU. These results served as an impetus for the ensuing multi-institutional trials.
CURRENT CLINICAL TRIALS
GITSG 9173 Trial
The GITSG was the first group that performed a prospective, randomized trial to evaluate the role of adjuvant chemoradiation in resected adenocarcinoma of the pancreas.7 The study was intended to be a cooperative effort among 14 different institutions, but in reality, most patients were accrued from two major institutions: Roswell Park Cancer Institute and the University of Miami. To have a better appreciation of the efforts required to perform the study, one needs to understand the prevailing viewpoint at that time. Most surgeons did not subscribe to the multidisciplinary approach of cancer treatment and were therefore reluctant to enroll patients onto an adjuvant study that had potential toxicity without known benefits. As a result, the accrual rate was very slow; however, some excitement was created when a doubling of the median survival time was observed in patients who had combined adjuvant therapy compared with surgery alone (20 vs. 11 months; P = .03). The study was terminated early when the interim analysis demonstrated a significant survival advantage in the adjuvant arm, in addition to the slow accrual rate. GITSG 9173 was originally designed to enroll 100 patients (50 in each arm) to have a 90% power to detect a doubling of survival time. Instead, only 43 patients were accrued over an 8-year period. Other problems with the trial included poor compliance and quality assurance. Twenty-four percent of the patients had treatment that was initiated beyond the defined protocol-specified time limit, and only 9% of the patients had the intended 2 years of chemotherapy. In addition, 32% of patients had violations of the scheduled radiation treatment. Selection bias was also a problem with the study, as it is with any adjuvant treatment. Those patients who recovered quickly were more likely to be included in the trial. Finally, the confidence intervals for median survivals of the two groups were large enough that they could easily overlap each other, but survivals at 5 and 10 years were significantly different (18% of treated patients vs. 0% of control patients).8
Given the previously described arguments against GITSG 9173, a confirmatory treatment trial was performed.9 In this trial, the investigators were able to accrue an additional 32 patients onto the treatment arm in 28 months and thereby verified their original findings. Unfortunately, this study also had several major drawbacks. There was no control arm, because the design was only to confirm and replicate the results of treatment in the controlled trial. Twenty percent of the patients in the confirmatory trial had severe hematological sequelae, compared with 14% in the first trial, despite the fact that the same chemotherapy regimen was used in both studies, although different institutions were participating. The confirmatory trial did have more patients with a better performance status than either group in the first trial. Although the authors acknowledged that there was a slight discrepancy in the Eastern Cooperative Oncology Group (ECOG) performance status of patients in the confirmatory trial (i.e., more patients had an ECOG performance status of 0 or 1 compared with the control or the treatment arms of the original study), they concluded that when only the fully ambulatory patients were evaluated, the median survivals for both the patients randomized to the treatment arm and the confirmatory patients were similar. One needs to take into account that this represents subset analysis performed in two underpowered studies. One should also realize that the first adjuvant trial demonstrated that ECOG (Zubrod performance scale) performance status was a significant predictor of overall survival, as has been the case in all chemotherapy trials.
Therefore, it can be concluded that the positive results from the GITSG trials may have been due to an alpha error. The authors, however, should be applauded for a monumental study that was perhaps ahead of its time.
Norwegian Trial
A second, less well known, controlled, prospective, randomized, multicenter trial of adjuvant therapy was performed by the Norwegian group.10 Sixty-one resected patients with adenocarcinoma of the head of the pancreas and papilla of Vater from 38 hospitals were randomized either to adjuvant chemotherapy or a control arm. The trial evaluated the efficacy of adjuvant combination chemotherapy—5-FU (500 mg/m2), doxorubicin (40 mg/m2), and mitomycin C (6 mg/m2)—once every 3 weeks for six cycles. Postoperative radiation was not used. The characteristics between the two arms were fairly similar. The study demonstrated that the time to recurrence in the first 2 years after surgery was prolonged in the treatment group and that the median survival doubled for those who received the adjuvant combination chemotherapy regimen versus the control group (23 vs. 11 months). However, there was no difference in the 2-year survival between the two groups (P = .10). It is interesting to note that the 5-year survival was worse by half in the treatment group when compared with the control group (4% in the treatment group vs. 8% in the control group), although this was not statistically significant. Compliance was a major problem because of the toxicity from the regimen. Only 13 (43%) of 30 patients were able to complete the 6 scheduled courses. There were five septic complications, of which one patient died from causes likely attributed to chemotherapy. Two patients each had cardiotoxicity and nephrotoxicity due to chemotherapy.
