Gastrointestinal Oncology - part 3 docx

occupa-I DENTIFICATION OF P ATIENTS AT R ISK FOR F AMILIAL P ANCREATIC C ANCERThe identification of patients at increased risk for familial pancreatic cancer can bechallenging because fa

20 years The roles of other potential risk factors such as race, gender, diet, or tion have yet to be examined in the familial setting Nonetheless, all patients with a fam-ily history of pancreatic cancer should be advised not to smoke

occupa-I DENTIFICATION OF P ATIENTS AT R ISK FOR F AMILIAL P ANCREATIC C ANCERThe identification of patients at increased risk for familial pancreatic cancer can bechallenging because familial aggregation of pancreatic cancer may be influenced by anumber of factors, including chance, family size, shared environmental exposures, andincomplete gene penetrance For example, a shared environmental factor could hypo-thetically result in multiple first-degree relatives with pancreatic cancer Conversely, if aparent with a pancreatic cancer gene with complete penetrance has 4 children, there is a6.25% chance that none of the children will be affected If this hypothetical gene is alsoincompletely penetrant, the chance that no offspring will be affected is even greater.Furthermore, it can be difficult or impossible to obtain accurate family history data forsuch a rapidly and uniformly fatal disease, and the clinician is often unable to elicit anaccurate family history for more than one or two generations of a family

For these reasons, there is currently no clear consensus on the definition of familialpancreatic cancer However, families known to segregate mutations associated with can-cer syndromes and those patients with hereditary pancreatitis should be considered atincreased risk It also appears that the absolute number of affected first-degree relatives

is correlated with increased cancer risk, and most experts would agree the families that

74 Chapter 5

Figure 5-1 Earlier onset and increased risk of pancreatic cancer in members of familial

pancreatic cancer kindreds who smoke (Reprinted from Gastroenterology, Vol 124(5),

1292-9, Rulyak SJ, et al, Risk factors for the development of pancreatic cancer in ial pancreatic cancer kindreds, Copyright 2003, with permission from the AmericanGastroenterological Association.)

famil-include 2 or more first degree relatives with pancreatic cancer should be considered tohave familial pancreatic cancer until proven otherwise Families that include multiplesecond-degree relatives with pancreatic cancer and/or those with an unusually young age

of onset (<50 years of age) may also be at increased risk, although these criteria are morecontroversial These latter criteria may be most useful in evaluating small families and/orthose with limited family history data

It is important for the clinician to recognize that the risk to an individual patient

varies by both the number and relationship of affected relatives For example, a patient

whose mother and brother have been diagnosed with pancreatic cancer is likely to be atmuch higher risk of developing pancreatic cancer than if his or her uncle or first cousinwere affected Table 5-4 gives criteria that may prompt a clinician to consider the diag-nosis of familial pancreatic cancer, although it is important to emphasize that estimation

of risk for an individual patient is likely to be imprecise and that listed criteria have notbeen validated prospectively

Identification and Management of Familial Pancreatic Cancer 75

Table 5-4

C RITERIA FOR THE I DENTIFICATION OF P ATIENTS AT R ISK

FOR I NHERITED P ANCREATIC C ANCER

Increased Risk for Inherited Pancreatic Cancer

1 Patients with 2 or more first- degree relatives with pancreatic cancer

2 Patients known to harbor mutations that result in cancer syndromes ated with pancreatic cancer

associ-• Familial breast cancer (BRCA2)

• Familial atypical multiple mole melanoma syndrome (CDKN2A)

• Peutz-Jeghers syndrome (STK 11)

• Familial adenomatous polyposis (APC)

• Patients with hereditary pancreatitis*

Possibly Increased Risk for Inherited Pancreatic Cancer

1 Individuals with one first-degree relative and one or more second-degree ative(s) diagnosed with pancreatic cancer

rel-2 Individuals with one first-degree relative diagnosed with pancreatic cancer at

an early age (under the age of 50)

