Radiation Oncology/Pancreas/Resectable

Front Page: Radiation Oncology | RTOG Trials | Randomized Trials

Resectable Pancreatic Cancer

Treatment Overview

• Historically, surgery alone was used to treat resectable pancreatic cancer. Unfortunately, 5-year OS at the time of the first GITSG trial was only 6%
• GITSG 9173 randomized patients to adjuvant chemo-RT with concurrent/adjuvant 5-FU vs. no adjuvant treatment. The study closed prematurely due to slow accural, but showed a dramatic benefit for DFS and OS. RT was given as split course (20 Gy + 20 Gy). Adjuvant chemo-RT became the standard of care in the US
• In a follow-up trial (RTOG 97-04), the role of induction and adjuvant gemcitabine compared with 5-FU as a sandwich to concurrent chemo-RT was evaluated. There was a trend for improved survival with gemcitabine (particularly for head of pancreas tumors) but also significantly more toxicity. This trend dissipated at 5 year follow up (p=0.12)
• In the US, NCCN.org considers both chemo-RT and chemotherapy to be valid choices (evidence 2A)
• In Europe, EORTC 40891 attempted to confirm the GITSG results. It randomized patients to adjuvant chemo-RT vs. no adjuvant treatment. RT schedule was the same; chemo schedule was concurrent C.I. 5-FU only, without 2 years of further adjuvant treatment. It also included patients with peri-ampullary cancers (45%), who have significantly better prognosis. There was no benefit for LRR or OS. On exploratory subset analysis, there was also no benefit for the pancreatic cancer subset. In this trial, ~80% of patients died of their disease, but 8% survived >10 years
• ESPAC-1 included 3 separate randomization (2x2, +/- chemo, +/- chemo-RT) trying to further determine the role of the various adjuvant options. On combined analysis, there was no survival benefit for adjuvant chemo-RT, but significant benefit for adjuvant chemotherapy. A further analysis of only the 2x2 patients again showed a survival benefit for adjuvant chemotherapy, and a deleterious effect of chemo-RT. However, this trial was widely criticized for its methodology
• A further meta-analysis of individual patient data from available randomized trials (Europe, Japan) confirmed a survival benefit for adjuvant chemotherapy and suggested trend to worse survival with chemo-RT. The one exception was patients with R1 resection, where chemo-RT appeared beneficial over chemotherapy alone
• Based on the combined results of EORTC 40891 and ESPAC-1, adjuvant chemo-RT is not indicated in Europe. Adjuvant chemotherapy is considered standard of care
• A German trial (CONKO-001) demonstrated a significant survival advantage for adjuvant gemcitabine over observation alone; there was no benefit in a Japanese trial of adjuvant cisplatin/5-FU
• Ongoing ESPAC-3 trial attempts to compare observation only to adjuvant 5-FU (ESPAC-1) to adjuvant gemcitabine (CONKO-001)
• to be continued...

Surgery

• Only ~10-20% of patients surgical candidates
• Based on Johns Hopkins experience (2007), ~20% will have N0 disease and ~80% will have N1 disease. Lymph node ratio (positive/examined) most powerful predictor of survival
• Perioperative mortality with modern techniques is low (<4%)
• Upon laparoscopy small liver mets and peritoneal implants are found in ~10% of head tumors and up to 40% of body/tail tumors, leading to aborted surgery

• Whipple procedure - invented in 1935 by Dr. Allen Whipple, an American surgeon.
  • Remove: head of pancreas, duodenum, proximal jejunum, gall bladder, distal common bile duct, regional lymph nodes
  • Reconstruction: pancreaticojejunostomy, choledochojejunostomy, gastrojejunostomy

• NCCN criteria for resectability:

Resectable - no distant mets, clear fat plane around celiac/SMA, patent SMV

Borderline Resectable - severe unilateral SMV/portal impingement, abut SMA, gastroduodenal artery encasement up to hepatic artery, limited involvement of IVC, colonic invasion, SMV occlusion

Unresectable - distant mets (including celiac/para-aortic), SMA/celiac encasement, SMV/portal occlusion, aortic/IVC invasion, SMV invasion below transverse mesocolon

Preoperative Biliary Drainage

• The Netherlands (2003-2008) -- preop biliary drainage vs. surgery only
  • 2010 PMID 20071702 -- "Preoperative biliary drainage for cancer of the head of the pancreas." (van der Gaag NA, N Engl J Med. 2010 Jan 14;362(2):129-37.)
  • Randomized. 202 patients, obstructive jaundice and bilirubin 2.3 - 14.6 mg/dl. Arm 1) Preoperative biliary drainage for 4-6 weeks vs. Arm 2) No drainage. Both arms followed by surgical resection. Drainage primarily by ERCP placed endoprosthesis. Primary outcome rate of serious complications
    • Outcome: Complications drain 74% vs. no drain 39% (RR 0.54, SS). Mortality and LOS no difference
    • Conclusion: Routine pre-op biliary drainage increases rate of complications

Surgical Results

• Kuala Lumpur, 1997 (1980-1993) PMID 9372373 -- "Carcinoma of the pancreas: resection outcome at the University Hospital Kuala Lumpur." (Shahrudin MD, Int Surg. 1997 Jul-Sep;82(3):269-74.)
  • Retrospective. 236 patients with curative intent. 75% PD, 15% total pancreatectomy, 10% distal pancreatectomy, 3% distal subtotal. Extra-pancreatic invasion in 75%, PNI 83%
  • Median OS: 14 months (vs. 5 months if no resection). 5-year OS: 10%. Op mortality 3.4%.
  • Recurrences: 10 patients recurred/died >5 years.