European Organization for Research and Treatment of Cancer Trial
After the positive results of the GITSG trials, the European Organization for Research and Treatment of Cancer (EORTC) initiated its own trial11 in 1987. This was a large, prospective, controlled, randomized phase III trial that attempted to either validate or refute the GITSG trials. The study was a cooperative effort by 29 centers in Europe and assessed the role of adjuvant radiotherapy and 5-FU for pancreatic head and periampullary adenocarcinomas; the latter are routinely included in pancreatic protocols in Europe but have a better prognosis than cancers of the pancreatic head.12 The chemoradiotherapy regimen differed from the GITSG trials. 5-FU was given as a continuous infusion during the period of radiotherapy instead of as a bolus injection after radiation. Although there was no quality control for radiotherapy, the EORTC group was able to achieve acceptable quality control, as well as compliance (93% of patients received the planned split course of 40 Gy of radiation), and there was no delay of treatment, because patients who had prolonged hospital stay were not randomized. The median age, performance status, location of the primary tumor, resection margin status, histology, and tumor and nodal stations were similar in the treatment and observations arms.
The results of this study demonstrated that although the regimen was safe and well tolerated, it did not have a significant effect on the overall 5-year survival. The median survivals for the treatment and the observation arms were 24 and 19 months, respectively, and the 5-year survivals for the treatment and observation arms were 28% and 22%, respectively (P = .208). It may be argued, however, that the study may have been markedly underpowered in that it did not accrue sufficient numbers of patients onto each of the treatment arms. This is particularly true when one recognizes that this trial recruited patients not only with pancreatic cancer, but also with other periampullary tumors. When the data are carefully scrutinized, there was a trend toward significance in favor of the treated group for cancer of the head of the pancreas. Both the median and the 2-year survivals favored the treatment arm (P = .099). Although these are legitimate concerns, one should be careful with them. One cannot theoretically extrapolate the data on the basis of a P value trending toward significance and then assume that had there been more patients accrued, the data would have demonstrated significance in favor of the treatment arm. From a purely statistical point of view, a value that trends toward significance should not have the same connotation as those values that are indeed statistically significant. Finally, although subgroup analysis for pancreatic head cancer reveals a doubling of the 5-year survival in the treatment arm, the confidence interval is so wide that it easily overlaps with the interval in the observation group.
Despite some of the study’s shortcomings, what is more intriguing is that not only did the regimen have no effect on survival, but it also did not alter the natural history of the disease—the percentage of documented progression (65% for the treatment arm vs. 66% for the observation arm), the site of first progression, and the site of distant progression were similar in both groups.
Johns Hopkins Data
Two years before the publication of the EORTC report, Yeo et al.13 from the Johns Hopkins hospital reviewed their experience with 174 patients with resected pancreatic cancer and substantiated the results of the GITSG trials. The study offered patients who had resected cancer one of three options: (1) standard adjuvant chemoradiation (modified from the GITSG protocol), (2) intensive adjuvant chemoradiation (external beam radiotherapy to the pancreatic bed with prophylactic hepatic irradiation followed by 5-FU plus leucovorin), or (3) no further therapy. The Cox proportional hazards survival analysis identified a superior survival advantage with the use of adjuvant therapy when compared with no further therapy. The median survival was 19.5 months for the treatment group and 13.5 months for the no-further-therapy group (P = .003). These results were very similar to those in the GITSG initial and confirmatory trials.7,9 Unfortunately, the trial was not a prospective, randomized trial and, consequently, introduced selection bias. Fifty-seven percent of patients selected to have the standard therapy, 12% had the intensive therapy, and 31% selected no treatment. Although the groups seem similar, one cannot discern why the patients would choose one treatment option over another. One can hypothesize that younger patients with better performance status who had uneventful recovery may have chosen to undergo more aggressive therapy, whereas older patients with prolonged recuperation periods may have opted for observation. In fact, the data demonstrated that older patients, those who had a longer postoperative stay, and those who had more postoperative complications tended to choose the no-therapy option. Although age, length of stay, and complication rate were shown by univariate analysis not to influence survival, this could be due to insufficient power to detect such a difference. Similar to the EORTC trial, this group also reported no alteration of the natural history of the disease in terms of disease progression with adjuvant chemoradiation. The study should be commended, however, because it was able to demonstrate that a difficult operation can be performed in a large cohort of patients with acceptable morbidity (36%) and mortality (.6%).