3 Individuals with 2 or more second-degree relatives with pancreatic cancer,one of whom developed it at an early age

4 Patients meeting clinical criteria for one of the above cancer syndromes without a clearly defined mutation

5 Members of HNPCC families

*Endoscopic screening is of uncertain utility as the findings of chronic pancreatitis are ilar to the findings of dysplasia

sim-S CREENING FOR P ANCREATIC C ARCINOMA

IN F AMILIAL P ANCREATIC C ANCERPancreatic cancer is most often asymptomatic until it has progressed to an incurablestage The nearly uniform lethality of pancreatic cancer is in part due to the lack of anacceptable screening test for pancreatic cancer At present, there are no adequate radi-ographic, endoscopic, or laboratory tests for use in population-based pancreatic cancerscreening However, the risk of familial pancreatic cancer approaches 50% in some kin-dreds, and methods for early detection of pancreatic cancer and precancer in such fam-ilies are clearly needed The goal of such strategies is to identify family members withpancreatic dysplasia or very early pancreatic cancers before they progress to invasive dis-ease

DYSPLASIA AS APRECURSOR TOPANCREATICCANCER

The discovery of a dysplastic precursor lesion has been critical to the development ofscreening and surveillance tests for other gastrointestinal cancers including those arising

in the colon, esophagus, and stomach Pancreatic adenocarcinomas also appear to arisefrom dysplasia, which is more correctly referred to as PanIN There are different grades

of PanIN, which appear to represent a continuum of neoplastic progression (Figure 2) PanIN-1 is also referred to as intraepithelial ductal hyperplasia and is characterized

5-by elongation of epithelial cells with abundant supranuclear mucin (PanIN-1A), times with papillary architecture (PanIN-1B) PanIN-2, or LGD, is defined by nuclearabnormalities including enlargement and crowding, hyperchromatism, and stratifica-tion PanIN-3, also referred to as HGD or carcinoma in-situ, is characterized by papil-lary projections, loss of nuclear polarity, nuclear atypia, and mitoses The prevalence ofPanIN in the general population is currently unknown, but a recent study of a referralpopulation of familial pancreatic cancer kindreds suggests that the prevalence of PanIN-

some-2 and PanIN-3 may be as high as 34% if stringent criteria are used to select family bers for screening

mem-76 Chapter 5

Figure 5-2 Histology of PanIN (A) PanIN-1 (open arrow) is characterized by elongation

of epithelial cells with abundant supranuclear mucin and PanIN-2 (solid arrow) is defined

by nuclear abnormalities including enlargement and crowding, hyperchromatism, andstratification (B) In PanIN-3, there are lush papillary projections, loss of nuclear polarity,

and nuclear atypia with mitoses (Photomicrographs courtesy of Dr Teresa Brentnall.) For

a full-color version, see page CA-IV of the Color Atlas.

There are several important limitations in the literature regarding PanIN lesions andtheir progression to cancer First, the absolute risk of pancreatic cancer in patients withpancreatic dysplasia has yet to be determined, as does the prevalence of PanIN in thegeneral population Second, the time course for progression from PanIN to cancer is asyet unknown In one series of 3 patients who underwent pancreatic surgery for non-malignant indications and were found to have dysplasia at the resection margins, allpatients progressed to cancer within 10 years Clearly, larger studies of the natural his-tory of PanIN are needed to confirm this finding Finally, it is important to note thatthe accurate pathologic diagnosis of PanIN can be difficult In one study, interobserveragreement among expert gastrointestinal pathologists was found to be only fair forPanIN 1 and PanIN 3 (kappa=0.43 and kappa=0.42, respectively) and poor for PanIN

2 lesions (kappa=0.14) Accurate diagnosis of PanIN is critical because PanIN is a tifocal or diffuse process and total pancreatectomy is necessary to ensure removal of alldysplastic tissue It is likely that the accuracy of PanIN diagnoses will improve as pathol-ogists accrue additional experience with this lesion

mul-While the natural history of pancreatic dysplasia is difficult to study and pletely characterized, PanIN represents a curable precursor lesion and thus it is possiblethat effective screening tests can be developed Potential screening methods may involvethe use of imaging modalities such as endoscopy or radiography Ultimately, a biomark-

incom-er in sincom-erum or pancreatic juice may be most attractive option for screening PanINs sharemany of the molecular genetic alterations found in adenocarcinomas Because thesemolecular alterations appear to be acquired in a step-wise fashion, it may eventually bepossible to detect the highest risk precursor lesions shortly before cancer develops