• Hopkins, 1995 (1970-1994) PMID 7794076 -- "Pancreaticoduodenectomy for cancer of the head of the pancreas. 201 patients." (Yeo CJ, Ann Surg. 1995 Jun;221(6):721-31; discussion 731-3.)
  • Retrospective. 201 patients with head of pancreas.
  • In-hospital mortality: 5%, but 0.7% for last 149 patients
  • 5-year OS: 21%. SM- 26% vs. SM+ 8% (SS). LN- 36% vs. LN+ 14% (SS). OS in 1990s 36%
  • Starting in 1991, 56 received RT (>45 Gy) + 5-FU, 22 declined: 2-year OS RT+ 35% vs. RT- 0%
  • Predictors of survival: diameter <3cm, LN-, SM-, diploid DNS

• MGH, 1993 (1978-1991) PMID 8094952 -- "Resection margins in carcinoma of the head of the pancreas. Implications for radiation therapy." (Willett CG, Ann Surg. 1993 Feb;217(2):144-8.)
  • Retrospective. 72 patients with head of pancreas. 39 received RT (40/22 - 50/28 Gy) + 5-FU
  • Margins: 51% SM+ (73% peripancreatic soft tissue, 38% pancreatic transection line). LN+: 65%
  • 5-year: OS 13%, LC 34%, FFDM 29%.
  • 2-year impact of margin (SM+ vs. SM-): OS 18% vs. 37%; LC 44% vs. 60%; FFDM 38% vs. 52%
  • No benefit to RT/5-FU

Patterns of failure

• Local failure 50-80%, sole site of failure ~25%
• Liver failure ~50%, commonly together with local failure, rarely as sole site
• Peritoneal failure ~30%
• Distant failure rare without local failure
• Note: most series are older without benefit of modern imaging and CA19-9 monitoring

• Tochigi, 2006 (Japan) PMID 16627216 -- "Patterns of recurrence after curative resection of pancreatic cancer, based on autopsy findings." (Hishinuma S, J Gastrointest Surg. 2006 Apr;10(4):511-8.)
  • Retrospective. Autopsy findings. 27 patients after curative resection. Recurrence classified as 1) local, 2) hepatic, 3) peritoneal, 4) para-aortic LN, 5) other distant mets
  • Recurrence: 75% local (15% undetectable on CT), 50% hepatic mets, 46% local+hepatic, peritoneal 33%, para-aortic LN 21%, other distant mets 75%
  • Local disease direct cause of death only in 17%

• Padua, 1997 (Italy) PMID 8995078 -- "Recurrence after resection for ductal adenocarcinoma of the pancreas." (Sperti C, World J Surg. 1997 Feb;21(2):195-200.)
  • Retrospective. 78 patients.
  • Recurrence: local component 72%, hepatic component 62%, both 97%. Majority within 24 months
  • Median DFS: 8 months; predictors: grade, size, SM, LN
  • Conclusion: surgery alone inadequate

• Kanazawa, 1993 (Japan) PMID 8104092 -- "An evaluation of radical resection for pancreatic cancer based on the mode of recurrence as determined by autopsy and diagnostic imaging." (Kayahara M, Cancer. 1993 Oct 1;72(7):2118-23.)
  • Retrospective. 45 patients. Recurrence classified as 1) retroperitoneal (local and/or LN), 2) hepatic, 3) peritoneal, 4) distant
  • 15 postmortem: local 80%, retroperitoneal LN 47%, hepatic 66%, peritoneal 53%
  • 15 antemortem: retroperitoneal 87%, hepatic 53% (almost all had also local recurrence)

• Lund, 1993 PMID 8406311 -- "Recurrence of exocrine pancreatic cancer--local or hepatic?" (Westerdahl J, Hepatogastroenterology. 1993 Aug;40(4):384-7.)
  • Retrospective. 74 patients with recurrent disease
  • Recurrence: 8% local only, 78% local + liver, 14% liver only

• U Penn, 1991 (1981-1989) PMID 1938511 -- "Adjuvant therapy of resected adenocarcinoma of the pancreas." (Whittington R, Int J Radiat Oncol Biol Phys. 1991 Oct;21(5):1137-43.)
  • Retrospective. 72 patients; initially 33 surgery only, then 19 surgery + RT, then 20 surgery + chemo-RT (CI 5-FU)
  • Local recurrence: 85% vs. 55% vs. 25%. Peritoneal 23%. Liver 23%.
  • 2-year OS: 35% vs. 30% vs. 43%; very poor survival if SM+

• Kansas, 1990 (1977-1987) PMID 2354408 -- "Patterns of failure after curative resection of pancreatic carcinoma." (Griffin JF, Cancer. 1990 Jul 1;66(1):56-61.)
  • Retrospective. 36 patients with curative resection of pancreatic CA, 26 Whipple, 7 total pancreatectomy, 3 distal pancreatectomy
  • Outcome: 26 recurred, 4 died of initial disease, 6 cured. Median OS 11 months, 5-year OS 17%
  • Initial failures: 27% local only (50% local component), 15% liver only (50% liver component), peritoneal 30%, distant 23% (never sole site)
  • Ultimate failures: 19% local only (73% local component), 12% liver only (62% liver component), peritoenal 42%, distant 27% (never sole site)
  • High local failure risk: age >60, T2-T3, location in body/tail
  • Adjuvant treatment: 10/36 received RT 45-60 Gy and 5-FU; had unfavorable characteristics (SM+, LN+, T2-3)

• MGH, 1976 (1963-1973) PMID 1260670 -- "Carcinoma of the pancreas: review of MGH experience from 1963 to 1973. Analysis of surgical failure and implications for radiation therapy." (Tepper J, Cancer. 1976 Mar;37(3):1519-24.)
  • Retrospective. 145 patients considered for definitive surgery, 21% with radical surgery, 43% palliative, 36% biopsy alone. 9 patients received RT (30-45 Gy); 5-FU given occasionally
  • Median OS: 4 months, 2-year OS: 7%. For radical surgery median OS: 10.5 months
  • Radical surgery failures: LR most common site of failure (50%), mets without LR 13%

