Chemotherapy and radiotherapy for pancreatic cancer: Adjuvant, neoadjuvant, and palliative

Adjuvant chemoradiotherapy

Early studies examining the role of external beam radiotherapy (EBRT) combined with 5-fluorouracil (5-FU) as a radio-sensitizing agent in advanced, unresectable, pancreatic adenocarcinoma identified increased median survival over radiotherapy alone. ^, One study published in 1979 by the Gastrointestinal Tumour Study Group (GITSG) demonstrated a survival benefit with radiotherapy and concurrent chemotherapy. In this trial, 43 patients were randomized, following potentially curative pancreatic resection, to either chemoradiotherapy (40 Gy EBRT, concurrent with 5-FU and subsequent 5-FU for up to 2 years) or no adjuvant treatment. Median, 2-year, and 5-year survival rates were all increased in the treatment group ( Table 66.1 ). However, recruitment was slow, and the trial enrolled only 43 patients after 8 years instead of the intended 150. Furthermore, this trial suffered from poor compliance and quality assurance because only 9% of patients completed the intended 2 years of therapy, and 32% had violations of the intended radiation therapy. An additional 30 patients were entered into a treatment group, without randomization (see Table 66.1 ). Several subsequent studies showed similar advantages, ^, although they were nonrandomized studies and included small numbers of patients.

Later, Klinkenbijl and colleagues, on behalf of the European Organization for Research and Treatment of Cancer (EORTC), randomized 218 patients after curative resection for pancreatic cancer to chemoradiotherapy (40 Gy EBRT with concurrent 5-FU) or no adjuvant treatment. This phase III trial enrolled 114 patients with pancreatic cancer and 92 with periampullary cancer. Patients with pancreatic cancer experienced a nonstatistically significant trend towards survival benefit in the treatment group; with median survival of 1 year (95% confidence interval [CI], 0.8–1.4) and 1.3 years (95% CI, 1.1–1.8) in the observation and treatment groups respectively ( P = .099). A similar trend had also been noted in the overall study population (see Table 66.1 ). About 20% of patients randomized to the chemoradiotherapy arm did not receive the intended treatment, raising the question of whether the benefit would have been significant if the study was adequately powered.

These early studies set the stage for the European Study Group for Pancreatic Cancer (ESPAC) to conduct a randomized controlled trial (RCT) of adjuvant chemotherapy and chemoradiotherapy for patients who had undergone resection of pancreatic cancer. In the ESPAC-1 trial, about half of the patients were randomized to chemotherapy (bolus 5-FU for 6 months) versus no chemotherapy and chemoradiotherapy (20 Gy concurrent with 5-FU) versus no chemoradiotherapy. A novel two-by-two (2 × 2) factorial design was utilized for the other half of patients, whereby they were randomized to chemoradiotherapy, chemotherapy, both, and observation. A total of 541 patients were enrolled from 61 centers in 11 countries, making it the largest adjuvant therapy trial in pancreatic cancer at the time. A total of 289 patients were randomized into the 2 × 2 factorial design, whereas 256 patients were randomized to either chemotherapy or chemoradiation versus observation outside the original design (ESPAC-1 plus). Results from the overall study failed to demonstrate a survival benefit for chemoradiotherapy over observation but revealed a statistically significant survival benefit for chemotherapy. These findings were maintained when data from only the 2 × 2 factorial design study was evaluated.

ESPAC-1 has been criticized for the 2 × 2 trial design. Variation in the amount of radiotherapy received and lack of adjuvant chemotherapy in the radiation arm raised doubts about the observed lack of benefit to radiotherapy. Nonetheless, data from both patient sets are relevant and indicate the benefit of chemotherapy ( Table 66.2 ), evident in both resection margin–positive (R1) and resection margin–negative (R0) patients. The 2 × 2 factorial design was used to answer the two most important questions for intervention: (1) Is there a benefit for adjuvant chemoradiotherapy? and (2) Is there a benefit for adjuvant chemotherapy? Meta-analyses have confirmed this survival advantage of adjuvant chemotherapy, with a reduced risk of death in pancreatic cancer patients after resection of 25% (95% CI, 0.64–0.90; P = .001; see Chapter 62 ). This benefit is more pronounced in patients with an R0 margin compared with those with an R1 margin. In addition, the reported survival advantage for adjuvant chemotherapy was maintained when adjusted for quality of life during the 24 months after resection. Overall, these studies suggested that chemotherapy, as opposed to radiotherapy, in the adjuvant treatment of resected pancreatic cancer provides significant survival impact.