European Study Group for Pancreatic Cancer Trial
The European Study Group for Pancreatic Cancer (ESPAC-1) is the largest randomized trial that evaluated the role of adjuvant chemotherapy alone and split-course chemoradiotherapy after resection of pancreatic adenocarcinoma but also included patients with ampullary cancers and those with cancer at the margin of resection.14 The study was a cooperative effort by 83 clinicians in 61 cancer centers across 11 countries. This report is an interim analysis on 541 patients, and the final analysis will be performed when all patients have been followed up for at least 2 years. The study set out to recruit 280 patients into a 2 x 2 factorial design to detect an excess of 20% deaths at 2 years between each main comparison at the 5% significance level with 90% power. Because the 2 x 2 factorial design restricted randomization, the trial was expanded to allow randomization options of chemoradiotherapy or chemotherapy. Comprehensive analysis were performed on all patients, and a separate analysis was performed on patients in the 2 x 2 factorial design. The randomization process was well balanced between treatment groups, and protocol violations occurred in only 9% of patients, but the choice of randomization options introduced an indeterminate degree of bias.
The interim results refuted the results from the GITSG trials, demonstrating no survival advantage for patients who had adjuvant chemoradiotherapy versus no chemoradiation. The median survival for the treatment group was 15.5 months (95% confidence interval, 13.5–17.4 months) versus 16.1 months for the no-chemoradiation group (95% confidence interval, 13.1–20.1 months). However, when the authors evaluated the effect of adjuvant chemotherapy versus no chemotherapy, there was a statistically significant difference in median survival (P = .005). The median survival for patients in the chemotherapy group was 19.7 months (95% confidence interval, 6.4–22.4 months) versus 14.0 months (95% confidence interval, 11.9–16.5 months) for the no-chemotherapy group. Surprisingly, however, when they evaluated the 2 x 2 factorial design, which was the authors’ original intent, there was no statistically significant survival advantage within the chemotherapy group (P = .19).
The study demonstrated that resection margin status was an independent factor only in the absence of tumor grade and nodal status.15 There was no survival advantage for using chemoradiation, irrespective of resection margin status. The median survival for the chemoradiation group was 15.5 months versus 16.1 months for the no-chemoradiation group (P = .24). However, there was an overall survival advantage of >5 months in favor of the chemotherapy group. The median survival for the chemotherapy group was 19.7 months versus 14.0 months for the no-chemotherapy group (P = .0005). However, when one analyzes the effect of chemotherapy for R0 versus R1 patients, the additional 5-month survival advantage was more for the R0 group than for the R1 group. The median survival for R1 patients who had chemotherapy was 11.0 months versus 10.3 months for the no-chemotherapy group, a negligible benefit of .7 months. Finally, one should cautiously weigh the benefit of the additional 5-month gains against the 27% risk of patients developing a serious grade 3 or 4 treatment-related toxicity. In fact, most of the toxicity (64%, or 28 of 44 patients) occurred in the chemotherapy-only group.
One of the major problems with the study is the complicated study design, which makes it very difficult to interpret the data. For example, the authors included a group of 70 patients who were randomized to the chemoradiotherapy group in their no-chemotherapy group analysis. Additionally, in their no-chemoradiotherapy group, they did not include the 96 patients who were randomized to the no-chemotherapy group. Moreover, radiotherapy quality control was suboptimal in the study because some of the facilities enrolling patients lacked modern radiotherapy equipment.16 In addition, there was latitude given to radiation oncologists who could administer up to 60 Gy, although >75% of patients received the median total radiation dose of 40 Gy.
It is interesting to note that the median survival for the surgery-only group in this trial (16 months) is very similar to the median survival for the adjuvant group in the GITSG trials (18–20 months), which were initiated almost 15 years earlier. Improved survival in the control group may account for the lack of a statistically significant benefit to the adjuvant chemoradiation group. The reason for such differences is not likely to be due to surgical techniques, because the extent of pancreatic resection, whether subtotal or total, did not influence survival,17,18 but it may be associated with improved long-term management of impaired motility, pancreatic insufficiency, chronic pancreatic fistula, and diabetes mellitus. One, however, should refrain from making too much of comparisons between the two studies because the study designs were very different. However, is it possible that, given the advances in the long-term care of resected patients, the need for adjuvant therapy could be bypassed?
