Optimal Adjuvant Therapy for Resected Pancreatic Cancer: Chemotherapy or Chemoradiotherapy?

Abstract

Despite recent research and advances in the understanding of the molecular and genetic basis of pancreatic cancer, the poor outcomes experienced by pancreatic cancer patients have changed little during the past 30 years. Adenocarcinoma of the pancreas is the fourth leading cause of cancer-related death in the United States, with only a small fraction of patients achieving long-term survival. According to data from the American Cancer Society, 5-year survival for pancreatic cancer patients is 5%; an estimated 33,730 newly diagnosed cases of pancreatic cancer will be nearly equaled by an estimated 32,300 pancreatic cancer deaths in the United States. This underscores the continued need to develop novel multimodality treatment approaches for this disease. Surgery has proved vital to a curative-intent treatment approach for these patients. However, only 10% to 20% of newly diagnosed patients present with potentially resectable nonmetastatic disease. In light of the minority of patients with resectable disease, there has been considerable debate over the potential advantages of adjuvant chemotherapy and radiotherapy. In recent decades, cooperative groups both in the United States and abroad have conducted randomized clinical trials seeking to define the potential benefit of adjuvant chemotherapy or adjuvant chemoradiotherapy vs. surgery alone for patients with resectable disease. Unfortunately, the results of these trials have been conflicting and no definitive consensus has yet been reached regarding optimal adjuvant therapy. This article reviews cooperative group data pertinent to this debate. It is suggested that (1) patients do benefit from adjuvant therapy, and (2) optimal adjuvant therapy should include gemcitabine-based chemoradiotherapy for select patient subgroups.

Although we have made progress toward understanding the molecular and genetic basis of pancreatic cancer, outcomes for pancreatic cancer patients have not improved significantly in the past 30 years. Adenocarcinoma of the pancreas represents the fourth leading cause of cancer-related deaths in the United States, with only a small fraction of patients achieving long-term survival. The American Cancer Society reports 5-year survival for pancreatic cancer patients to be 5%, and the estimated 33,730 newly diagnosed cases of pancreatic cancer in the United States will be nearly matched by an estimated 32,300 pancreatic cancer deaths.¹ Clearly, there is a pressing need to develop novel multimodality approaches to treating this disease.

Surgery has proved vital to a curativeintent treatment approach for these patients, but only 10% to 20% of newly diagnosed patients present with nonmetastatic, potentially resectable disease.^1,2 There has been considerable debate over the potential advantages of adjuvant chemotherapy and radiotherapy for the minority of patients with resectable disease. In recent decades, cooperative groups both in the United States and abroad have conducted randomized studies attempting to define the potential benefit of adjuvant chemotherapy or adjuvant chemoradiotherapy vs. surgery alone. Due in part to conflicting results of these trials, no definitive consensus regarding optimal adjuvant therapy has been reached.

Considering current national treatment recommendations in the United States, the National Comprehensive Cancer Network (NCCN) advises that clinical trials are preferred for patients in the adjuvant setting and that either chemotherapy or chemoradiotherapy is appropriate, which serves to highlight the controversy over optimal adjuvant therapy.³ In this review, we will focus on cooperative group data pertinent to this debate.

RATIONALE FOR ADJUVANT THERAPY

The relatively high rate of both locoregional and distant disease recurrence in pancreatic cancer following surgery provides a strong rationale for the addition of adjuvant therapy.^4,5 Though a consensus on what defines optimal adjuvant therapy has not yet been reached, one can make a strong argument that it will likely prove to be a combination of chemotherapy, radiotherapy, and probably biologic therapy. If one considers adenocarcinomas of other gastrointestinal sites, the improved locoregional control conveyed by postoperative chemoradiotherapy has been demonstrated to improve overall survival in phase III cooperative group studies, and indeed, postoperative chemoradiotherapy is considered the standard of care for those disease sites.^6,7

Chemoradiotherapy is effective at eradicating microscopic residual disease in the locoregional field of treatment. Consequently, patients likely to derive a survival benefit from adjuvant chemoradiotherapy are those whose residual microscopic disease is limited to the locoregional area, typically included in the radiation field. The retroperitoneal location of the pancreas and its proximity to major vessels often makes gross total resections with wide margins difficult. Patients with close or positive microscopic margins are those most likely to benefit from the addition of radiotherapy in the adjuvant setting. In considering adjuvant chemotherapy vs. adjuvant chemoradiotherapy, future trial designs should include strategies (such as sequence of therapy and use of prognostic factors) to select for the patient subgroups expected to benefit most from adjuvant multimodality therapy. A review of the existing cooperative group randomized trial data provides insight into what likely represents optimal adjuvant therapy.

