Radiation Oncology: A Question-Based Review

Pancreatic and Periampullary Adenocarcinoma

John P. Christodouleas and Joseph M. Herman

image Background

Approximately how many pancreatic adenocarcinoma (PCA) pts are diagnosed per yr in the U.S.?

The incidence of PCA is ~30,000 cases/yr in the U.S.

Where does PCA rank in cancer incidence in the U.S.? Cancer mortality?

As of 2009, PCA is the 10th most common cancer Dx but the 4th most common cause of cancer death in the U.S.

Is there a racial or gender predilection for PCA?

YesBlacks are more commonly affected than whites; however, the incidence is similar among males and females.

In what decades of life does PCA incidence peak?

The peak age of PCA is in the 6th–7 th decades of life.

What are 3 environmental exposures associated with PCA?

Most common environmental risk factors for PCA:

1.     Tobacco smoking

2.     2-naphthylamine

3.     Benzidine

What % of PCA is familial?

~5% of PCA is familial.

What 2 genetic mutations have most frequently been associated with familial PCA?

p16 and BRCA2 are the 2 most common familial associated genetic changes found in PCA.

Is chronic pancreatitis associated with increased risk of PCA?

No. Chronic pancreatitis is not associated with risk of PCA. Historically, there appeared to be an association, but this can be explained by confounding factors.

What % of PCA arise in the head, body, and tail of the pancreas?

Common PCA sites are 75% in the head, 15% in the body, and 10% in the tail.

What % of PCA pts have metastatic Dz at Dx?

~50% of PCA pts have DM at Dx.

What % of PCA pts have regional node+ Dz at Dx?

~25% of PCA pts have regional node+ Dz at Dx.

For PCA, what are the 3 most common sites of DM?

Common sites of DM for PCA include the liver, peritoneal surface, and lungs.

Is there a role for screening in PCA?

No. There is no current role for PCA screening. There are studies evaluating the role of screening 1st-degree relatives of PCA with EUS, but this is still experimental.

What % of pancreatic tumors are from the exocrine pancreas?

~95% of PCA are from the exocrine pancreas.

What are the 4 most common pathologic subtypes of exocrine pancreatic tumors?

Most common subtypes of exocrine pancreatic tumors:

1.     Ductal adenocarcinoma (80%)

2.     Mucinous cystadenocarcinoma

3.     Acinar cell carcinoma

4.     Adenosquamous carcinoma

What is the most common oncogene in PCA?

The K-ras oncogene is present in ~85% of PCA.

What are the 2 most common presenting Sx of PCA?

Common presenting Sx of PCA are pancreatic/biliary duct obstruction, jaundice, and abdominal pain.

What Dz is commonly diagnosed 1–2 yrs prior to a PCA Dx?

60%–80% of PCA pts are diagnosed with diabetes 1–2 yrs prior to Dx. However, only a small proportion of diabetic pts develop PCA.

Periampullary cancers refer to tumors arising from what 3 structures?

Periampullary tumors are those arising from the ampulla of Vater, distal common bile duct (CBD), and adjacent duodenum.

image Workup/Staging

What is the DDx of a pancreatic mass?

The DDx of a pancreatic mass includes exocrine cancer, islet cell/neuroendocrine cancer, cystic adenomas, papillary cystic neoplasms (e.g., intraductal papillary mucinous tumor), lymphoma, acinar cell carcinoma, and metastatic cancer.

Name 4 appropriate procedures for obtaining tissue from a suspicious pancreatic mass.

Procedures to obtain tissue from a suspicious pancreatic mass:

1.     EUS-guided FNA

2.     CT-guided FNA

3.     Endoscopic retrograde cholangiopancreatography (ERCP)

4.     Pancreatic resection (i.e., histologic Dx is not required before surgery)

What is the major advantage of EUS-guided FNA over CT-guided FNA of a pancreatic mass?

EUS-guided FNA is associated with lower risk of peritoneal seeding (2% vs. 16%).

What is the workup for suspected PCA?

Suspected PCA workup: H&P, CBC, CMP, CA 19-9, triphasic thin-sliced CT abdomen (pancreas protocol), chest imaging, +/− ERCP/EUS FNA for Dx and/or stent placement

In what circumstance will a PCA pt not excrete any CA 19-9?

