INTRODUCTION About 33,730 people develop pancreatic adenocarcinoma each year in the United States, and almost all are expected to die from the disease . Surgical resection offers the only chance of cure. However, only 15 to 20 percent of patients have resectable disease at diagnosis; approximately 40 percent have metastatic disease, and another 40 percent have locally advanced unresectable tumors. Typically these patients have tumor adherence or invasion into adjacent structures, particularly the celiac and superior mesenteric vasculature. Median survival is 8 to 12 months for patients with locally advanced disease, and less for those with metastatic disease at presentation. Optimal treatment for locally advanced unresectable pancreatic cancer is controversial. This topic review will focus on the therapeutic options, which include radiation therapy (RT) alone, chemotherapy alone, and combined chemoradiotherapy with or without surgery. CRITERIA FOR UNRESECTABILITY As a general rule, pancreatic cancers are considered unresectable if any of the following are present :
An assessment of resectability is usually made based upon a preoperative staging computed tomography (CT) scan and less commonly at the time of laparotomy. A variety of other imaging modalities, including MRI, endoscopic ultrasound (EUS), as well as staging laparoscopy, may be needed in some circumstances. This topic is addressed in detail elsewhere. EXTERNAL BEAM RT ALONE Locally advanced pancreatic tumors can usually be encompassed in a conventional external beam RT (EBRT) portal. However, in most settings, EBRT alone does not provide optimal palliation or tumor control; local failure rates are as high as 72 percent . These dismal results have spawned efforts to improve outcomes by the concurrent administration of drugs that act as radiation sensitizers, including 5-fluorouracil (5-FU), gemcitabine, and most recently, paclitaxel. CONCOMITANT CHEMORADIOTHERAPY As has been seen in a variety of tumors arising in the gastrointestinal tract, the addition of concurrent chemotherapy to EBRT improves outcomes compared to EBRT alone in patients with locally advanced pancreatic cancer. From the standpoint of symptom palliation, cancer-related pain is diminished in 35 to 65 percent of patients, and cachexia and obstructive symptoms may also improve. However, the survival benefit from chemoradiotherapy relative to EBRT alone is modest, and local control remains problematic. 5-FU-based approaches Most studies have evaluated combined 5-FU and EBRT, which are synergistic in other gastrointestinal malignancies. While the majority have compared chemoradiotherapy to EBRT alone, the control arm in a few studies was supportive care only or chemotherapy alone. Early trials Two early randomized trials directly comparing EBRT with and without concomitant 5-FU-based chemotherapy came to opposite conclusions:
Infusional 5-FU was used in a trial from the Eastern Cooperative Oncology Group (ECOG) which randomly assigned 114 patients to RT (59.4 Gy) alone or with concurrent infusional 5-FU (1000 mg/m2 daily on days 2 through 5 and 28 to 31) plus mitomycin (10 mg/m2 on day 2) . In contrast to the GITSG trial, the addition of chemotherapy to RT increased toxicity without providing any benefit in terms of response rate (9 versus 6 percent), median disease-free survival (5 versus 5.1 months) or overall survival (7.1 versus 8.4 months, respectively). The method of 5-FU administration (intermittent high-dose rather than the more contemporary method of continuous infusions of lower doses such as 200 mg/m2 daily during EBRT) as well as the inclusion of mitomycin C might have contributed to the relatively poor outcomes in the experimental group. Medicare/SEER report A benefit for chemoradiotherapy as compared to RT alone was further supported by a retrospective but large series of 1700 elderly patients with locally advanced pancreatic cancer derived from a linked Medicare/SEER database . Forty-four percent of patients received some form of therapy, and the adjusted mean survival durations for patients undergoing chemoradiotherapy, RT alone, chemotherapy alone, and no therapy were 47, 29, 27, and 15 weeks, respectively. Although not derived from a randomized trial, these data support the view that chemoradiotherapy provides benefit over either modality alone, and that treatment appears to be better than supportive care. The authors performed a sophisticated statistical analysis using propensity score methods to minimize the impact of selection bias. However, the possibility that patients who are considered good candidates for chemoradiotherapy may survive longer due to other unmeasurable factors can only be excluded in a randomized trial. Versus supportive care alone The relatively modest benefits of treatment also raise the question of whether combined chemotherapy and EBRT are better than supportive care alone. In order to address this issue, Japanese investigators conducted a small study in which 31 patients with locally unresectable pancreatic cancer were randomly assigned to EBRT (50.4 Gy in 28 fractions) with concurrent 5-FU or