Despite these shortcomings, the lack of any clear benefit from adjuvant chemoradiation in this trial and the EORTC trial, both of which used the same chemoradiotherapy regimen, is clear. The future ESPAC trial will not include radiotherapy in any of its treatment arms.
Meta-Analysis of Adjuvant Chemoradiotherapy and Chemotherapy
The results of a meta-analysis of adjuvant chemoradiotherapy and chemotherapy were recently published in the 2002 Proceedings of the American Society of Clinical Oncology.19 The analysis included data from the GITSG, EORTC, Norwegian Pancreatic Cancer Trials Group, and ESPAC-1 trials. The meta-analysis confirmed the ESPAC-1 results, showing no overall benefit for chemoradiation (risk reduction, -4%; P = .71) and a benefit for chemotherapy alone (risk reduction, 35%; P < .001). The results, however, should be interpreted with caution because they were heavily dependent on the ESPAC-1 trial because of its large size.
CURRENT PERSPECTIVES ON ADJUVANT THERAPY
Comparison of the trials that have influenced our perception of the benefit of adjuvant therapy after resection of pancreas cancer is difficult. These trials have different designs and inclusion criteria for what is deemed as locally resectable disease. In the original GITSG trial, only one third of the cases had cancer confined to the pancreatic gland, whereas the rest had spread to adjacent nodes or contiguous structures, mainly the duodenum. In the EORTC trial, T1/2N0/1aM0 pancreatic head or T1–3N0/1aM0 periampullary cancers were included, whereas involvement of adjacent organs was excluded. The current classification of locally resectable versus locally advanced cancers depends on radiological results and on the surgeon’s intraoperative assessment. The generally accepted notion that a Whipple procedure performed by one surgeon is similar to a Whipple procedure performed by another surgeon should be questioned. This factor is fraught with variability because it depends on the surgeon’s personal experience. What may deemed to be an unresectable situation by one surgeon (i.e., close adherence to the portal vein) may be deemed resectable by another. Many series now include patients in whom portal vein resections have been performed.20,21 Consequently, any large study involving more than one surgeon that lacks rigorous surgical and pathologic quality assurance is likely to result in the inclusion of a heterogeneous population of patients, and this biases the data.
The inclusions of subtypes of cancer also differ. The GITSG trials included cancers located at the head and body of the pancreas and excluded periampullary cancers. The Johns Hopkins group and the ESPAC-1 group excluded duodenal and ampullary cancers, whereas the EORTC and the Norwegian group included periampullary cancers in their analysis.
Furthermore, chemotherapy administration was also not uniform. Unlike the GITSG trials, the EORTC and the ESPAC-1 trials gave 5-FU during and not after radiotherapy. In addition, chemotherapy was given as a continuous infusion in the EORTC and ESPAC-1 trials rather than by bolus injection. Finally, unlike the GITSG trials, chemotherapy was not given over a 2-year period in the EORTC and ESPAC-1 trials.
With the current financial constraints in medical care, financial prudence is paramount when clinical protocols are designed. The total cost of pancreatic cancer care in the United States is approximately $2.6 billion.2 Therefore, financial consideration is integral if we are to concentrate our efforts in the care of these patients. Currently, there are several competing single-institutional protocols that diminish the accrual rate and prolong study times. The National Cancer Institute listed >100 current clinical trials performed nationwide on pancreatic cancer.22 The question of optimal resource utilization and the need to have a multi-institutional approach should be the driving forces when clinical trials are constructed. It has taken clinicians literally decades to embrace the concept of a multidisciplinary approach to cancer management, and it is hoped that such time can be bypassed for the concept of a multi-institutional approach. Established cooperative groups are handicapped by the prolonged gestation time of protocol development.
Finally, to put things in perspective, let us look at the hard numbers. Approximately 30,000 new cases of pancreatic adenocarcinoma are diagnosed per year in the United States. However, only approximately 15%, or 4500 cases, are resectable. Let us assume that all 4500 patients qualify for the combined adjuvant therapy; only 20% (900 cases per year) of patients are estimated to be alive in 5 years. Approximately half of these patients will survive beyond 5 years, and the rest will die of recurrent or metastatic disease.23 Therefore, the theoretical maximal benefit of combined therapy is 10% to 20% (450–900 per 4500 cases) in patients with resectable disease.