If a pt is red cell Lewis antigen A–B negative, then the pt cannot excrete CA 19-9. The Lewis antigen negative phenotype is present in 5%–10% of the population.

What is the significance of a post-resection CA 19-9 >90 U/mL?

In RTOG 9704, 53 pts (14%) had CA 19-9 >90 U/mL, and only 2 of these pts survived up to 3 yrs.

What is the NCCN 2010 classification scheme for PCA?

PCA are classified in 4 categories (and per AJCC staging):

1.     Resectable (T1-3N0 or N+) (stages I–II)

2.     Borderline resectable (T4NX) (stage III)

3.     Locally advanced (T4NX) (stage III)

4.     Metastatic (TXNXM1) (stage IV)

What is the AJCC 7th edition (2009) T and N staging for PCA?

1.     T1: limited to pancreas and ≤2 cm

2.     T2: limited to pancreas and >2 cm

3.     T3: extends beyond pancreas but without celiac axis or superior mesenteric artery (SMA) involvement

4.     T4: celiac axis or SMA involvement

5.     N1: regional node involvement

What are the AJCC 7th edition (2009) stage groupings for PCA?

1.     Stage 0: Tis

2.     Stage IA: T1N0M0

3.     Stage IB: T2N0M0

4.     Stage IIA: T3N0M0

5.     Stage IIB: T1-3N1M0

6.     Stage III: T4NXM0

7.     Stage IV: TXNXM1

Per the NCCN, what 3 criteria are necessary for a primary pancreatic tumor to be resectable?

NCCN resectability for PCA is defined as follows:

1.     Patent superior mesenteric vein (SMV)/portal vein confluence

2.     Clear fat plane around celiac artery and SMA

3.     No nodal mets or other mets beyond field of resection

Per the NCCN, what pancreatic head/body lesions are considered “borderline resectable”?

NCCN definition of borderline resectability for PCA:

1.     Severe unilat SMV/portal confluence impingement

2.     Tumor abutment on SMA

3.     Gastroduodenal artery encasement up to hepatic artery

4.     Tumors with limited involvement of IVC

5.     Short segment SMV occlusion with patent vein both proximally and distally

6.     Colon or mesocolon invasion

What pancreatic tail lesions are considered “borderline resectable”?

Invasion into the adrenal gland, colon, mesocolon, or kidney are considered borderline resectable for PCA tail lesions.

What location of PCA is associated with higher rates of resectability: head, body, or tail?

PCA head tumors are more resectable b/c they cause Sx early (and therefore present with earlier-stage Dz).

What % of pts with resectable PCA tumors by CT imaging will be resectable at the time of surgery?

~80% of PCA pts deemed resectable by CT are resectable at the time of surgery.

What is the stage of a PCA pt with positive cytology at time of laparoscopy?

Positive cytology is stage IV (M1).

Does the AJCC 7th edition (2009) TNM staging for periampullary adenocarcinoma differ from PCA?

Yes. There is a separate AJCC TNM staging for ampulla of Vater carcinoma, distal CBD carcinoma, and duodenal carcinoma.

image Treatment/Prognosis

What surgical procedure is required to resect a pancreatic head lesion?

Surgery utilized for pancreatic head resection includes pylorus-preserving pancreaticoduodenectomy (PPPD) or classic pancreaticoduodenectomy (Whipple procedure).

What anastomoses are performed in the classic pancreaticoduodenectomy (Whipple)?

There are 3 anastomoses performed for the Whipple procedure:

1.     Pancreaticojejunostomy

2.     Choledochojejunostomy (hepaticojejunostomy)

3.     Gastrojejunostomy

What are the 4 most favorable prognostic factors after resection?

Most favorable prognostic factors after resection of PCA:

1.     −Margins (R0)

2.     Low grade (G1)

3.     Small tumor size (<3 cm)

4.     N0 status

Is there a benefit to R1 or R2 resection over definitive CRT for PCA?