ADJUVANT CHEMOTHERAPY

Randomized trials of adjuvant chemotherapy vs. observation in patients with resected pancreatic cancer are relatively limited. Early trials included 5-fluorouracil (5-FU)–based chemotherapy, whereas more recent trials have evolved toward gemcitabine-based regimens. Criticisms of these studies have included lack of sufficient statistical power and inclusion of nonpancreatic cancers. A summary of the outcomes of these studies is provided in Table 1.

Table 1.

Randomized multicenter phase III trials of observation vs. adjuvant chemotherapy following surgical resection.

Author	No. pts.	Regimen	Overall survival (actuarial)				Median survival (months)
			1 year	2 year	3 year	5 year
Bakkevold et al8	30	5-FU+DOX+MMC	70%	NR	27%	4%	23*
	31	Control	45%	NR	30%	8%	11
Takada et al9	81	5-FU+MMC	NR	NR	NR	12%	NR
	77	Control	NR	NR	NR	18%	NR
Neoptolemos (ESPAC-1 )10	147	5-FU+LV	NR	40%	NR	21%	20.1*
	142	Control	NR	30%	NR	8%	15.5
Oettle et al (CONKO-001)11	179	GEM	72%	41%	20%	8%	22.1†
	177	Control	72%	37%	14%	5%	20.2

Abbreviations: 5-FU = 5-fluorouracil; DOX = doxorubicin; GEM = gemcitabine; LV = leucovorin; MMC = mitomycin C; NR = not reported.

^*Statistically significant

^†Trend toward significance

Norwegian Trial

In the earliest prospective, randomized multicenter trial, Bakkevold et al reported the results of 61 Norwegian patients with pancreatic (n = 47) or periampullary (n = 14) cancers who underwent resection and were subsequently randomized to observation vs. chemotherapy.⁸ The chemotherapy consisted of 5-FU (500 mg/m²), doxorubicin (40 mg/m²), and mitomycin C (6 mg/m²) given once every 3 weeks for a total of six cycles. Chemotherapy resulted in an increase in median survival (23 vs. 11months, P = .02) and a delay to recurrence, but ultimately, long-term survival was not statistically different in the two arms. The observed 1-, 2-, 3-, and 5-year survivals in the treatment group were 70%, 43%, 27%, and 4% compared with 45%, 32%, 30%, and 8% in the control group, respectively.

Toxicity of adjuvant therapy was substantial, with less than half of the patients in the treatment arm able to complete the planned six cycles of therapy. Furthermore, since the analysis did not stratify for pancreatic cancer and peripampullary disease (which is associated with better outcomes), it is difficult to draw conclusions from this study regarding the benefit of adjuvant chemotherapy for patients with true pancreatic cancer.

Japanese Trial

In Japan, Takada et al conducted a randomized multicenter trial that included pancreatic cancer patients as well as patients with tumors of the gall bladder, bile duct, or ampulla of Vater.⁹ Unlike the Norwegian study, though, the results of this trial were stratified by disease type; resected pancreatic cancer patients represented 173 of the 508 patients enrolled. Following resection, patients were randomized to observation vs. chemotherapy. Chemotherapy consisted of mitomycin C (6 mg/m² IV) at the time of surgery and 5-FU (310 mg/m² IV) given weekly during postoperative weeks 1 and 3. Thereafter, patients received maintenance oral 5-FU (100 mg/m²) until disease recurrence. Based on the analysis of 158 eligible patients, there appeared to be no benefit to adjuvant chemotherapy. The difference in 5-year survival between the two groups—11.5% for patients receiving adjuvant chemotherapy (n = 81) vs. 18% for the control group (n = 77)—was not statistically significant.