To evaluate the potential role of gemcitabine in conjunction with adjuvant chemoradiotherapy, the Radiation Therapy Oncology Group (RTOG) 9704 trial compared gemcitabine versus 5-FU administered before and after 5-FU–based chemoradiotherapy. Chemotherapy was given for 3 weeks before and 10 to 12 weeks after chemoradiotherapy (50.4 Gy EBRT concurrent with infusional 5-FU). Four hundred and fifty-one patients were stratified by tumor size, nodal involvement, and resection margin status. Median survival was not improved in the gemcitabine arm (20.5 months) compared with the 5-FU arm (16.9 months; hazard ratio [HR] = 0.84; 95% CI, 0.67–1.05; P = .12; Table 66.3 ). Analysis of the 5-year data of RTOG 9704 demonstrated no changes to the original inferences drawn.

ESPAC-1 reported a survival benefit for adjuvant chemotherapy, without exploring the question of whether chemoradiotherapy could improve on that benefit. The EORTC 40013/FFCD/GERCOR Phase II trial randomized 90 patients to receive chemotherapy alone with four cycles of gemcitabine, or chemoradiotherapy with two courses of gemcitabine followed by 50.4 Gy EBRT concurrent with weekly gemcitabine for 5 to 6 weeks. This trial included only R0 resection patients, who received gemcitabine-based chemoradiotherapy after systemic chemotherapy. Toxicity was the primary endpoint and was comparable in both arms. There were no differences between the two groups for the secondary endpoints of overall survival (24 months in both arms; see Table 66.3 ) and disease-free survival (12 months in chemoradiotherapy arm and 11 months in chemotherapy-alone arm). Local relapse rates were lower in the chemoradiotherapy arm (11% vs. 24%), although there was no difference in distant metastases rates (42% vs. 40%). This study suggested that chemoradiotherapy added to chemotherapy can reduce the incidence of local recurrence even though it does not demonstrate an overall survival benefit. Use of a more intense chemotherapy regimen before chemoradiotherapy did not increase survival as observed in a Phase II RCT that compared 3 months of induction chemotherapy using cisplatin, epirubicin, 5-FU, and gemcitabine (PEFG) versus gemcitabine. Chemotherapy was followed by chemoradiotherapy (54–60 Gy RT, concurrent with infusional 5-FU).

A systematic review and network meta-analysis identified nine high-quality studies with 3033 total patients who had been randomized to one of six treatment arms including chemotherapy and/or chemoradiotherapy. The analysis demonstrated that adjuvant chemotherapy with either 5-FU or gemcitabine reduces mortality by at least one-third in patients with resected pancreatic adenocarcinoma. It was unclear whether chemoradiation plus chemotherapy is more effective than chemotherapy alone in prolonging survival; however, the addition of radiotherapy added toxicity. Nonetheless, in many parts of North America, the use of adjuvant chemoradiation continues, in contrast to Europe and Asia, where adjuvant chemotherapy is considered the standard of care. Guidelines in North America favor adjuvant chemotherapy, while making allowance for the addition of chemoradiotherapy, particularly for settings with high risk of recurrence (such as positive resection margins). Interferon-based adjuvant protocols have evaluated and demonstrated no improvement in survival and with significant toxicity.

Adjuvant chemotherapy

The ESPAC-1 trial provided the first evidence that survival benefit can be derived from adjuvant chemotherapy (see Table 66.2 ), when based on regimens of biomodulated 5-FU. The benefit of chemotherapy was evident in both R1 and R0 resections. As noted previously, two meta-analyses confirmed a reduced mortality risk of 25% (95% CI, 0.64–0.90; P = .001) and an advantage of adjuvant chemotherapy in R0 margin versus R1 margin patients. The reported survival advantage for adjuvant chemotherapy was maintained when adjusted for quality of life during the 24-month period after resection. The advance made by ESPAC-1 has driven the next generation of adjuvant chemotherapy trials, notably ESPAC-3 (version 2) and ESPAC-4, as well as translational runoffs.