ENDOSCOPICAPPROACHES FORSCREENING

Endoscopic techniques have been used to detect PanIN or early pancreatic cancer,and endoscopic screening protocols for high-risk members of familial pancreatic cancerkindreds are ongoing at several institutions in the United States and abroad At present,EUS appears to be the most promising screening test for identifying pancreatic dyspla-sia and early pancreatic cancer because it is relatively non-invasive and has the ability todiscriminate subtle abnormalities in the pancreatic parenchyma The pancreatic abnor-malities on EUS that are associated with dysplasia are the same findings that are present

on EUS in patients with chronic pancreatitis These changes include the presence ofhyperechoic strands, hyperechoic foci, lobules (sometimes referred to as hypoechoicnodules), echogenic pancreatic duct walls, and irregularity of the pancreatic duct (Figure5-3) Several of these findings have been reported to be more common in relatives ofpatients with pancreatic cancer compared to controls undergoing EUS for another indi-cation (Table 5-5) More importantly, Brentnall and colleagues have shown that thesefindings can be used to identify members of familial pancreatic kindreds withhistopathologically-confirmed PanIN on pancreatic biopsy specimens

The findings seen on EUS should not be mistaken to represent endosonographicimaging of PanIN lesions, as EUS does not permit resolution of cellular details Instead,these findings should be thought of as endoscopic markers associated with dysplasia.These changes may represent inflammation, fibrosis, or fatty replacement of the pancreasassociated with PanIN, although the precise anatomic correlates of these findings have yet

to be determined One important limitation of EUS in screening is that identical ings can be found in patients with chronic pancreatitis, patients who consume alcohol,and occasionally in otherwise healthy adults undergoing EUS for other indications Identification and Management of Familial Pancreatic Cancer 77

find-ERCP may also have utility in detecting pancreatic dysplasia, although few wouldadvocate this as a first-line screening test given the potential for complications, includ-ing pancreatitis As with EUS findings, the ERCP findings associated with dysplasia canalso be seen in patients with chronic pancreatitis However, in selected members offamilial pancreatic cancer kindreds without clinical evidence for pancreatitis, severalERCP findings have been associated with dysplasia These include irregularity of main

78 Chapter 5

Figure 5-3 EUS findings

associated with pancreatic

dysplasia in familial

pan-creatic cancer kindreds

Radial imaging of the

pan-creas reveals hyperechoic

strands (small arrows) and

hypoechoic lobules (larger

arrows) in a patient

subse-quently diagnosed with

PanIN 3 (Image courtesy

of Dr Michael Kimmey.)

Table 5-5

F INDINGS OF EUS E XAMINATIONS OF M EMBERS OF

F AMILIES W ITH 3 OR M ORE M EMBERS W ITH P ANCREATIC

C ANCER ( N =30) C OMPARED TO C ONTROLS U NDERGOING

EUS FOR O THER I NDICATIONS ( N =103)

Finding FPC relatives (%) Controls (%) p-value

*These findings, also known as "lobules," are multifocal hypoechoic areas throughout the

pancreatic parenchyma measuring between 2 mm and 8 mm The term lobule is used to

distinguish these findings from those of a discrete mass lesion within the pancreasAdapted from Jagannath S, et al Endoscopic ultrasound abnormalities in at-risk relativesfrom familial pancreatic cancer kindreds: a prospective, controlled pilot cohort study

Gastrointest Endosc 2002;56(Suppl):S120.

pancreatic duct or side branches, ectasia of pancreatic duct side branches, sacculations ofpancreatic ducts, and early acinarization of pancreatic head during pancreatic duct injec-tion with inability to completely fill the pancreatic tail (Figure 5-4) ERCP should beused primarily as a confirmatory test when abnormalities on EUS are identified,although one potential advantage of ERCP is ability to obtain pure pancreatic juice orpancreatic duct brushings, which may provide material for screening if an accuratemolecular screening assay can be developed It is important to emphasize that the use ofendoscopic screening tests in familial pancreatic cancer is a field under development, andall screening should be conducted in the context of an approved clinical protocol