ESPAC-1

Neoptolemos et al reported the results of the European Study Group for Pancreatic Cancer (ESPAC-1) trial, which employed a 2 × 2 factorial design to compare the effects of chemoradiotherapy and chemotherapy on survival following resection for adenocarcinomas of the pancreas.¹⁰ The design of this trial was somewhat complex in that patients were randomly assigned to receive chemoradiotherapy or chemotherapy, neither treatment or both treatments. The chemotherapy consisted of IV bolus 5-FU (425 mg/m²) and leucovorin (20 mg/m²) given for 5 days every 28 days for a total of six cycles. Considering only the patients in the 2 × 2 design of observation (n = 142) vs. chemotherapy (n = 147), there was a statistically significant advantage in median survival associated with adjuvant chemotherapy (20.1 months) when compared with observation (15.5 months). This trial is discussed in greater detail later in this article.

CONKO-001

Mature results of the latest prospective, randomized multi-institutional phase III Charité Onkologie trial (CONKO-001) were recently published by Oettle et al.¹¹ In the largest study reported to date, 368 patients were randomized to receive either adjuvant chemotherapy with gemcitabine (n = 179) or observation (n = 177) following gross resection (R0 or R1) of their pancreatic cancer. Of note, no surgical or histologic standards were identified in this study. Gemcitabine was given at a dose of 1,000 mg/m² on days 1, 8, and 15 of a 4-week cycle for a total of 6 months. Patients randomized to adjuvant chemotherapy began treatment within 6 weeks of their surgery. The primary end point of this study was disease-free survival, and the study was powered to demonstrate a 6-month improvement with gemcitabine.

Eighty percent of the patients had an R0 resection, 70% were node-positive, and the two groups were well matched for all demographic, clinical, and tumor characteristics. At a median follow-up of 53 months, there was a significant improvement in median disease-free survival for patients receiving gemcitabine (13.4 vs. 6.9 months, P < .001). Median overall survival trended better for patients receiving gemcitabine, but did not reach statistical significance (22.1 vs. 20.2 months, P = .06).

In summary, two of three 5-FU–based randomized trials demonstrated a median survival advantage with adjuvant chemotherapy vs. observation. The only randomized gemcitabine-based study demonstrated an improvement in median disease-free survival and a trend toward improved overall survival (though the study was powered for the former end point). Taken together, these data suggest that adjuvant therapy in pancreatic cancer conveys a survival advantage over observation. The potential advantage of adding radiotherapy to chemotherapy in the adjuvant setting is considered in the following section.

ADJUVANT CHEMORADIOTHERAPY

Randomized cooperative group trials of adjuvant chemoradiotherapy in pancreatic cancer are also limited. There are three trials of chemoradiotherapy vs. observation: GITSG 9173, EORTC 40891, and ESPAC-1. All of these trials employed a suboptimal “split-course” radiotherapy schedule to what most would consider an inadequate total dose (40 Gy) with respect to microscopic disease control. Nevertheless, the results of these trials have provided the foundation and momentum to further refine multimodality adjuvant therapy as represented in the design of the recently reported RTOG 97-04 trial. The results of the above trials are summarized in Table 2 and are discussed in detail below.

Table 2.

Randomized multicenter phase III trials evaluating adjuvant chemoradiotherapy.

Overall survival (actuarial)
Author	No. pts.	Regimen	2 year	5 year	Median survival (months)
GITSG13	21	5-FU + RT(sc)	43%	19%	21.0*
	22	Control	19%	5%	10.9
Klinkenbijl et al (EORTC)14	60	5-FU + RT(sc)	37%	20%	17.1†
	54	Control	23%	10%	12.6
Neoptolemos et al (ESPAC-1)18	Pooled Data Analyses
	145	CRT	29%	10%	15.8
	144	no CRT	41%	20%	17.8*
	147	CT	40%	21%	20.1*
	142	no CT	30%	8%	15.5
	2 × 2 Subanalysis
	69	Control	NR	11%	16.9
	75	CT	NR	29%	21.6
	73	CRT	NR	7%	13.9
	72	CRT + CT	NR	13%	14.2
Regine et al (RTOG 97-04)23	187	5-FU + CRT + 5-FU	NR	NR	16.9
	194	G + CRT + G	NR	NR	20.6*

Abbreviations: 5-FU = 5-fluorouracil; CT = chemotherapy (5-FU); CRT = chemoradiotherapy (5-FU + RT); G = gemcitabine; LV = leucovorin; NR = not reported; RT(sc) = radiation therapy, split course.