Extent of surgical resection

• Extent of surgery
  • Initial Whipple procedure included distal gastrectomy
  • Randomized trial from Johns Hopkins of peri-ampullary tumors concluded that pylorus-sparing pancreaticoduodenectomy has comparable outcomes to pancreaticoduodenectomy with gastrectomy and extended lymph node dissection
• Extent of lymph node dissection
  • Extended lymph node dissection initially described in 1973, since standard LND doesn't remove LNs frequently involved with microscopic disease and many patients die with local recurrence
  • It became widely adopted in Japan, but because of complexity and no apparent improvement in survival, was not adopted by Western surgeons
  • Japanese retrospective series showed superior survival rates to Western series, and prospective randomized trials were undertaken
  • Three randomized trials were published, and none showed a survival benefit for extended LND. Meta-analysis published in 2007 also didn't find a survival benefit, and found a trend to worse morbidity
  • Lymph node ratio (positive/examined) appears to be the most potent prognostic factor for survival after surgery

• Meta-Analysis; 2007 PMID 17318801 -- "Systematic review and meta-analysis of standard and extended lymphadenectomy in pancreaticoduodenectomy for pancreatic cancer." (Michalski CW, Br J Surg. 2007 Mar;94(3):265-73.)
  • Meta-analysis. 4 randomized trials identified, 3 analyzed. Standard vs extended procedures
  • Outcome: Weighted mean HR for OS 0.93 (NS)
  • Toxicity: Comparable, trend to higher delayed gastric emptying with extended
  • Conclusion: Extended lymphadenectomy doesn't benefit OS, and may trend to increased morbidity

• Johns Hopkins; 2007 (1995-2005) PMID 17462460 -- "Prognostic relevance of lymph node ratio following pancreaticoduodenectomy for pancreatic cancer." (Pawlik TM, Surgery. 2007 May;141(5):610-8. Epub 2007 Mar 23.)
  • Retrospective. 905 patients, pancreaticoduodenectomy (classic 29%, pylorus-preserving 71%). R0 resection in 59%. Median LN removed 17. Median F/U 2 years
  • Outcome: 21% N0 vs. 79% N1 disease. Median OS 17 months, N0 25 months vs N1 16 months (SS). Lymph node ration most compelling independent predictor of survival: LNR=0 25 months, LNR <0.2 22 months, LNR 0.2-0.4 15 months, LNR >0.4 12 months (SS)
  • Extent of dissection: N0 patients and <12 LNs 22 months vs. >12 LNs 33 months (p=0.09, but fairly small sample size). N0 and <12 LN survival comparable to N1
  • Conclusion: Lymph node ratio powerful predictor of survival; patients with <12 lymph nodes may be understaged

• Mayo Clinic (1997-2003) -- standard vs. extended LND
  • Randomized. 79 patients with resectable pancreatic head cancer, peri-ampullary and islet cell excluded. Distal gastrectomy, pylorus preservation not allowed. Arm 1) standard LND (D1) removed first-echelon lymph nodes en bloc vs. Arm 2) extended LND (D2) which in addition dissected second-echelon lymph nodes
  • 2005 PMID 16269290 -- "A prospective randomized trial comparing standard pancreatoduodenectomy with pancreatoduodenectomy with extended lymphadenectomy in resectable pancreatic head adenocarcinoma." (Farnell MB, Surgery. 2005 Oct;138(4):618-28; discussion 628-30.)
    • Outcome: Median OR time D1 6.2 hrs vs. D2 7.6 hrs (SS), blood transfusions 22% vs. 44% (SS), median LN resected 15 vs. 36 (SS), LOS ~11 days (NS). Median OS 2.2 years vs. 1.6 years (NS), 1-year OS 82% vs. 71% (NS)
    • Toxicity: Diarrhea, body appearance, bowel control better (SS) with D1 resection
    • Conclusion: No difference, extended LND unattractive for further evaluation

• Johns Hopkins (1996-2001) - Pancreaticoduodenectomy: pylorus preservation vs. distal gastrectomy with retroperitoneal LN dissection
  • Randomized. 299 patients with periampullary carcinoma (57% pancreatic, 22% ampullary, 17% distal bile duct, 3% duodenal). Initially treated with pylorus-preserving pancreaticoduodenectomy then Arm 1) pylorus preservation and standard LND (distal gastrectomy only if inadequate margin) vs. Arm 2) radical resection with gastrectomy and extended LND
  • 2002 PMID 12192322 -- "Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma, part 2: randomized controlled trial evaluating survival, morbidity, and mortality." (Yeo CJ, Ann Surg. 2002 Sep;236(3):355-66; discussion 366-8.)
    • Complication rate: standard 29% vs. radical 43%
    • LN+ 74%, SM+ 10% (standard 21% vs. radical 5%)
  • 5-years; 2005 PMID 16332474 -- "Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma--part 3: update on 5-year survival." (Riall TS, J Gastrointest Surg. 2005 Dec;9(9):1191-204; discussion 1204-6.). Median live F/U 5.3 years
    • Outcome: 5-year OS standard 25% vs. radical 31% (NS); pancreatic CA - standard 13% vs. radical 29% (NS). SM+ standard 21% vs. radical 5% (SS)
    • Conclusion: Pylorus-preserving pancreaticoduodectomy without retroperitoneal lymphadenectomy procedure of choice, but higher SM+ in standard resection

• Milan (Italy)(1991-1994) -- standard LND vs extended LND
  • Randomized. 81 patients with resectable head of pancreas CA. Pancreatoduodenectomy, then Arm 1) standard LND vs. Arm 2) extended LND and retroperitoneal soft-tissue clearance (hepatic hilum, along aorta from the diaphragm to the IMA, laterally to both renal hila, with clearance of the celiac axis and SMA). IORT in 19 patients (NS)
  • 1998 PMID 9790340 -- "Standard versus extended lymphadenectomy associated with pancreatoduodenectomy in the surgical treatment of adenocarcinoma of the head of the pancreas: a multicenter, prospective, randomized study. Lymphadenectomy Study Group." Pedrazzoli S, Ann Surg. 1998 Oct;228(4):508-17.)
    • Outcome: No difference in peri-operative outcomes. LN sampled standard 13 vs. extended 20 (SS). Median OS standard 11.0 months vs. 16.4 (NS)
    • Post-hoc analysis: N1 patients significantly better survival if extended vs standard; no difference in N0 patients
    • Predictors of survival: Grade, size (>2.0 cm), N1, >=4 transfused units of blood
    • Conclusion: Extended resection doesn't improve survival, although trend in N1 patients