The quality of life for treated patients with adenocarcinoma of the pancreas has been addressed in the past by several investigators.24–26 Quality of life is an important issue and should be included in any future studies because any survival benefit from therapy needs to justify the quality-of-life costs for the patient. Although there is general consensus that adjuvant therapy is well tolerated, toxicity may not be negligible. The ESPAC-1 trial reported a 27% incidence of serious toxic effects with the adjuvant regimen. A quality-adjusted time without symptoms or toxicity analysis has been demonstrated to be an excellent method to compare treatment outcomes for different cancers,27–29 and similar analysis can also be performed on the available data for patients with adenocarcinoma of the pancreas.
CONCLUSIONS
Given the currently available clinical trials, one can surmise that the argument for adjuvant therapy for resected adenocarcinoma of the pancreas is not strongly supported by the data (Table 1). It would seem that the current practice of administering adjuvant therapy for resected adenocarcinoma of the pancreas arises from the clinician’s known bias of its traditionally poor outcome. Of the six prospective clinical trials, the GITSG trials are the only randomized, prospective trials that demonstrated a survival benefit with adjuvant chemoradiation. Dissection of other trials revealed flaws that would make it very difficult to make an argument for standard adjuvant therapy. The final results of the ESPAC-1 trial are still pending, but the interim results, although confusing because of patient allocations, are not very encouraging for advocating adjuvant chemoradiation. In fact, the authors contend that chemoradiation might even have an adverse effect on the benefits of chemotherapy because of the delay of its first use.15 This group favors the use of chemotherapy, but critical analysis of their data makes it not very palatable. The Johns Hopkins data are retrospective, and, therefore, their positive results should be viewed with caution. Finally, the EORTC trial, which was successful in achieving well-balanced groups for comparison, demonstrated not only that there was no benefit in adjuvant therapy, but also that the natural history of the disease was not altered by the therapy. Given our current understanding of the disease, where do we go from here?
|
Footnotes
There is a general consensus that adjuvant chemoradiation should be considered as a standard of care for patients with resected adenocarcinoma of the pancreas. A critical review of seven clinical studies demonstrates that such a practice is likely to be based on clinicians’ biases of its traditional poor outcome.
Received for publication June 20, 2002. Accepted for publication January 9, 2003.
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  M. M. Corsini, R. C. Miller, M. G. Haddock, J. H. Donohue, M. B. Farnell, D. M. Nagorney, A. Jatoi, R. R. McWilliams, G. P. Kim, S. Bhatia, et al. Adjuvant Radiotherapy and Chemotherapy for Pancreatic Carcinoma: The Mayo Clinic Experience (1975-2005) J. Clin. Oncol., July 20, 2008; 26(21): 3511 - 3516. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Javle, J. F. Gibbs, K. K. Iwata, Y. Pak, P. Rutledge, J. Yu, J. D. Black, D. Tan, and T. Khoury Epithelial-Mesenchymal Transition (EMT) and Activated Extracellular Signal-regulated Kinase (p-Erk) in Surgically Resected Pancreatic Cancer Ann. Surg. Oncol., December 1, 2007; 14(12): 3527 - 3533. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. S. Chadha, T. Khoury, J. Yu, J. D. Black, J. F. Gibbs, B. W. Kuvshinoff, D. Tan, M. G. Brattain, and M. M. Javle Activated Akt and Erk Expression and Survival After Surgery in Pancreatic Carcinoma Ann. Surg. Oncol., July 1, 2006; 13(7): 933 - 939. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Talamonti, W. Small Jr., M. F. Mulcahy, J. D. Wayne, V. Attaluri, L. M. Colletti, M. M. Zalupski, J. P. Hoffman, G. M. Freedman, T. J. Kinsella, et al. A Multi-Institutional Phase II Trial of Preoperative Full-Dose Gemcitabine and Concurrent Radiation for Patients With Potentially Resectable Pancreatic Carcinoma Ann. Surg. Oncol., February 1, 2006; 13(2): 150 - 158. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Yeo The Recent Past and Future of Adjuvant Therapy for Pancreatic Cancer Ann. Surg. Oncol., June 1, 2003; 10(5): 488 - 489. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