^*Statistically significant

^†Trend toward significance

GITSG 9173

The first prospective, randomized multicenter trial of adjuvant chemoradiotherapy vs. observation was performed by the Gastrointestinal Tumor Study Group (GITSG). Resected patients with negative margins were randomly assigned to splitcourse radiotherapy (n = 21) or no adjuvant treatment (n = 22). Adjuvant therapy consisted of radiotherapy (40 Gy in 6 weeks, with a 2-week break after the first 20 Gy) combined with chemotherapy (5-FU during weeks 1 and 5 of radiotherapy followed by maintenance 5-FU for 2 years or until disease progression). Accrual was slow, but a positive interim analysis demonstrated a significant benefit to treatment with adjuvant chemoradiotherapy. Patients randomized to receive adjuvant treatment exhibited statistically significant improvements in median survival compared with the control group (21 vs. 11 months), 2-year survival (43% vs. 19%), and 5-year survival (19% vs. 5%).¹² Therapy in this trial was well tolerated, but the power of the study was limited. Criticisms of the trial included poor patient accrual (limited power), lack of radiotherapy quality assurance, and the fact that over a third of the patients in the treatment arm did not complete maintenance therapy as intended.

In response to criticisms leveled at the initial study, including inadequate sample size, GITSG sought to confirm the results of the randomized trial with an additional cohort of 32 patients assigned to the identical adjuvant chemoradiotherapy treatment. This nonrandomized cohort of patients achieved outcomes similar to those in the randomized trial, with median, 2-year, and 5-year survivals of 18 months, 46%, and 17%, respectively.¹³ The findings of this randomized trial led to the adoption of adjuvant chemoradiotherapy as the standard of care in the United States.

EORTC 40891

In the second randomized, multicenter trial, the European Organisation for Research and Treatment of Cancer (EORTC) randomized resected patients to chemoradiotherapy or observation. This trial was similar to the GITSG design, save for the following: (1) nonpancreatic periampullary adenocarcinomas were included in the study, (2) no maintenance chemotherapy was included in the treatment arm, (3) patients with positive surgical margins were allowed in the study without stratification, and (4) chemotherapy consisted of continuous infusion 5-FU instead of bolus 5-FU. When compared with patients randomly assigned to observation, the 110 patients assigned to adjuvant chemoradiotherapy showed no significant improvements in median survival, 2-year survival, or 5-year survival.¹³ As in the GITSG trial, adjuvant treatment was well tolerated with no grade 4 or 5 toxicities.

There have been numerous criticisms of this trial, including (1) the inclusion of patients with positive margins without stratification, (2) the lack of radiotherapy quality assurance, (3) the absence of maintenance chemotherapy, (4) the inclusion of nonpancreatic periampullary cancers (which are associated with better prognosis), and (5) inappropriate statistical design. To address the criticism of the inclusion of the periampullary cancers, the authors reported a subanalysis including only the patients with pancreatic head cancers (n = 114). Median survival for patients with pancreatic head cancers treated with adjuvant chemoradiotherapy vs. observation was 17.1 months vs. 12.6 months, 2-year survival rates were 37% vs. 23%, and 5-year survival¹⁴ rates were 20% vs. 10% in favor of adjuvant chemoradiotherapy. Statistical significance was not reached (P = .099); however, the power to detect a statistical significance was diluted as is the case with most subanalyses. Further, it has been demonstrated that with a more appropriate statistical design (onesided log-rank test), EORTC 40891 would have been statistically significant as a positive trial for adjuvant chemoradiotherapy in patients with pancreatic cancer.¹⁵

ESPAC-1

As mentioned previously, the largest reported phase III trial was conducted by the European Study Group for Pancreatic Cancer. The ESPAC-1 trial included patients with resected adenocarcinomas of the pancreas (regardless of surgical margin status) and sought to establish the benefit of adjuvant chemotherapy or chemoradiotherapy when compared with observation. A total of 541 patients were randomized to one of three arms: (1) chemotherapy vs. no chemotherapy, (2) chemoradiotherapy vs. no chemoradiotherapy, or (3) a 2 × 2 factorial design of observation vs. chemotherapy vs. chemoradiotherapy with maintenance chemotherapy. An early intent-to-treat analysis of this trial at a median follow up of 10 months was reported by Neoptolemos et al and suggested no benefit to chemoradiotherapy (compared with no chemoradiotherapy) and a benefit to chemotherapy (compared with no chemotherapy) with pooled data from all three arms.¹⁶