Emerging evidence in the 1990s demonstrated the value of gemcitabine for treatment of pancreas cancer, initially in the setting of advanced disease. ^, The CONKO-001 multicenter Phase III RCT demonstrated improved median disease-free survival (DFS) with adjuvant gemcitabine (13.4 months; 95% CI, 11.4–15.3) compared with observation (6.9 months; 95% CI, 6.1–7.8; P < .001; Table 66.4 ). The estimated DFS at 3 years and 5 years was 23.5% and 16.5%, respectively, in the gemcitabine group versus 7.5% and 5.5% in the observation group. There was, however, no significant improvement in median overall survival (OS) with gemcitabine (22.1 months; 95% CI, 18.4–25.8), compared with observation alone (20.2 months; 95% CI, 17–23.4; P = .06). Estimated OS at 3 years and 5 years was 34% and 22.5%, respectively, for gemcitabine patients versus 20.5% and 11.5% for observation patients. Subsequent analyses of the 5- and 10-year data continued to demonstrate significantly improved median OS: 20.7% and 12.2%, respectively, in the gemcitabine arm versus 10.4% and 7.7% in the observation arm. Two smaller studies provided similar survival estimates for gemcitabine, with a survival benefit over observation that was not further improved by the addition of fluoropyrimidines (see Table 66.4 ). ^,

As previously illustrated, CONKO-001 showed that gemcitabine improved survival compared with observation in resected pancreatic cancer patients, and Yoshitomi et al. found that the addition of fluoropyrimidine to gemcitabine did not improve on survival. It remained unclear, however, whether gemcitabine was superior to 5-FU for adjuvant treatment of patients with resected pancreatic cancer. The ESPAC-3 trial set out to answer this question. This study recruited 1088 patients with pancreatic ductal adenocarcinoma and reported no difference in OS between 5-FU/leucovorin (LV; also known as folinic acid [FA]) and gemcitabine by treatment ( P = .39), treatment effect by resection status ( P = .56), or by progression-free survival ( P = .44). There were, however, key advantages to the use of gemcitabine, such as its significantly reduced side effect and toxicity profile (see Table 66.4 ). Evaluating ESPAC-3 to determine the optimal duration and time to initiate chemotherapy, Valle and colleagues concluded that completion of all six cycles of adjuvant chemotherapy was an independent prognostic factor after resection for pancreatic adenocarcinoma. No survival disadvantage was seen by delaying chemotherapy initiation for up to 12 weeks after resection. Interestingly, within the sub-group that completed less than 6 months of chemotherapy, those that started chemotherapy after 8 weeks fared better than those that started before 8 weeks after resection. These sub-analyses suggest a benefit to delaying chemotherapy for up to 12 weeks to allow for adequate postoperative recovery. Caution should be taken, however, in the interpretation of data obtained from subgroup analyses that were not pre-specified, as in this case.

The JASPAC-01 phase III multi-institutional trial randomized 385 patients in Japan to receive either gemcitabine or S-1, an orally active fluoropyrimidine. S-1 contains a prodrug of 5-FU (tegafur) along with two other agents designed to minimize gastrointestinal and systemic toxicity (gimeracil and oteracil), effectively increasing the tolerable tumor dose of 5-FU. In advanced pancreatic cancer, S-1 has been shown to be noninferior to gemcitabine, ^, so this study set out to evaluate noninferiority in the adjuvant setting. After the interim analysis for efficacy, the study was stopped early in accordance with the prespecified criteria for discontinuation. At the time of stoppage, all protocol treatments had been completed and the data analysis reported 5-year OS of 44.1% and 24.4% in the S-1 and gemcitabine groups, respectively. Median OS was observed to be 46.5 and 25.5 months, respectively, in the per protocol analysis (see Table 66.4 ). From these results, S-1 was determined to be noninferior to gemcitabine in Japanese patients. Because S-1 is not approved in the West, it is uncertain how generalizable these findings are outside of Japan.