Extensive Peritoneal Lavage

• Kumamoto, 2005 PMID 16211228 -- "EIPL (extensive intraoperative peritoneal lavage) therapy significantly reduces peritoneal recurrence after pancreatectomy in patients with pancreatic cancer." (Yamamoto K, Int J Oncol. 2005 Nov;27(5):1321-8.)
  • Retrospective. 39 consecutive patients, treated without EIPL (1995-2000) or with EIPL (2001-2003)
  • Recurrence (EIPL vs. non-EIPL): local 13% vs. 29%, LN 27% vs. 29%, hepatic 40% vs. 50%, peritoneal 7% vs. 46%, distant 13% vs. 21%

Neoadjuvant Chemo-RT

Proton Therapy

• Harvard; 2010 (2006-2008) PMID 20421151 -- "Phase I Study of Preoperative Short-Course Chemoradiation with Proton Beam Therapy and Capecitabine for Resectable Pancreatic Ductal Adenocarcinoma of the Head." (Hong TS, Int J Radiat Oncol Biol Phys. 2010 Apr 24. [Epub ahead of print])
  • Phase I. 15 patients, localized, resectable, head of pancreas adenoCA. Proton therapy dose-escalation with capecitabine 825 mg/m2 BID x10 days. Level 1 30/10, Level II 25/5 in 2 weeks, Level III 25/5 in 1.5 weeks, Level IV 25/5 in 5 days
  • Outcome: Surgical resection in 11/15, others developed mets (n=4) or were unresectable (n=1)
  • Toxicity: No dose-limiting toxicity. Grade 3 in 4 patients (pain 1, stent obstruction/infection 3)
  • Conclusion: Preop chemo-RT with 1 week proton beam is feasible

Adjuvant Therapy Meta-Analysis

• UK Clinical Trials Unit
  • Surgical margin; 2008 PMID 18209156 -- "Influence of resection margins and treatment on survival in patients with pancreatic cancer: meta-analysis of randomized controlled trials." (Butturini G, Arch Surg. 2008 Jan;143(1):75-83; discussion 83.)
    • Individual data from 4 randomized trials (GITSG, Norway, EORTC, Japan, ESPAC-1), 875 patients (R0 68%, R1 32%). Pooled hazards ratios calculated. Median F/U 3.7 years
    • Outcome: Surgical margin involvement not predictive for survival (HR 1.10, NS), median OS R0 16 months vs. R1 14 months (NS), 5-year OS 16% vs. 15%
    • Chemo-RT: If R0 pooled HR 19% increased risk of death with chemo-RT (NS), median OS chemo-RT 16 months vs. no chemo-RT 16 months (NS); if R1 pooled HR 28% reduction in risk of death (NS), median OS chemo-RT 15 months vs. no chemo-RT 11
    • Chemotherapy: If R0 pooled HR 35% reduction in risk of death with chemo (SS), median OS chemo 21 months vs. no chemo 14 months (SS); if R1 pooled HR 4% increased risk of death with chemo, median OS chemo 15 months vs. 13 months
    • Predictors of outcome: tumor size, grade, N-stage, not surgical margin
    • Conclusion: Adjuvant chemotherapy alone should be standard of care for either R0 or R1 resections. There may be a role for chemo-RT in R1 resection, but meta-analysis underpowered
    • Comment (same issue): widely disparate patients (IA-IIB), some subgroups don't make biological sense (eg R0 no chemo worse outcome than R1 no chemo)
  • 2005 PMID 15812554 -- "Meta-analysis of randomised adjuvant therapy trials for pancreatic cancer." (Stocken DD, Br J Cancer. 2005 Apr 25;92(8):1372-81.)
    • 5 randomized trials (GITSG, Norway, EORTC, Japan, ESPAC-1), 4 recent trials (excluding GITSG) provided individual patient data (n=875). Previously unpublished F/U data on 261 ESPAC-1 patients. Overall median F/U 3.7 years
    • Chemo-RT: pooled HR 9% increase in risk of death with chemo-RT (NS), median OS chemo-RT 16 months vs. no chemo-RT 15 months; 2-year OS 30% vs. 34%, 5-year OS 12% vs. 17%
    • Chemotherapy: pooled HR 25% reduction in risk of death with chemotherapy (SS), median OS chemo 19 months vs. no chemo 13 months; 2-year OS 38% vs. 28%, 5-year OS 19% vs. 12%
    • Subgroup analysis: Chemo-RT may be more effective in patients with SM+
    • Conclusion: Chemotherapy is effective adjuvant treatment, chemoradiation is not. There may be benefit for chemo-RT in SM+

Surgery +/- Chemo-RT

• Erasmus MC (2000-2007) -- intra-arterial chemo-RT vs observation
  • Randomized. 120 patients, pancreatic head (52%) or periampullary (48%) cancer. After pancreaticoduodenectomy, randomized to Arm 1) intra-arterial mitoxantrone, 5-FU, leucovoring, and cisplatin x6 cycles with RT 54/30 after first cycle vs Arm 2) observation
  • 2008 PMID 19092348 -- "Adjuvant intra-arterial chemotherapy and radiotherapy versus surgery alone in resectable pancreatic and periampullary cancer: a prospective randomized controlled trial." (Morak MJ, Ann Surg. 2008 Dec;248(6):1031-41.) Median F/U 1.5 years
    • Outcome: Median OS 19 months vs 18 months (NS), PFS 37% vs 20% (SS). Significantly fewer liver mets on subgroup analysis
    • Conclusion: Intra-arterial chemo-RT prolongs time to progression, but no impact on survival
  • QoL; 2010 PMID 20029974 -- "Quality of life after adjuvant intra-arterial chemotherapy and radiotherapy versus surgery alone in resectable pancreatic and periampullary cancer: a prospective randomized controlled study." (Morak MJ, Cancer. 2010 Feb 15;116(4):830-6.)
    • Subset analysis, 103 patients, 355 responses. QoL C30 questionnaires
    • Outcome: No impact on physical, emotional, or social functioning. Less pain, nausea and vomiting with chemo-RT
    • Conclusion: Intra-arterial chemo-RT improved QoL compared with observation