The results of this early analysis came under criticism for a multitude of reasons.¹⁷ The mature results of ESPAC-1 were reported in 2004, and analysis was restricted to patients in the third arm of the trial (the 2 × 2 factorial design).¹⁸ In the 2 × 2 subanalysis, patients were either assigned to observation, chemotherapy, chemoradiotherapy, or chemoradiotherapy and further chemotherapy. After a median follow-up of 47 months, the estimated 5-year survival of patients randomized to chemoradiotherapy vs. no chemoradiotherapy was 10% vs. 20% (P = .05). The patients who received chemotherapy had a significantly higher 5-year survival rate when compared with those who did not receive chemotherapy (21% vs. 8%, P = .009). Detailed results of this complex trial can be found in Table 2.

The investigators concluded that the use of adjuvant chemotherapy resulted in a significant improvement in overall survival and that postoperative chemoradiotherapy worsened survival because it delayed the administration of chemotherapy. The results of this trial generated substantial controversy, resulting in several letters to the editor expounding on the problems in the design and conduct of this study.^19–22 One example of the criticism was lack of centralized quality assurance of radiotherapy dose and fields. Per protocol, patients were intended to receive a total of 40 Gy. However, only 75% of patients received this dose. Some patients received less than this target dose, while others received up to 60 Gy. In addition, the 2 × 2 factorial design allowed for “background therapy” (chemoradiotherapy or chemotherapy), which can directly influence compliance with future randomized treatments and thus potentially influence results.

Irrespective of the criticisms of this trial, it should be noted that the controversial results of ESPAC-1 stand alone with respect to the conclusion that adjuvant chemoradiotherapy worsens survival. The authors concluded from this trial that adjuvant chemotherapy should be considered standard adjuvant therapy.

RTOG 97-04

The Radiation Therapy Oncology Group (RTOG) 97-04 study represents the most recent randomized cooperative group trial evaluating adjuvant chemoradiotherapy and was reported by Regine et al at the 2006 annual meeting of the American Society of Clinical Oncology.²³ This Intergroup trial was designed to evaluate pre- and post-chemoradiotherapy with 5-FU vs. pre- and post-chemoradiotherapy with gemcitabine. This study was a multi-institutional collaborative effort involving the RTOG, the Eastern Cooperative Oncology Group, and the Southwest Oncology Group between July 1998 and July 2002.

To avoid some of the criticisms of past trials, namely, lack of radiotherapy quality control, all radiation treatment fields and diagnostic studies (imaging films, reports, and operative notes) were centrally reviewed and approved prior to the start of therapy. As previously mentioned, the ESPAC-1 trial has been criticized for lack of radiotherapy quality assurance (permitting a wide range of doses, up to as high as 60 Gy). An audit of the Intergroup’s own phase III trial in gastric cancer revealed that 35% of patients had major or minor deviations with respect to RT fields and dose, underlying the importance of prospective, centralized quality assurance.²⁴

A total of 538 patients were enrolled, of which 442 were eligible for participation. Failure to send serum for CA-19-9 analysis (n = 22) and treatment starting more than 8 weeks after surgery (n = 19) were major causes of patient ineligibility. All patients enrolled in the study had resected pancreatic adenocarinoma of pathologic stages T1-4N0-1M0. Patients were stratified according to surgical margins (positive vs. negative), tumor diameter (< 3 cm vs. ≥ 3 cm), and nodal status (involved vs. uninvolved). Patients were randomized to one of two treatment arms: (1) 5-FU, followed by 5-FU–based chemoradiotherapy, followed by further 5-FU, or (2) gemcitabine, followed by 5-FU–based chemoradiotherapy, followed by further gemcitabine. Chemotherapy (3 weeks prior to and 3 months after surgery) consisted of either gemcitabine 1,000 mg/m² or 5-FU 250 mg/m²/day. The chemoradiotherapy component of the treatment was the same in both arms and consisted of continuouscourse (not split-course) fractionated radiotherapy to a total dose of 50.4 Gy combined with concomitant 5-FU (250 mg/m²/day) delivered by continuous infusion throughout the 5.5 weeks of radiotherapy.