No. Retrospective evidence suggests that survival is similar between PCA pts who had R1 or R2 resection and definitive CRT. Therefore, planned resections should be done in pts where R0 resections are likely. Debulking surgery does not improve outcome over definitive CRT.

Should pts with resectable PCA undergo extended retroperitoneal lymphadenectomy?

No. Resectable PCA pts should not undergo an extended retroperitoneal lymphadenectomy. There is no survival benefit to extended lymphadenectomy by an RCT (5-yr 25% vs. 31%, NSS). (Riall TS et al., J Gastrointest Surg 2005)

Can definitive CRT replace surgical resection for resectable PCA?

No. Surgery alone is superior to CRT alone for pts with resectable PCA per the Japanese PCA Study Group in an RCT of surgery alone vs. definitive CRT (50.4 Gy with continuous infusion [CI] 5-FU). The trial was stopped early due to the benefit of surgery: MS was 12 mos vs. 9 mos, and 5-yr OS was 10% vs. 0%. (Doi R et al., Surg Today 2008)

What are adj Tx options for a PCA pt s/p resection?

Adj Tx options after a pancreaticoduodenectomy:

1.     Adj gemcitabine (CONKO-001)

2.     Adj gemcitabine alone → 5-FU/RT→ gemcitabine alone (RTOG 9704)

3.     Adj 5-FU/RT (GITSG 91-73); consider maintenance gemcitabine afterward

4.     Adj 5-FU → 5-FU/RT → 5-FU (RTOG 9704)

5.     Observation alone

What is the standard total dose and fractionation for PCA after surgical resection?

Standard adj RT volumes and doses for adj PCA are as follows: tumor bed and at-risk regional nodes plus a 1–2-cm margin for motion and set-up error to 45 Gy. CD to tumor bed and a margin to 50.4–54 Gy in 1.8 Gy/fx.

What American study 1st reported a benefit of adj CRT vs. no additional Tx for resected PCA? Describe the arms of this study and the major results.

The Gastrointestinal Tumor Study Group (GITSG 91–73) trial 1st reported benefit to adj CRT for PCA in 1985.

1.     Standard arm: postop observation

2.     Experimental arm: Adj CRT using split-course RT to 40 Gy (2-wk break after 20 Gy) with intermittent bolus 5-FU → 2 full yrs of adj 5-FU alone

3.     Improved MS (20 mos vs. 11 mos) and 2-yr OS (42% vs. 15%) in the adj CRT arm (Kalser MH et al., Arch Surg 1985)

Did the EORTC 40891 study on PCA support or contest the benefit of adj CRT?

Support. The EORTC 40891 trial included T1-T2, N0-N1 PCA or T1-T3, N0-N1 periampullary adenocarcinoma. The same randomization was used as that in GITSG 1985, except the Tx arm did not rcv maintenance adj 5-FU for 2 yrs. Median PFS was 17 mos (CRT) vs. 16 mos (observation), NSS; MS was 24 mos (CRT) vs. 19 mos (observation), NSS. For the subset of PCA pts, 5-yr OS was 20% (CRT) vs. 10% (observation), p = 0.09 (Klinkenbijl JH et al., Ann Surg 1999). The authors concluded that routine adj CRT was not warranted, though statistical reanalysis of this study found a significant survival benefit with adj therapy (Garofalo MC et al. Ann Surg 2006).

Did the ESPAC-1 study on PCA support or contest the benefit of adj CRT?

Contest. ESPAC-1 included pts with grossly resected adenocarcinoma of the pancreas. The study used a 2 × 2 factorial design; surgery +/− CRT and +/− adj chemo. Adj CRT was similar to GITSG. Adj chemo was 6 mos of 5-FU. MS was 15 mos (CRT) vs. 16 mos (no CRT), NSS; OS was 20 mos (chemo) vs. 14 mos (no chemo), p = 0.0005.

Criticisms: Physicians could have pts randomize into 2 × 2 or directly into 1 of the 2 randomizations. “Background Tx” was allowed (i.e., observation pts may have rcv chemo and/or RT). There was no central RT quality assurance. (Neoptolemos JP et al., Lancet 2001)

Note: Analysis of 2 × 2 subset suggests that CRT had a deleterious effect; 5-yr OS was 10% (CRT) vs. 20% (no CRT), p = 0.05.