Several studies have attempted to explore the potential benefit of multi-agent fluoropyrimidine-based chemotherapy in the adjuvant setting. Bakkevold and colleagues randomized patients who had undergone pancreatic resection (61 with radical pancreatic resection, including 14 patients with periampullary cancers) to 5-FU, doxorubicin (Adriamycin), and mitomycin C (FAM) or observation. Although there was a significant increase in median survival of 23 months versus 11 months ( P = .02) favoring the FAM-treated group, there was no improvement in 3-year or 5-year survival. Increased toxicity was also noted, and only 56% of the FAM treatment group was able to complete the six courses of chemotherapy ( Table 66.5 ). In another study, Takada, and colleagues enrolled 508 patients with mixed pancreaticobiliary cancers ( n = 158 pancreatic cancer) who were randomized to surgery alone or surgery with combined mitomycin C and 5-FU. The 5-year survival rates were not significantly different, at 18% versus 11.5%, respectively. Of note, this study used oral 5-FU, which has lower bioavailability compared with intravenous administration because of the phenomenon of first-pass hepatic metabolism. Kosuge and colleagues undertook a multicenter RCT of surgery versus surgery and combination cisplatin and 5-FU, with 60% of patients also receiving 30 Gy of intraoperative radiotherapy. They concluded that although cisplatin with 5-FU was safe and well tolerated, it conferred no survival benefit (see Table 66.5 ).

Gemcitabine combined with the orally active 5-FU prodrug capecitabine (GemCap) has produced a small improvement in survival compared with gemcitabine alone in the advanced disease setting. Subsequently, GemCap was evaluated in the adjuvant setting in the ESPAC-4 trial. This study enrolled 732 patients who had undergone curative resection for pancreatic ductal adenocarcinoma and randomized them to gemcitabine ( n = 366) or GemCap ( n = 364). The median OS was 28 months (95% CI, 23.5–31.5) in the GemCap group and 25.5 months (95% CI, 22.7–27.9) in the gemcitabine group, with an HR of 0.82 (95% CI, 0.68–0.98; P = .032). In an updated analysis, this benefit was maintained with a median OS of 30.2 months (95% CI, 25.8–33.5) and 27.9 months (95% CI, 24.8–29.9), respectively. This study established GemCap as an option for adjuvant therapy, and it is now typically used for older patients and/or for patients with an ECOG performance status greater than 1.

More recently, the value of adjuvant-modified FOLFIRINOX (mFOLFIRINOX; 5-FU, oxaliplatin, leucovorin, irinotecan) has been established as the standard adjuvant recommendation for fit patients who have undergone curative surgery for pancreatic cancer based on the results of the PRODIGE (Parternariat de Recherche en Oncologie Digestive)/ACCORD (Action Concertées dans les Cancers Colorectaux et Digestifs) 24/CCTG (Canadian Clinical Trials Group) PA.6 trial. In total, 493 patients were randomized to either adjuvant mFOLFIRINOX or adjuvant gemcitabine for 6 months. Specific eligibility included World Health Organization (WHO) performance status 0 to 1, CA19-9 less than 180 U/mL, and either an R0 or R1 resection. At a median follow-up of 33.6 months for surviving patients, the median DFS was 21.6 months in the mFOLFIRINOX group and 12.8 months in the gemcitabine group (HR 0.58; 95% CI, 0.46–0.73; P < .001). The median OS was notable at 54.4 months in the mFOLFIRINOX and 35.0 months in the gemcitabine group, respectively (HR 0.64; 95% CI 0.48–0.86; P = .03). There was a higher rate of grade 3 to 4 adverse events in the mFOLFIRINOX arm of 75.9%, and 52.9% in the gemcitabine group. These data immediately changed practice and support the recommendation for adjuvant mFOLFIRINOX for good performance status patients.

The APACT (Adjuvant PAnCreas Trial) results have been preliminarily reported and provide insight into the role of gemcitabine and nab-paclitaxel in the adjuvant setting. This large adjuvant study randomized 866 patients to either gemcitabine and nab-paclitaxel or gemcitabine alone for six cycles. Eligibility criteria were like the adjuvant PRODIGE trial. The primary endpoint was DFS adjudicated by blinded independent central review. At a median follow-up of 38.5 months, the DFS for gemcitabine and nab-paclitaxel was 19.5 months versus 18.8 months for the gemcitabine group (HR 0.88; 95% CI 0.729–1.063; P = .18). The interim median OS was 40.5 months for the doublet and 36.2 months for the gemcitabine group (HR 0.82; 95% CI 0.68–0.99; P = .045). Mature OS data are awaited because the interim analysis did not show significant benefit with the doublet regimen. Currently, the use of gemcitabine and nab-paclitaxel in the adjuvant setting remains investigational.

Overall key factors that are moving the benchmark for adjuvant therapy for pancreatic cancer include improved patient selection (such as limiting by CA19-9 level), integration of optimal supportive therapy measures, and improvements in therapy.