• ESPAC-1 (1994-2000) -- 2x2 surgery +/- chemotherapy and +/- chemo-RT
  • Randomized, 2x2 design + 2 other randomizations to improve accrual. 541 patients total, treated with pancreaticoduodenectomy (R0 81%, R1 19%). Physicians could choose a randomization. 285 patients randomized to 2x2 factorial (observation, chemotherapy x6 cycles, chemo-RT, chemo-RT followed by adjuvant chemo x6 cycles). 68 patients randomized to +/- adjuvant chemo-RT. 188 patients randomized to +/- adjuvant chemo. Patients could be given other "background" therapy, which could be chemo or RT. RT AP/PA split course 20/10 + 20/10 (GITSG regimen, although up to 60 Gy could be given). Chemotherapy concurrent bolus 5-FU 500 mg/m2 + leucovorin, adjuvant 5-FU 425 mg/m2 x6 months
  • All patients; 2001 PMID 11716884 -- "Adjuvant chemoradiotherapy and chemotherapy in resectable pancreatic cancer: a randomised controlled trial." (Neoptolemos JP, Lancet. 2001 Nov 10;358(9293):1576-85.) Median F/U of alive patients 10 months
    • Analysis: Combined group chemo-RT (n=178) vs. no chemo-RT (n=175); combined group chemo (n=235) vs. no chemo (n=238).
    • Outcome: median OS chemo-RT 15 months vs. no chemo-RT 16 months (NS); median OS chemo 20 months vs. no chemo 14 months (SS)
    • Conclusion: No survival benefit for adjuvant chemo-RT, significant benefit for adjuvant chemo
    • Comment: Physicians were able to choose which randomization to put a patient on. Also, physicians were able to give additional "background" therapy which means that patientss randomized to no chemotherapy could still get chemotherapy and those randomized to chemotherapy only could still get RT. No central audit.
  • 2x2 subset; 2004 PMID 15028824 — "A randomized trial of chemoradiotherapy and chemotherapy after resection of pancreatic cancer" (Neoptolemos JP, N Engl J Med. 2004 Mar 18;350(12):1200-10.)
    • Analysis of 2x2 factorial patients only. 289 patients, not powered for individual group comparison. 82% deaths recorded. Median F/U 3.9 years
    • Outcome: median OS observation 17 months vs. chemo-RT 14 months vs. 22 months chemo vs. 20 months chemo-RT + chemo (not powered for statistics)
      • Chemo-RT comparison: median OS chemo-RT 16 months vs. no chemo-RT 18 months (NS), 5-year OS chemo-RT 10% vs. no chemo-RT 20% (p=0.05); no benefit regardless of chemo
      • Chemo comparison: median OS chemo 20 months vs. no chemo 15 months (SS), 5-year OS chemo 21% vs. no chemo 8% (p=0.009); benefit regardless of RT
    • Pattern of failure: LR 35%, LR+DM 27%, DM 34%
    • Conclusion: Adjuvant chemotherapy significant survival benefit; adjuvant chemo-RT had deleterious effect on survival
    • Editorial: PMID 15028829: Analysis of 2x2 not straightforward due to low power, worse survival of chemo-RT group probably due to toxicity
    • Commentary: PMID 15752874 — "A challenge to the therapeutic nihilism of ESPAC-1." (2005): Substandard RT used, inappropriate toxicity reporting, longer time-to-treatment in chemo-RT group, inclusion of R1 patients

• EORTC 40891 (1987-95) -- surgery +/- chemo-RT
  • Randomized. 218 patients. Pancreatic T1-2 N0-1a (55%) or periampullary (ampulla, distal CBD, duodenum) T1-3 N0-1a (45%). Arm 1) Chemo-RT vs. Arm 2) observation. RT split course 20/10 + 20/10 (GITSG regimen). Chemo 5-FU C.I. 25 mg/kg over 5 days at the beginning of each RT course. No maintenance (vs. 2 years in GITSG). SM+ 21%, LN+ 50%
  • 1999 PMID 10615932 — "Adjuvant radiotherapy and 5-fluorouracil after curative resection of cancer of the pancreas and periampullary region: phase III trial of the EORTC gastrointestinal tract cancer cooperative group." (Klinkenbijl JH, Ann Surg. 1999 Dec;230(6):776-82.)
    • For all tumors: Median PFS chemo-RT 17 months vs observation 16 months (NS); median OS 24 months vs 19 months (NS). 2-year OS 51% vs 41%, 5-year OS 28% vs 22% (NS)
    • Pancreas tumor subset: median OS 17 months vs 13 months (NS). 2-year OS 37% vs 23%, 5-year OS 20% vs 10% (p=0.09).
    • Periampullary tumor subset: median OS 39 months vs 40 months (NS). 5-year OS 38% vs. 36% (NS)
    • Site of failure: LR both arms 15% (NS), LRR+DM ~20%, DM ~30%
    • Conclusion: Routine use of post-op chemo-RT not warranted
  • 12-years; 2007 PMID 17968163 -- "Long-term survival and metastatic pattern of pancreatic and periampullary cancer after adjuvant chemoradiation or observation: long-term results of EORTC trial 40891." (Smeenk HG, Ann Surg. 2007 Nov;246(5):734-40.) Median F/U 11.7 years
    • Outcome: Death reported in 79% (malignant disease cause of death in chemo-RT 79% vs. observation 86%, but only 1 death >7 years); median OS chemo-RT 1.8 years vs. observation 1.6 years (NS), 5-year OS 25% vs. 22% (NS), 10-year OS 17% vs. 18% (NS) - pancreatic subset 8% vs. periampullary subset 29%; median PFS 1.5 years vs. 1.2 years (NS)
    • Site of failure: LR both arms 20%, LR+distant 30% (LR a component in 50% of failures), DM ~50% (liver 50%)
    • Conclusion: No benefit for adjuvant chemo-RT
  • Criticisms: No maintenance chemotherapy. About 20% of pts assigned to chemo-RT arm did not receive that treatment, although 93% received RT. Benefit was diluted by inclusion of favorable periampullary cancers, although on exploratory pancreatic cancer subset analysis there was no difference.