How does the presence of a +margin after resection for PCA influence the decision for adj CRT?

UK Clinical Trials Unit meta-analysis of 5 RCTs, including individual data from 4 RCTs, found that the benefit of adj CRT was greater in R1 pts compared to R0 pts, though the difference was not SS. Also, the benefit of adj chemo alone decreased in R1 pts compared to R0 pts, suggesting that CRT may have more important role in R1 pts. (Butturini G et al., Arch Surg 2008)

What study supports the role for adj gemcitabine for resected PCA over best supportive care? What subset of pts were excluded from this trial?

CONKO-001 included T1-T4, N0-N1 pts in a RCT of observation vs. adj gemcitabine. Outcomes favored adj gemcitabine: MS was 23 mos vs. 20 mos (SS), and 5-yr OS was 21% vs. 9% (SS) (Neuhaus P et al., ASCO Abstract 2008). Note: Pts with CA 19-9 >90 were excluded from this trial.

What study compared adj 5-FU to gemcitabine following surgical resection?

ESPAC-3 showed a MS of 23 mos (95% confidence interval: 21.1, 25.0 mos) for pts treated with 5-FU and 23.6 mos (95% confidence interval: 21.4, 26.4 mos) for pts treated with gemcitabine. There was no significant difference between the groups. (ASCO abstract 2009)

Did the study RTOG 9704 on PCA support or contest a benefit of gemcitabine-based adj CRT?

This is controversialRTOG 9704 randomized R0 and R1 PCA pts to CI 5-FU (250 mg/m2/d) CRT (50.4 Gy) and pre- and post-CRT with either additional 5-FU or gemcitabine. Among all eligible pts, there were no differences. In a preplanned subset analysis of pts with pancreatic head tumors, trends favored gemcitabine: MS was 20.5 mos vs. 16.7 mos, and 3-yr OS was 31% vs. 22%, but results were NSS (p = 0.09). The 3-yr LR was significantly better for the gemcitabine arm (23% vs. 28%). (Regine W et al., JAMA 2008)

What is the Picozzi regimen for pts with PCA?

The Picozzi regimen is adj CRT with α-interferon, 5-FU, cisplatin, and RT (45–54 Gy). 42% of pts were admitted during CRT; however, outcomes were very promising: 5-yr OS was 55%, and MS was not reached (Picozzi VJ et al., Am J Surg 2003). Subsequent study results are pending.

What is the Tx paradigm for borderline resectable PCA?

Borderline resectable PCA Tx paradigm: Consider staging laparoscopy, stent placement if jaundice, and neoadj CRT → resection.

What neoadj CRT regimen should be used for borderline resectable PCA?

There is no standard neoadj Tx for PCA. Use similar paradigms as for locally advanced cases: (a) 5-FU/RT, (b) gemcitabine/RT, or (c) induction gemcitabine/Taxotere/Xeloda (GTX) → 5-FU/RT, with RT to 45–50.4 Gy in 1.8–2Gy/fx or 30 Gy in 3Gy/fx per the MDACC paradigm (Breslin TM et al., Ann Surg Oncol 2001).

What is the Tx paradigm for locally advanced PCA?

Locally advanced unresectable PCA Tx paradigm: Biliary stent (if jaundice) can be done 1st → (a) definitive CRT, (b) induction chemo → restage → CRT, (c) gemcitabine + RT, or (d) gemcitabine-based chemo.

What definitive CRT regimen should be used for locally advanced PCA?

Standard regimen for definitive CRT: 5-FU (CI 250 mg/m2/d) + RT, with RT to 50–60 Gy in 1.8–2 Gy/fx or 30 Gy in 3 Gy/fx.

What study established the role of definitive CRT vs. RT alone in locally advanced PCA? What were the study arms and survival outcomes?