no chemoradiation . Despite the small number of patients, treatment was associated with statistically significant improvements in median survival (13.2 versus 6.4 months), one-year survival (53 versus 0 percent), average monthly Karnofsky score (77.1 versus 65.5), and the number of days spent in the hospital per month of survival (12.3 versus 19.0 days). The treatment arm in this trial reflects the most frequently used method of concurrent 5-FU administration, at least in the United States. Capecitabine as a substitute for infusional 5-FU As noted above, largely based upon the experience in other gastrointestinal malignancies such as rectal cancer, infusional rather than bolus 5-FU has become the most commonly used approach for radiation sensitization. Accumulating data support the view that oral capecitabine can safely replace continuous infusional 5-FU as a radiation sensitizer in patients treated for locally advanced pancreatic cancer . However, randomized trials comparing this approach to standard infusional 5-FU regimens are not available. This approach will likely gain momentum if an ongoing trial by the National Surgical Breast and Bowel Project (NSABP) of capecitabine versus infusional 5-FU in patients undergoing RT for rectal cancer shows therapeutic equivalence. (See "Neoadjuvant chemoradiotherapy for rectal cancer", section on Oral fluoropyrimidines). Gemcitabine-based chemoradiotherapy The observation that gemcitabine has potent radiation sensitizing effects, in conjunction with its demonstrated clinical benefit in metastatic pancreatic cancer provided the rationale for investigating its use in patients with locally advanced disease A variety of gemcitabine treatment schedules and doses have been studied, limiting the ability to compare outcomes among clinical trials.
On the basis of these results, the Cancer and Leukemia Group B conducted a phase II trial of low-dose twice weekly gemcitabine (40 mg/m2 twice weekly) plus concurrent RT (50.4 Gy) in 43 patients with locally unresectable nonmetastatic pancreatic cancer . Treatment was relatively toxic: grade 3 or 4 hematologic toxicity in 60 percent, grade 3 or 4 gastrointestinal toxicity in 42 percent, and one treatment-related death attributed to sepsis. Although median overall survival was a disappointing 8.2 months, two patients were still alive at 35 and 41 months of posttreatment follow-up.
Gemcitabine versus 5-FU These small trials have not addressed which agent (gemcitabine versus 5-FU) is optimal for use with concurrent RT in patients with locally advanced disease. This issue was addressed in a retrospective series from MD Anderson Cancer Center which compared the outcomes of 53 patients with locally advanced pancreatic cancer receiving concurrent gemcitabine (250 to 500 mg/m2 weekly) and a relatively low dose EBRT (30 Gy in 10 fractions) with 61 similar patients undergoing concurrent infusional 5-FU (200 to 300 mg/m2 daily) with the same dose of EBRT . Acute treatment-related toxicity was more common in the patients treated with gemcitabine (23 versus 2 percent), and the rates of local progression (62 versus 61 percent), distant metastases (55 versus 47 percent), one-year survival (42 versus 28 percent), and median survival (11 versus 9 months) were not significantly different. The only randomized trial directly comparing both approaches randomly assigned 34 patients with locally advanced disease to chemoradiotherapy with either concurrent gemcitabine (gemcitabine 600 mg/m2 weekly for six weeks plus three dimensional conformal EBRT 50.4 to 61.2 Gy) or bolus 5-FU (5-FU 500 mg/m2 daily for three days every two weeks for six weeks plus the same EBRT regimen) . There were no significant differences in the toxicity profile between the two groups. Patients receiving gemcitabine were significantly more likely to have their pain controlled (39 versus 6 percent), and they also had lived significantly longer (median survival 14.5 versus 6.7 months). The small size of this study and the unconventional schedule of 5-FU administration makes it difficult to interpret the differences in outcome. Thus, there is little evidence to support better outcomes in patients undergoing chemoradiotherapy for locally advanced pancreatic cancer with concomitant gemcitabine as compared to concurrent daily infusional 5-FU (the usual regimen, at least in the United States). Combining gemcitabine and 5-FU The ECOG attempted a trial in which 5FU, gemcitabine, and RT were combined in the treatment of locally unresectable pancreatic cancer; however, excess toxicity forced early closure after only seven patients were treated . On the other hand, a preliminary report of a subsequent phase II trial of this approach conducted by the CALGB suggested acceptable toxicity and a median survival of 11.4 months . Further experience with this approach is needed. Paclitaxel plus EBRT Paclitaxel is a potent radiation sensitizer, and may be particularly beneficial for the many pancreatic cancers that harbor p53 mutations Experience with this approach is limited but promising:
Summary Conventional EBRT plus concomitant daily infusional 5-FU chemotherapy provides a modest survival benefit compared to supportive care alone. Despite the lack of definitive trials that substantiate benefit relative to RT alone, combined modality therapy with infusional 5-FU is often considered the standard of care for locally advanced pancreatic cancer in the United States. However, given the marginal survival benefit of such therapy, it is unclear whether chemotherapy alone with gemcitabine would demonstrate similar results. Possible role for surgery The role of surgery in patients undergoing chemoradiotherapy for locally advanced disease is unclear. The available data are scant, and consist mainly of small single institution series The three largest series are described in detail:
Despite these encouraging results, neoadjuvant chemoradiotherapy is not yet a standard approach for patients with locally advanced unresectable disease. While patients should be reevaluated for potential resectability after chemoradiotherapy, the frequency with which this culminates in a complete resection is low, particularly in those initially deemed unresectable. In addition, long-term survival appears to be exceedingly rare in this circumstance. Whenever possible, such patients should be offered treatment in the context of a clinical trial. CHEMOTHERAPY ALONE Chemotherapy alone is an increasingly utilized treatment option for patients with locally advanced disease. The data to support the use of chemotherapy in this setting come mainly from two sources: chemotherapy trials conducted in mixed populations of patients with both locally advanced and metastatic disease, and the few trials that have directly compared 5-FU-based chemoradiotherapy versus 5-FU or gemcitabine chemotherapy alone. Contemporary trials evaluating different chemotherapy combinations in mixed populations of patients with locally advanced and metastatic pancreatic cancer suggest that the impact of gemcitabine-based chemotherapy on survival may be of approximately the same magnitude as that achieved by chemoradiotherapy.
Although it is inherently hazardous to perform cross-study comparisons, these median survival values are similar to those reported in series of patients treated with chemoradiotherapy In the setting of advanced pancreatic cancer, whether the addition of any agent to gemcitabine provides benefit over gemcitabine alone remains a controversial issue. Of the many randomized trials that have been conducted using various gemcitabine doublets, only one (erlotinib plus gemcitabine) has shown a survival benefit over gemcitabine alone, and this was extremely modest (median survival 6.4 versus 5.9 months) . As a result, at many institutions, gemcitabine alone is considered the standard approach when chemotherapy alone is indicated. Trials comparing chemotherapy and chemoradiotherapy A few trials have directly compared chemoradiotherapy to chemotherapy alone (either 5-FU-based or gemcitabine), none of which provide a definitive answer as to which approach is better. 5-FU-based chemotherapy Two early studies comparing chemoradiotherapy to 5-FU-based chemotherapy conducted in the 1980s came to opposite conclusions:
Gemcitabine-based chemotherapy In contrast to these results, a survival benefit could not be shown for chemoradiotherapy followed by gemcitabine compared to gemcitabine alone in the French FFCD-SFRO trial, presented at the 2006 American Society of Clinical Oncology meeting . The chemoradiotherapy arm consisted of EBRT (60 Gy in 30 fractions) plus concomitant 5-FU (300 mg/m2 over 24 hours five days per week) and cisplatin (20 mg/m2 on days 1 through 5 during weeks 1 and 5) followed by gemcitabine (1000 mg/m2 weekly for three of every four weeks) until progression. The chemotherapy alone group received gemcitabine (1000 mg/m2 weekly for seven of the first eight weeks, then for three of every four weeks), also until progression. The trial was halted because of poor accrual after 109 of the planned 176 patients were enrolled. In a preliminary report (median 16-month follow-up), the group receiving gemcitabine alone had significantly better one-year (82 versus 51 percent) and median survival (14.3 versus 8.4 months), as well as significantly less treatment-related (particularly hematologic) toxicity. Interpretation of these results is limited by the fact that they have been presented only in abstract form, the relatively high dose of RT (60 Gy compared to the more usual 54 Gy in the setting of locally advanced disease), and the likely contribution of concurrent cisplatin to the toxicity profile. Thus, this trial should be interpreted as showing that RT to 60 Gy with concurrent 5-FU and cisplatin appears to be inferior to gemcitabine alone; it does not demonstrate that gemcitabine alone is superior to all types of conventional 5-FU-based or gemcitabine-based chemoradiotherapy. Unfortunately, a randomized ECOG trial (ECOG 4201) of gemcitabine versus gemcitabine plus RT in patients with localized but unresectable pancreatic cancer was opened, but terminated due to poor accrual. Chemotherapy followed by chemoradiotherapy A newly proposed strategy for limiting the