Based on 381 patients with pancreatic head tumors eligible for analysis, a statistically significant survival advantage was seen for patients treated with gemcitabine-based therapy. The median survival and 3-year overall survival for patients treated with gemcitabine-based chemoradiotherapy compared with 5-FU–based chemoradiotherapy was 20.6 months vs. 16.9 months and 32% vs. 21%, respectively. With prospective radiotherapy quality assurance on RTOG 97-04, only 5% of patients treated on protocol were found to have an unacceptable protocol variance. Abrams et al previously reported the effect of radiotherapy quality assurance and compliance on survival in RTOG 97-04 in abstract form at the 2006 American Society for Therapeutic Radiology and Oncology annual meeting.²⁵

On multivariate analysis, nodal status (positive vs. negative) and tumor diameter (< 3 cm vs. ≥ 3 cm) were statistically significant prognostic factors for survival. Although hematologic toxicity was greater in the gemcitabine-based arm, 85% to 90% of patients were able to complete therapy as per the study design. Based on these results, the authors concluded that gemcitabine was superior to 5-FU and that future adjuvant chemoradiotherapy trials should build upon a gemcitabine-based foundation.

DISCUSSION

Although the data from randomized cooperative group trials in pancreatic cancer are not free of flaws, in aggregate they do suggest a benefit for adjuvant therapy in pancreatic cancer when compared with observation alone. Three of the four randomized trials of adjuvant chemotherapy vs. observation are positive for a median survival advantage.^8,10,11 Further, two of the three cooperative group trials comparing chemoradiotherapy to observation are positive for adjuvant chemoradiotherapy.^13–15

Several trials have helped to define the optimal chemotherapy basis upon which to build future trials. Single-agent gemcitabine has been established to be superior to 5-FU based on a randomized trial in 128 patients with symptomatic, locally advanced pancreatic cancer.²⁶ Patients treated with gemcitabine had a significantly better clinical benefit response when compared with those treated with 5-FU (23.8% vs. 4.8%). Median survival and response rate were also superior for gemcitabine. These data from the locally advanced population, taken together with the preliminary results of the CONKO-001 trial and the RTOG 97-04 trial, suggest that a gemcitabine-based regimen is effective and superior to a 5-FU–based approach when considering optimal adjuvant therapy in the resectable population.

The collective results of the GITSG, EORTC, and RTOG trials provide compelling evidence that optimal adjuvant therapy for patients with cancers of the pancreatic head should consist of chemoradiotherapy built around gemcitabine. The regimens in all three of these trials were remarkably well tolerated, and the RTOG trial demonstrated that almost 90% of patients could complete therapy as per protocol. Given the heterogeneity of the results of the adjuvant chemoradiotherapy trials, it is likely that differences in patient groups or risk factors contributed at least in part to the differences seen with respect to the benefit of adjuvant chemoradiotherapy.

Tumor size, surgical margin status, and lymph node status are characteristics that have been previously reported to influence survival.²⁷ The recently reported results of the RTOG 97-04 trial have helped to validate these prognostic factors on multivariate analysis. Such factors should be incorporated into future trial designs to help define the patient subgroups most likely to benefit from adjuvant chemoradiotherapy. Further, centralized prospective quality assurance should be incorporated into all future trials to limit the degree to which heterogeneity in treatment contributes to differences in outcomes. Radiotherapy should consist of modern, fractionated, three-dimensionally planned treatment delivered in continuous-course fashion to biologically effective doses.

Although the results of existing randomized trials suggest a benefit to chemoradiotherapy, several questions remain unanswered. Optimal sequencing of adjuvant chemoradiotherapy has yet to be defined. One consideration for future trial design would be to use sequencing as a strategy to select for patients most likely to benefit from aggressive locoregional treatment with chemoradiotherapy. An example of this strategy would be to treat resected pancreatic cancer patients with three cycles of gemcitabine and then reimage to assess for evidence of distant disease progression. In the absence of evidence of progression, patients would be randomized to either chemoradiotherapy or further gemcitabine. With this treatment approach, patients destined for early distant metastases (and therefore unlikely to benefit from aggressive locoregional therapy) are identified prior to potential “overtreatment,” and only those patients most likely to benefit from aggressive locoregional chemoradiotherapy go on to receive treatment. In addition to sequencing, novel biologics such as bevacizumab and erlotinib have shown promise and should be explored in combination with a gemcitabine-based approach.

Well-conceived randomized trials conducted within the cooperative group setting will continue to be vital in optimizing multidisciplinary care, and continued efforts toward improving patient selection will be critical in resolving the controversy surrounding adjuvant treatment of pancreatic cancer.