The GITSG 9273 trial (pts enrolled in the 1970s). Arm 1: RT alone, split-course 60 Gy in 2 Gy/fx; arm 2: 5-FU + RT, split-course RT to 40 Gy in 2 Gy/fx; arm 3: 5-FU + RT, split-course RT to 60 Gy in 2 Gy/fx. All arms rcv maintenance 5-FU × 2 yrs. MS favored the CRT arms: 5.3 mos (arm 1) vs. 9.7 mos (arm 2) vs. 9.3 mos (arm 3). 1-yr OS favored the CRT arms: 10% (arm 1) vs. 35% (arm 2) vs. 46% (arm 3). There were no statistical differences between the CRT arms. (Moertel CG et al., Cancer 1981)

What study suggests that gemcitabine alone may be superior to 5-FU–based CRT in locally advanced PCA? What were the study arms and survival outcomes?

FFCD/SFRO study (French). Arm 1: RT (60 Gy) + CI 5-FU + intermittent cisplatin → maintenance gemcitabine; arm 2: induction gemcitabine → maintenance gemcitabine. Induction CRT was more toxic and had worse survival outcomes. MS was 8.6 mos vs. 13 mos, and 1-yr OS was 32% vs. 53%. Criticism: The induction CRT was not standard and was very poorly tolerated. (Chauffert B et al., Ann Oncol 2008)

What study suggests that concurrent gemcitabine-based CRT is superior to gemcitabine alone for locally advanced PCA?

ECOG 4201 (abstract: ASCO 2008). The trial closed early due to slow accrual. 71 pts (69 evaluable). Arm 1: gemcitabine alone; arm 2: gemcitabine + RT. MS was better in CRT (11 mos vs. 9.2 mos, p = 0.034) as well as 2-yr OS (12% vs. 4%).

What is the Tx paradigm for metastatic PCA?

Metastatic PCA Tx paradigm: gemcitabine alone, gemcitabine/erlotinib, or a clinical trial

What study suggested that gemcitabine/erlotinib may be superior to gemcitabine alone in metastatic PCA?

NCI Canada study of gemcitabine +/− erlotinib. The addition of erlotinib was associated with improved MS (5.9 mos vs. 6.2 mos) and 1-yr OS (17% vs. 23%). (Moore MJ, et al., JCO 2007)

What study demonstrated the benefit of adding bevacizumab (Avastin) to gemcitabine/erlotinib in metastatic PCA?

AViTA trial (ASCO 2008): gemcitabine/erlotinib/bevacizumab vs. gemcitabine/erlotnib. There was better PFS (4.6 mos vs. 3.6 mos) but not OS (7.1 mos vs. 6 mo, NSS) with the addition of bevacizumab.

For pts with resected PCA, LF is the site of 1st failure for what % of pts treated with adj CRT? Distant failure as the 1st site?

Based on RTOG 9704, LF was site of 1st failure in 23%–28% of PCA and distant failure was 1st site in 71%–77%.

Estimate the MS and 3-yr OS for pts with resectable PCA at Dx.

Outcomes for resected PCA, if treated with adj CRT: MS was ~20 mos and 3-yr OS was ~30% based on RTOG 9704; if untreated, MS was 11–12 mos.

Estimate the MS and 1-yr OS for pts with locally advanced PCA at Dx.

Outcomes for locally advanced PCA: MS was ~9 mos and 1-yr OS was ~35% based on GITSG 9273.

Estimate the MS and 1-yr OS for pts with metastatic PCA at Dx.

Outcomes for metastatic PCA: MS was ~6 mos, and 1-yr OS was ~20% based on an NCIC study.

In PCA pts, what are 3 Tx options for tumor-associated biliary obstruction?

Tx options for tumor-associated biliary obstruction:

1.     Endoscopic biliary stent

2.     Percutaneous biliary drainage with subsequent internalization

3.     Open biliary-enteric bypass

In PCA pts, what are 3 Tx options for tumor-associated gastric outlet obstruction?

Tx options for tumor-associated outlet obstruction:

1.     Gastrojejunostomy

2.     Enteral stent

3.     PEG tube

In PCA pts, what Tx should be considered for tumor-associated severe abdominal pain when medicinal therapy is no longer effective?

Celiac plexus neurolysis is an effective option for Tx-refractory pain.

Approximately at what vertebral bodies are the pancreas, celiac axis, and SMA located?