• GITSG 9173 (1974-82) -- surgery +/- chemo-RT
  • Randomized. Terminated early due to poor accrual. 42 patients with adenocarcinoma of the pancreas (ampulla excluded), who were resected, margins negative, no peritoneal mets. Head of pancreas 95%, contigous disease 38%, LN+ 28%, Grade 3 5%. Subtotal Whipple 68%, Stage I 35%. Randomized to Arm 1) RT + 5-FU vs Arm 2) observation. RT was supravoltage 40/20 split course (2 week break after 20/10). Fields included the pancreatic bed and regional nodes AP/PA. 5-FU (500 mg/m2) bolus on days 1-3 of each cycle. Adjuvant chemotherapy given weekly x 2 years beginning 1 month after RT.
  • Original study; 1985 PMID 4015380 — "Pancreatic cancer. Adjuvant combined radiation and chemotherapy following curative resection." (Kalser MH, Arch Surg. 1985 Aug;120(8):899-903.) Median F/U 5.5 years
    • Median OS: chemo-RT 20 months vs observation 11 months (SS). 2-year OS 42% vs 15%; 5-yr OS vs 15% and 5%; median DFS 11 months vs. 9 months (SS)
    • Recurrence: overall 71% vs. 86%; local 47%; hepatic 32% vs. 50%; DM 40% vs 52%.
    • Conclusion: Adjuvant chemo-RT may prolong survival time
  • Confirmatory study; 1987: PMID 3567862 — "Further evidence of effective adjuvant combined radiation and chemotherapy following curative resection of pancreatic cancer." Gastrointestinal Tumor Study Group. Cancer. 1987 Jun 15;59(12):2006-10.
    • Not randomized. 30 additional patients, same characteristics, received chemo-RT protocol.
    • Outcome: MS 18 months. 2-yr OS 46%, 5-yr OS 17%. Similar results to chemo/RT arm from randomized study.
    • Conclusion: Use of adjuvant chemo-RT is preferred to no adjuvant treatment

Non-randomized

• Hopkins, 2008 (1993-2005) PMID 18640931 -- "Analysis of fluorouracil-based adjuvant chemotherapy and radiation after pancreaticoduodenectomy for ductal adenocarcinoma of the pancreas: results of a large, prospectively collected database at the Johns Hopkins Hospital." (Herman JM, J Clin Oncol. 2008 Jul 20;26(21):3503-10.)
  • Retrospective review 616 patients with adenoCA of head/neck/uncinate. 345 (56%) did not receive chemoRT, while 271 (44%) recieved 5-FU based chemoRT (median dose 50 Gy), including 2-6 months maintenance 5-FU.
  • Outcomes: chemoRT had superior MS (21.2 mo vs 14.4 mo), 2-yr OS (43.9% vs 31.9%) and 5-yr OS (20.1% vs 15.4%) (all SS)
  • Poor predictors: tumor >3cm, SM+, and postop complications
  • Pts receiving chemoRT were younger, less comorbid disease, and less perioperative complications
  • Competing risk factor MVA performed, chemoRT still demonstrated significant beneficial effect
  • Conclusion: suggests that adjuvant concurrent FU-based CRT significantly improves survival after Whipple for resectable pancreatic CA.

• Mayo, 2008 (1975-2005) PMID 18640932 -- "Adjuvant radiotherapy and chemotherapy for pancreatic carcinoma: the Mayo Clinic experience (1975-2005)." (Corsini MM, J Clin Oncol. 2008 Jul 20;26(21):3511-6.)
  • Retrospective review of 472 pts who underwent R0 resection (+ margins, unresectable, indolent tumor types excluded)
  • Median f/u 32.4 mo. Median RT dose was 50.4 Gy in 28 fractions; 98% of RT patients also received concurrent 5-FU.
  • Median OS improved w/ adjuvant chemoRT (25.2 vs 19.2 months) (SS). Two-year OS was 50% versus 39%, and 5-year OS was 28% versus 17%.
  • ChemoRT arm had more LN-positive, high-grade tumors and more tumors with extension beyond the pancreas
  • MVA revealed lack of adjuvant therapy to be an independent negative prognostic factor
  • Conclusion: OS better in patients who received adjuvant chemoRT.