RADIOGRAPHIC APPROACHES FOR SCREENING

If the goal of a screening program is to detect PanIN, it is unlikely that currentlyavailable radiographic imaging modalities such as helical CT or magnetic resonancecholangiopancreatography (MRCP) have the ability to resolve either the histopatholog-

ic abnormalities seen in PanIN or the fibroinflammatory changes detected with EUS.However, improvements in technology may eventually allow detection of pancreaticcancers at a very early stage where the potential of cure is reasonable However, until theclinical utility of radiographic tests is demonstrated, their use outside of research proto-col is inadvisable unless patients report symptoms that warrant cross-sectional imaging

MOLECULAR APPROACHES TO SCREENING

The detection of molecular alterations in either blood or pancreatic juice may have

the potential for the early detection of pancreatic cancer or PanIN Mutations in k-ras have been the most widely studied of these alterations While k-ras can be detected in

pancreatic juice, pancreatic brushings, or fine needle aspirates of the pancreas, none ofthese assays have adequate specificity for the diagnosis of adenocarcinoma or dysplasia

because k-ras mutation are also present in benign conditions, including chronic atitis p53 mutations can also be detected in pancreatic juice However, p53 mutations

pancre-appear to be a late event in pancreatic carcinogenesis and these mutations are absent in

up to 50% of adenocarcinomas, thereby limiting sensitivity for the detection of

pancre-atic cancers or dysplasia Mutations in CDKN2A (p16) and SMAD4 (DPC4) can be

Identification and Management of Familial Pancreatic Cancer 79

Figure 5-4 ERCP findings

associated with

pancreat-ic dysplasia in familialpancreatic cancer kin-dreds Markedly abnormalpancreatogram reveals anirregular, dilated mainpancreatic duct with mul-tiple side branch saccula-tions within the pancreat-

ic head (Image courtesy

of Dr Michael Kimmey.)

identified from pancreatic juice, but one recent study suggests that testing for these

mutations adds little additional diagnostic information to the detection of K-ras It is

possible that combinations of screening markers may improve the accuracy of lar testing for pancreatic cancers There has been one study to suggest that combining

molecu-assays for K-ras and p53 may improve diagnostic accuracy for pancreatic

adenocarcino-ma However, no published studies have utilized combinations of molecular tests for thediagnosis of PanIN lesions

Telomere dysfunction is another promising molecular marker for screening of ial pancreatic cancer kindreds because it is an early and nearly ubiquitous alteration inPanIN Telomere length shortening has been reported to be present in 96% in PanINspecimens Assays of telomere length currently require histologically normal tissue as acontrol, and such tissue cannot be readily obtained by sampling pancreatic juice or byfine needle aspiration because PanIN is a diffuse and frequently multifocal process.However, telomerase activity is a surrogate for telomere shortening, and preliminaryreports suggest that telomerase activity can be determined from either pancreatic juice

famil-or fine needle aspirates It is impfamil-ortant to recognize that telomere dysfunction is

present-ly not onpresent-ly seen in low grade dysplasia (PanIN-2) and high grade dysplasia (PanIN-3),but also in pancreatic duct hyperplasia (PanIN-1) Therefore, any potential assay fortelomere dysfunction is likely to have high sensitivity but low specificity for the diagno-sis of dysplasia, although sensitivity is arguably the most important feature of a screen-ing assay If a suitable assay for telomere dysfunction can be developed, it may be mostuseful to select a subset of patients who warrant more intensive screening with EUS orother modalities

In patients with a family history suggestive of one of the defined cancer syndromesthat result in pancreatic cancer, testing for germline mutations may represent an impor-tant tool for identifying family members at risk for pancreatic cancer The yield of rou-tine testing of familial pancreatic cancer kindreds for most of these mutations is likely to

be low One exception may be testing for BRCA2, as recent studies suggest that up to 19% of familial pancreatic cancer kindreds may harbor a BRCA2 mutation in the

absence of a family history of breast or ovarian cancer However, future studies to

con-firm the prevalence of BRCA2 mutations among familial pancreatic cancer kindreds are

needed before such testing can be widely recommended

As the pathways that lead to pancreatic adenocarcinoma are better understood,molecular approaches may assume an increasing role in screening and surveillance ofhigh-risk family members With the advent of proteomic technology, it is likely thatnewer markers with improved sensitivity and specificity will become available Whilemolecular techniques hold great promise, there are presently few studies to suggest thattesting for single molecular genetic alterations or combinations thereof have utility inclinical detection of PanIN or early pancreatic cancer As with other screening modali-ties, their use should be restricted to investigational protocols