1.     Pancreas: L1-2

2.     Celiac axis: T12

3.     SMA: L1

In pts with adenocarcinoma of the pancreatic head receiving adj CRT, what structures and regional nodes should be covered by the RT field?

Classic adj fields for head lesions cover the tumor bed, pancreaticoduodenal nodes, local suprapancreatic nodes (not entire pancreas), celiac nodes, porta hepatis nodes, and SMA/SMV nodes. For unresectable tumors or neoadj Tx, the trend is now toward smaller fields that treat the tumor plus a small margin (McGinn CJ et al., JCO 2001).

In the absence of 3D planning, what are the borders for the initial AP and lat RT fields for PCA of the head?

Classic pancreatic head fields rcv 45 Gy in 1.8 Gy/fx.

1.     AP superior: T10-11 interspace

2.     AP inferior: L3-4 interspace

3.     AP left: 2 cm from left border of vertebral body

4.     AP right: 2 cm to right of preop duodenum

5.     Lateral superior-inferior: same as AP

6.     Lateral posterior: 2 cm into vertebral body

7.     Lateral anterior: 2 cm ant to preop gross Dz or duodenum

Is there any data on stereotactic body radiation therapy (SBRT) for pancreatic cancer?

There are several emerging studies that have evaluated SBRT for unresectable pancreatic cancer. Early reports suggest excellent LC; however, a significant proportion of pts experience duodenal toxicity.

Stanford (Chang DT et al., Cancer 2009): 77 pts with unresectable PCA rcv 25 Gy x 1 with CyberKnife SBRT. The 6- and 12-mo isolated LR rate was 5% and 5%, respectively. The PFS at 6 and 12 mos were 26% and 9%, respectively. The 1-yr OS was 21%. At 12 mos, the ≥G2 late toxicity was 25%.

Beth Israel Deaconess (Mahadevan A et al., IJROBP 2010). 36 pts with unresectable PCA rcv 24–36 Gy CyberKnife SBRT in 3 fx. Gemcitabine was given after SBRT. The LC rate was 78%, and MS was 14.3 mos. 39% developed ≥G2 toxicity (25% G2, 14% G3).

Is there a potential for dose escalation with the use of IMRT in pancreatic cancer?

YesSpalding et al. described potential benefits of IMRT in the delivery of RT for pancreatic malignancies. The median increase was from 52–66 Gy, achieving a boost dose upward of 85 Gy. (Med Phys 2007)

Do Tx paradigms differ between pancreatic and periampullary adenocarcinoma?

Yes. Tx paradigms can differ between pancreatic vs. periampullary cancers. Consider observation for completely resected T1-T2, N0 ampulla of Vater carcinoma. Retrospective reviews suggest high OS (5-yr OS ~80%) with observation alone (Willett C et al. Surg Gynecol Obstet 1993). Otherwise, periampullary adenocarcinoma generally follows PCA paradigms, especially for T3-T4 (Krishnan S et al., IJROBP 2008) or node+ pts (Zhou J et al., Radiother Oncol 2009).

image Toxicity

What are the expected acute and late RT toxicities associated with Tx of resected and unresectable PCA?

1.     Acute toxicities: nausea, diarrhea, small bowel obstruction, weight loss

2.     Late toxicities: small bowel obstruction/stenosis/perforation, gastric/small bowel ulceration and/or bleeding, biliary stenosis obstruction, 2nd malignancies

What are the standard dose limitations for the liver, kidneys, spinal cord, and stomach and small bowel for pancreatic RT?

Standard dose limits for normal structures during pancreatic cancer RT:

1.     Liver: 50% <30 Gy, limit whole liver to ≤20 Gy.

2.     Kidneys: at least two thirds of 1 kidney <18 Gy

3.     Spinal cord: ≤45 Gy

4.     Stomach and small bowel: up to 50 Gy to small volumes

What is the typical follow-up schedule after Tx of pancreatic cancer?

Pancreatic cancer follow-up schedule after Tx: surveillance every 3–6 mos × 2 yrs, then annually. At each visit, perform full H&P for Sx assessment, and consider CA 19-9 levels and contrast CT scan chest/abdomen. (NCCN 2010)