Surgery + Chemo-RT: Adjuvant Chemo Alternatives

• RTOG 97-04 / Intergroup (1998-2002) -- sandwich 5-FU vs. gemcitabine
  • Randomized. 451 patients, with complete GTR of pancreas (R0 or R1). Head of pancreas 86%. SM+ 34%, SM- 42% and unknown 25%. Arm 1) CI 5-FU (250 mg/m2) x3 weeks -> chemo-RT -> CI 5-FU (250 mg/m2) x12 weeks vs. Arm 2) gemcitabine (1000 mg/m2) x3 weeks -> chemo-RT -> gemcitabine (1000 mg/m2) x12 weeks. Chemo-RT was same for both arms (RT 50.4 Gy + concurrent 5-FU 250 mg/m2), 1-2 weeks after induction chemo. Compliance with both >85%
  • 3-years; 2008 PMID 18319412 -- "Fluorouracil vs gemcitabine chemotherapy before and after fluorouracil-based chemoradiation following resection of pancreatic adenocarcinoma: a randomized controlled trial." (Regine WF, JAMA. 2008 Mar 5;299(9):1019-26.) Median F/U 1.5 years for all, 4.7 years for surviving
    • Outcome: median OS 5-FU 17 months vs. gemcitabine 20 months (p=0.09); 3-year OS 22% vs. 31%
    • Toxicity: Grade 4 hematologic 5-FU 1% vs. gemcitabine 14% (SS)
    • Conclusion: Trend to survival for gemcitabine (but NS), but GEM significantly more toxic
  • CA 19-9; 2008 PMID 19029412 -- "Postresection CA 19-9 Predicts Overall Survival in Patients With Pancreatic Cancer Treated With Adjuvant Chemoradiation: A Prospective Validation by RTOG 9704." (Berger AC, J Clin Oncol. 2008 Nov 24. [Epub ahead of print])
    • Secondary outcome. 385 patients with assessable CA 19-9. Analyzed as dichotomized variable (<=90 vs >90 and <180 and >=180). Overall, >180 in 9% and >90 in 14%
    • Outcome: CA 19-9 <=90 significantly better (HR 3.4 SS); CA 19-9 <180 also (HR 3.5, SS). No patients with CA 19-9 >180 survived 3 years
    • Conclusion: Post-op CA 19-9 highly significant predictor of OS
  • 5-years; 2011 PMID 21499862 -- "Fluorouracil-based chemoradiation with either gemcitabine or fluorouracil chemotherapy after resection of pancreatic adenocarcinoma: 5-year analysis of the U.S. Intergroup/RTOG 9704 phase III trial." (Regine WF, Ann Surg Oncol. 2011 May;18(5):1319-26.)
    • Outcome: No difference in OS. Pancreatic head patients had a median survival of 20.5 months and 5-year OS of 22% with gemcitabine versus 17.1 months and 18% with 5-FU.
    • Conclusion: No difference in OS between Gemcitabine and 5-FU.
    • Editorial: Argues that post-op data difficult to interpret due to surgical and path issues resulting in lack of standardization of trial data.

Non-randomized

• Picozzi - single instution experience using XRT, 5FU, and alpha-interferon regimen following surgery
  • PMID12727570 — Interferon-based adjuvant chemoradiation therapy after pancreaticoduodenectomy for pancreatic adenocarcinoma., Picozzi et al. Am J Surg. 2003 May;185(5):476-80.
  • 42% were hospitalized during chemoradiation, virtually all due to gastrointestinal toxicity.
  • At mean follow-up time of 31.9 months, 67% of the patients are alive.
    • Actuarial overall survival for the 1-, 2-, and 5-year periods was 95%, 64%, and 55%.

Surgery +/- Adjuvant Chemo

• CONKO-001 (Germany)(1998-2004) - adjuvant gemcitabine vs observation
  • Randomized. 354/368 patients. Pancreas carcinoma T1-4 N0-1 M0 s/p gross complete resection (R0 or R1 allowed). Arm 1) adjuvant gemcitabine (1000 mg/m2) x 6 cycles vs Arm 2) no adjuvant treatment. R0 resection in 80%
  • 2007 PMID 17227978 — "Adjuvant Chemotherapy With Gemcitabine vs Observation in Patients Undergoing Curative-Intent Resection of Pancreatic Cancer." (Oettle H, JAMA. 2007 Jan 17;297:267-277.) Median F/U 4.4 years
    • Relapse: chemo 74% vs control 92%; LR (with/without DM) in 34% vs 41%. DM alone in 56% vs 49%.
    • Survival: Median DFS 13 vs 7 months (SS); 1-year DFS 58% vs. 31%, 3-year 23% vs. 7%, 5-year 16% vs. 5%. Median OS 22 months vs. 20 months (NS) but 5-year OS 22% vs. 11% (p=0.06)
    • Subgroup analysis shows benefit for gemzar for DFS for all subgroups. In subgroup analysis, OS benefit thus far for gemzar for R0 resection and T3-4 and N- pts.
    • Conclusion: Adjuvant Gemzar significantly improved DFS (primary endpoint), and support its use as adjuvant therapy
  • Final; 2008 ASCO Abstract "CONKO-001: Final results of the randomized, prospective, multicenter phase III trial of adjuvant chemotherapy with gemcitabine versus observation in patients with resected pancreatic cancer (PC)." (Neuhaus P, J Clin Oncol 26: 2008 (May 20 suppl; abstr LBA4504))
    • Outcome: 83% deaths. Median DFS GEM 13.4 months vs OBS 6.9 months (SS), benefit in all subgroups. Median OS 23 months vs. 20 months (SS), 3-year OS 37% vs. 20%, 5-year OS 21% vs. 9% (SS)
    • Conclusion: Adjuvant Gemcitabine x6 months significantly increases DFS and OS compared with observation

• JSAP (Japan)(1992-2000) -- adjuvant cisplatin/5-FU vs. observation
  • Randomized. 85 patients, s/p complete resection (R1 not allowed). Arm 1) adjuvant cisplatin 80 mg/m2 + 5-FU 500 mg/m2 x2 cycles. Intraop RT used but not discussed.
  • 2006 PMID 16490736 -- "A multicenter randomized controlled trial to evaluate the effect of adjuvant cisplatin and 5-fluorouracil therapy after curative resection in cases of pancreatic cancer." (Kosuge T, Jpn J Clin Oncol. 2006 Mar;36(3):159-65. Epub 2006 Feb 20.)
    • Outcome: median OS chemo 12 months vs. observation 16 months (NS); 5-year OS chemo 26% vs. observation 15%
    • Prognostic factors: Grade 1, no LN+. T-stage, size not prognostic
    • Toxicity: Grade 3+ 16%
    • Conclusion: Postop adjuvant cisplatin/5-FU safe, but no survival benefit