A PPROACH TO P ATIENTS W ITH A F AMILY H ISTORY

OF P ANCREATIC C ANCERThe management of members of familial pancreatic cancer kindreds presents a for-midable challenge to the clinician because clinical experience with screening amonghigh-risk pancreatic cancer patients is limited and currently evolving However, severalcenters across the United States have initiated protocols to study the role of screening

80 Chapter 5

and surveillance for high-risk family members, and early results from these studies arepromising One protocol using endoscopic screening has been developed at theUniversity of Washington Medical Center in Seattle To date, 73 patients from 50 dif-ferent familial pancreatic cancer kindreds have been enrolled and follow-up extends to 9years From this cohort, 9 patients have been diagnosed with PanIN 3 and 7 have beendiagnosed with PanIN 2 Thirteen of these patients have elected to undergo total pan-createctomy, with no operative deaths Most importantly, no patient enrolled in the pro-tocol has developed pancreatic cancer while under surveillance

While these results are encouraging, it must be emphasized that experience withendoscopic screening is limited, and endoscopic tests such as EUS are operator depend-ent Therefore, screening with EUS should only be entertained in patients at greatest riskfor pancreatic dysplasia, and patients should be well-informed, active participants in thedecision to proceed with screening The findings of EUS should be corroborated byother studies, including ERCP and ultimately laparoscopic biopsy of the pancreatic tail,prior to proceeding with total pancreatectomy Finally, it cannot be overemphasized thatthe natural history of pancreatic dysplasia is uncertain and the performance characteris-tics of endoscopic tests such as EUS and ERCP have yet to be determined However, arecent decision analysis that assumed a 90% sensitivity of EUS for dysplasia suggestedthat a strategy employing EUS to screen members of familial pancreatic cancer kindreds

is cost-effective, although the benefit appears to be limited to patients with a pre-testprobability of pancreatic dysplasia of 16% or greater

In all patients with a family history of pancreatic cancer, evaluation should beginwith a careful history and physical examination in order to determine if alarm symptomssuch as abdominal or back pain, diarrhea, weight loss, or recent onset diabetes are pres-ent A history of smoking, alcohol consumption, or occupational exposures should also

be elicited All patients with a family history of pancreatic cancer should be counselednot to smoke A careful family history should be obtained in order to construct adetailed family pedigree, which can then be used to select patients who may be candi-dates for screening Patients with 2 or more first-degree family members diagnosed withpancreatic cancer are considered at increased risk of pancreatic cancer, as are patientswho are known to harbor a germline mutation associated with pancreatic cancer.Patients with less compelling family histories may also be at increased risk of pancreaticcancer, although the utility of screening in such patients is far less certain Table 5-4 listsproposed criteria for pancreatic cancer risk stratification, with the understanding thatsuch criteria have yet to be prospectively validated Patients with hereditary or acquiredchronic pancreatitis represent a particular challenge, because the EUS and ERCP find-ings in patients with these conditions are indistinguishable from the findings of dyspla-sia Therefore, endoscopic screening patients with chronic pancreatitis cannot be wide-

ly recommended, although experience with this population of patients is limited Once a patient is deemed to be at increased risk for pancreatic cancer, extensivecounseling regarding the uncertainties about the natural history of PanIN and lack ofclearly validated methods for screening and surveillance for pancreatic neoplasia shouldthen be undertaken Counseling must include a frank discussion about the outcome ofscreening if abnormalities are uncovered, specifically total pancreatectomy In patientswith family histories suggestive of a known cancer syndrome, genetic counseling andmutational analysis should be considered Patients who remain interested after a thor-ough discussion of risks and benefits are then offered screening