use of chemoradiotherapy to those who are most likely to benefit includes initial gemcitabine-based chemotherapy, followed by chemoradiotherapy for those without evidence of disease progression. This strategy was evaluated in a retrospective series of 181 patients with locally advanced, unresectable but nonmetastatic pancreatic cancer who had been treated with gemcitabine-based chemotherapy alone as part of phase II and III trials conducted by the European Groupe Cooperateur Multidisciplinaire en Oncologie (GERCOR) . In each protocol, chemotherapy was initially given for three months, and the decision to continue it or administer chemoradiotherapy (55 Gy external beam RT with concurrent infusional 5-FU) was left to the discretion of the investigator. Among the 128 patients who did not progress after three months of chemotherapy and who retained an adequate performance status, 72 received chemoradiotherapy while 56 continued with chemotherapy. When the two groups were compared, chemoradiotherapy was associated with significant improvement in median progression-free survival (10.8 versus 7.4 months with chemotherapy) and overall survival (15 versus 11.7 months). These data suggest that chemoradiotherapy could improve survival relative to continued chemotherapy alone after a period of initial disease control with chemotherapy. Another interpretation is that early chemotherapy allows the selection of patients who do not have occult micrometastatic disease, thus limiting RT to those who might benefit from improved local control. Prospective evaluation of this treatment strategy is needed to validate the benefit of this approach; such a phase III trial is underway in Europe. SYMPTOM PALLIATION Although chemoradiotherapy can frequently relieve pain and may successfully palliate obstructive symptoms, specific treatment may be needed to provide relief of obstructive jaundice, gastric outlet obstruction, and unrelieved pain. Here we will provide a brief overview of palliation for specific symptoms. Obstructive jaundice Randomized trials demonstrate that endoscopically placed stents are as successful as surgical bypass procedures in relieving the obstructive jaundice caused by pancreatic cancer, but with lower morbidity and procedure-related mortality rates. Most jaundiced patients can be successfully managed using stents, and relatively few require surgical biliary-enteric bypass as a planned palliative procedure. Stenting can be performed either percutaneously or endoscopically. In a randomized trial compared these two approaches in 75 patients with malignant bile duct obstruction, patients treated endoscopically had a higher success rate for relief of jaundice (81 versus 61 percent), and a significantly lower 30-day mortality rate (15 versus 33 percent) . However, for some patients, especially those who have undergone pancreaticoduodenectomy, endoscopic stent placement may not be feasible, and percutaneous approaches are needed for adequate palliation. The preferred endoscopic stent for palliation is an expandable metal stent, which is superior to plastic stents for long-term palliation. Metal stents should only be used when the possibility of surgical resection has been excluded. Plastic stents (usually made of polyethylene) are available and can be removed, but have the disadvantage of requiring frequent stent change (usually every three months). Metal stents may not provide permanent palliation. Stent occlusion develops in 7 to 42 percent of patients, the most common cause being tumor ingrowth. The addition of a coating over the wire mesh (such as polyurethane) may reduce the rate of occlusion, although studies examining this possible benefit have produced mixed results. Occluded stents are usually best managed by endoscopic insertion of a second metal stent or a plastic stent. Gastric outlet obstruction Gastroduodenal obstruction caused by pancreatic cancer can be managed by gastrojejunostomy, although experience is increasing with endoscopically placed expandable metal stents. Patients who have both gastric outlet obstruction and obstructive jaundice may be best served by surgical gastric/biliary bypass. Palliation of pain Palliation of pain caused by pancreatic adenocarcinoma can usually be successfully achieved by drug treatment alone. Celiac plexus block may be required for pain control in selected patients and may be achieved by a surgical, radiographically guided percutaneous or endosonographically guided approach. SUMMARY AND RECOMMENDATIONS Approximately 40 percent of patients with pancreatic cancer present with locally advanced, unresectable nonmetastatic disease. The optimal management of these patients is controversial. Therapeutic options include radiation therapy (RT) alone, chemoradiotherapy, and chemotherapy alone.
Whenever possible, such patients should be encouraged to enroll on clinical trials testing new approaches. Ongoing clinical research efforts are focusing upon the evaluation of new systemic agents to combine with EBRT, and improving staging methods to better select those patients who might benefit from a combined perioperative approach. |