• Japan (1986-1992) -- adjuvant Mitomycin C/5-FU vs. observation
  • Randomized. 508 patients (pancreatic n=173, cholangio n=139, gallbladder n=140, ampulla n=56), s/p surgery. Stage II-IV, pancreas LN+ 80%, liver+ 10%, curative resection 58%. Arm 1) adjuvant chemo with mitomycin C + 5-FU C.I. x2 courses, followed by 5-FU until recurrence vs. Arm 2) observation
  • 2002 PMID 12365016 -- "Is postoperative adjuvant chemotherapy useful for gallbladder carcinoma? A phase III multicenter prospective randomized controlled trial in patients with resected pancreaticobiliary carcinoma." (Takada T, Cancer. 2002 Oct 15;95(8):1685-95.)
    • Outcome: 5-year OS pancreatic chemo 11% vs. observation 18%; for curative resection chemo 18% vs. observation 27% (NS)
    • Conclusion: No benefit for adjuvant systemic chemotherapy in pancreatic cancer

• Norway (1984-1987) -- adjuvant AMF vs. observation
  • Randomized. 61 patients with pancreatic (77%) or papilla (23%), s/p radical resection. Arm 1) adjuvant chemo (AMF) x6 cycles vs. Arm 2) control.
  • 1993 PMID 8471327 -- "Adjuvant combination chemotherapy (AMF) following radical resection of carcinoma of the pancreas and papilla of Vater--results of a controlled, prospective, randomised multicentre study." (Bakkevold KE, Eur J Cancer. 1993;29A(5):698-703.)
    • Outcome: median OS chemo 23 months vs. observation 11 months (SS); 2-year OS 43% vs. 32% (NS); 5-year OS 4% vs. 8% (NS)
    • Conclusion: Adjuvant chemo postpone incidence of recurrence during first 2 years; no difference on ultimate cure rate

Surgery + Chemo: Adjuvant Chemo Alternatives

• ESPAC-3 (2000-2007) -- adjuvant 5-FU vs. adjuvant gemcitabine
  • 1088 pts, R0 or R1 resection. No RT.
  • Trial originally included observation arm, which was later closed to accrual.
  • 2009 - ASCO Abstract 2009 -- "ESPAC-3(v2): A multicenter, international, open-label, randomized, controlled phase III trial of adjuvant 5-fluorouracil/folinic acid (5-FU/FA) versus gemcitabine (GEM) in patients with resected pancreatic ductal adenocarcinoma" (Neoptolemos J, Journal of Clinical Oncology, 2009 ASCO Annual Meeting Proceedings, Vol 27, No 18S (June 20 Supplement), 2009: LBA4505)
    • No OS diff. Median OS 23.0 mo (5-FU) vs 23.6 mo (GEM), NS. Improved toxicity profile with gem (but higher thrombocytopenia rate).

Ongoing:

• ESPAC-4 - adjuvant gemcitabine + capecitabine vs gemcitabine alone

Surgery vs. Chemo-RT

• Japan Pancreatic Cancer Study Group -- Surgery vs. Chemo-radiation
  • Randomized. Stopped early after survival benefit to surgery. 42/150 patients, locally advanced but resectable pancreatic cancer, no involvement of SMA or common hepatic artery, no para-aortic LN+. Arm 1) Surgery (pancreaticoduodenectomy or distal pancreatectomy + dissection of regional lymphatics) vs. Arm 2) Concurrent RT 50.4/28 and C.I. 5-FU 200 mg/m2 to primary tumor + 1-3 cm margin to cover regional lymph nodes
  • 5-years; 2008 PMID 18958561 -- "Surgery versus radiochemotherapy for resectable locally invasive pancreatic cancer: Final results of a randomized multi-institutional trial." (Doi R, Surg Today. 2008 Nov;38(11):1021-1028. Epub 2008 Oct 29.) Minimum F/U alive patients 5 years
    • Outcome: median OS surgery 12 months vs. chemo-RT 9 months (NS); 3-year OS 20% vs. 0% (SS); 5-year OS 10% vs. 0% (NS)
    • Conclusion: Locally invasive resectable pancreatic cancer is treated more effectively by resection than by chemo-radiation

Adjuvant Chemo vs. Adjuvant Chemo-RT

• Intergroup EORTC 40013-22012 / FFCD-9203 / GERCOR (2004-2007) -- gemcitabine vs gemcitabine followed by chemo-RT
  • Randomized Phase II. 90 pts with R0 resection randomized to: 1) Control arm - gemcitabine x 4 cycles, or 2) gemcitabine x 2 cycles followed by chemo-RT 50.4 Gy with weekly gemcitabine
  • 2010 PMID 20837948 -- "Adjuvant Gemcitabine Alone Versus Gemcitabine-Based Chemoradiotherapy After Curative Resection for Pancreatic Cancer: A Randomized EORTC-40013-22012/FFCD-9203/GERCOR Phase II Study" (Van Laethem JL, J Clin Oncol. 2010 Oct 10;28(29):4450-6.)
    • Primary objectives: treatment completed per protocol in 86.7% vs 73.3%. Grade 4 toxicity 0% vs 4.7%.
    • Grade 3 late toxicity in 3 pts in RT arm. Median DFS 11 months vs 12 months. Median OS 24 months in both arms. First local recurrence 24% vs 11%.
    • Conclusion: adjuvant gemcitabine-based chemo-RT is feasible and well-tolerated.

• ESPAC-1
  • See full details above at #Surgery +/- Chemo-RT

Official Guidelines

• 2009 - American Hepato-Pancreato-Biliary Association
  • PMID 19390900 PDF -- "Combined modality treatment of resectable and borderline resectable pancreas cancer: expert consensus statement." (Abrams RA, Ann Surg Oncol. 2009 Jul;16(7):1751-6.)

Radiation Technique

• RTOG Postoperative Pancreas Atlas

Other Resources

• ARRO Case - Borderline Resectable Pancreatic Cancer

• ARRO Contour Borderline Resectable Pancreas

• Multiple Choice Questions on Pancreas