Identification and Management of Familial Pancreatic Cancer 81

The algorithm used at the University of Washington for screening patients who meetcriteria for familial pancreatic cancer is shown in Figure 5-5 Interested patients areoffered EUS as the initial screening test, beginning either at the age of 50 years or 10years prior to the earliest age at which pancreatic cancer was diagnosed in an affectedfamily member If EUS is abnormal and there is no recent history of alcohol use, thepatient is then offered ERCP to confirm the pancreatic abnormalities noted on EUS.However, if the patient actively consumes alcohol, he or she is advised to abstain fromalcohol for 6 months and the EUS is repeated to confirm that abnormalities persist Ifboth EUS and ERCP display characteristic abnormalities, the patient is referred for

82 Chapter 5

Figure 5-5 Approach to screening for pancreatic cancer in familial pancreatic cancer

kindreds

laparascopic distal pancreatectomy to confirm the presence of dysplasia Patients withhigh-grade dysplasia (PanIN-3) are advised to undergo completion total pancreatecto-

my, which is required to ensure removal of all dysplastic tissue Postoperatively, thesepatients require lifelong pancreatic enzyme replacement and close follow-up by a dia-betologist for diabetes mellitus Experience with the management of patients with LGD(PanIN-2) limited, and patients are offered the choice of pancreatectomy or continuedclose surveillance To date, several patients with PanIN-2 have elected to undergo pan-createctomy, while others have chosen to continue surveillance with EUS and/or ERCP.Patients with normal findings on EUS are advised to undergo repeat examination on

a yearly basis, as at least one case of progression from a normal EUS to cancer in lessthan 12 months has occurred It remains to be determined whether the surveillanceinterval may be lengthened after several normal examinations Patients with an abnor-mal EUS examination but normal findings at ERCP are recommended to undergorepeat EUS in 1 year with close comparison of findings Patients with progression ofendosonographic changes undergo repeat ERCP, with laparoscopic biopsy reserved forpatients with abnormalities on ERCP However, if the abnormal EUS findings are stable

in such patients, the examination is repeated on a yearly basis

C ONCLUSIONApproximately 10% of pancreatic cancers are inherited A number of genetic syn-dromes can predispose mutation carriers to pancreatic cancer, including FAP, FAMMM,Peutz-Jeghers syndrome, hereditary breast/ovarian cancer syndrome, hereditary pancre-atitis, and possibly HNPCC However, the gene or genes involved in the majority ofinherited pancreatic cancers have yet to be determined, although the molecular pathways

of pancreatic tumorigenesis are beginning to be better understood Smoking appears to

be an important modifier of familial pancreatic cancer risk, but the role of other ronmental factors in such families has yet to be determined Screening and surveillanceprotocols using endoscopic techniques such as EUS and ERCP have been developed, buttheir use should be restricted to approved research protocols Molecular genetic testingmay have the greatest potential for screening and surveillance, although tests withimproved sensitivity and specificity will need to be developed Most importantly, what

envi-is learned from familial pancreatic cancer kindreds will have important implications forscreening and prevention of pancreatic cancer in the general population

B IBLIOGRAPHYBanke MG, Mulvihill JJ, Aston CE Inheritance of pancreatic cancer in pancreatic cancer-

prone families Med Clin North Am 2000;84(3):677-690.

Brat DJ LK, Yeo CJ, Warfield PB, Hruban RH Progression of pancreatic intraductal

neo-plasias to infiltrating adenocarcinoma of the pancreas Am J Surg Pathol

1998;22:163-169

Brentnall TA, Bronner MP, Byrd DR, Haggitt RC, Kimmey MB Early diagnosis and

treat-ment of pancreatic dysplasia in patients with a family history of pancreatic cancer Ann Intern Med 1999;131(4):247-255.

Eberle MA, Pfutzer R, Pogue-Geile KL, et al A new susceptibility locus for autosomal

dom-inant pancreatic cancer maps to chromosome 4q32-34 Am J Hum Genet 2002;70(4):

1044-1048

Identification and Management of Familial Pancreatic Cancer 83

Giardiello FM, Welsh SB, Hamilton SR, et al Increased risk of cancer in the Peutz-Jeghers

syndrome N Engl J Med 1987;316:1511-1514.

Giardiello FM, Offerhaus GJ, Lee DH, et al Increased risk of thyroid and pancreatic

carci-noma in familial adecarci-nomatous polyposis Gut 1993;34(10):1394-1396.

Goldstein AM, Fraser MC, Struewing JP, et al Increased risk of pancreatic cancer in

melanoma-prone kindreds with p16INK4 mutations N Engl J Med 1995;333(15):

970-974

Goldstein AM, Struewing JP, Chidambaram A, Fraser MC, Tucker MA type relationships in U.S melanoma-prone families with CDKN2A and CDK4 muta-

Genotype-pheno-tions J Natl Cancer Inst 2000;92(12):1006-1010.

Hahn SA, Greenhalf B, Ellis I, et al BRCA2 germline mutations in familial pancreatic

carci-noma J Natl Cancer Inst 2003;95(3):214-221.

Hruban RH, Adsay NV, Albores-Saavedra J, et al Pancreatic intraepithelial neoplasia: a new

nomenclature and classification system for pancreatic duct lesions Am J Surg Pathol.

2001;25(5):579-586

Kimmey MB, Bronner MP, Byrd DR, Brentnall TA Screening and surveillance for hereditary

pancreatic cancer Gastrointest Endosc 2002;56(4 Suppl):S82-86.

Klein AP, Beaty TH, Bailey-Wilson JE, Brune KA, Hruban RH, Petersen GM Evidence for

a major gene influencing risk of pancreatic cancer Genet Epidemiol

2002;23(2):133-149

Lowenfels AB, Maisonneuve P, DiMagno EP, et al Hereditary pancreatitis and the risk of

pancreatic cancer International Hereditary Pancreatitis Study Group J Natl Cancer Inst.

1997;89:442-446

Lowenfels AB, Maisonneuve P, Whitcomb DC, Lerch MM, DiMagno EP Cigarette smoking

as a risk factor for pancreatic cancer in patients with hereditary pancreatitis JAMA.

2001;286(2):169-170

Lowenfels AB, Maisonneuve P Epidemiologic and etiologic factors of pancreatic cancer

Hematol Oncol Clin North Am 2002;16(1):1-16.

Lynch HT, Smyrk T, Kern SE, et al Familial pancreatic cancer: a review Semin Oncol 1996;

23:251-255

Lynch HT, Voorhees G, Lanspa S, McGreevy PS, Lynch JF Pancreatic carcinoma and

hered-itary nonpolyposis colorectal cancer: A family study Br J Cancer 1985;52:27-28.

Lynch HT, Brand RE, Hogg D, et al Phenotypic variation in eight extended CDKN2Agermline mutation familial atypical multiple mole melanoma-pancreatic carcinoma-prone families: the familial atypical mole melanoma-pancreatic carcinoma syndrome

Cancer 2002;94(1):84-96.

Murphy KM, Brune KA, Griffin C, et al Evaluation of candidate genes MAP2K4, MADH4,ACVR1B, and BRCA2 in familial pancreatic cancer: deleterious BRCA2 mutations in

17% Cancer Res 2002;62(13):3789-3793.

Neglia JP, FitzSimmons SC, Maisonneuve P, et al The risk of cancer among patients with

cys-tic fibrosis Cyscys-tic Fibrosis and Cancer Study Group N Engl J Med 1995;332:494-499

Ozcelik H, Schmocker B, DiNicola N, et al Germline BRCA2 617delT mutations in

Ashkenazi Jewish pancreatic cancer patients Nat Genet 1997;16:17-18.

Rulyak SJ, Brentnall TA Inherited pancreatic cancer: surveillance and treatment strategies for

affected families Pancreatology 2001;1:477-485.

Rulyak SJ, Lowenfels AB, Maisonneuve P, Brentnall TA Risk factors for the development of

pancreatic cancer in familial pancreatic cancer kindreds Gastroenterology 2003;124(5):

1292-1299

84 Chapter 5

Rulyak SJ, Kimmey MB, Veenstra DL, Brentnall TA Cost-effectiveness of pancreatic cancer

screening in familial pancreatic cancer kindreds Gastrointest Endosc 2003;57(1):23-29.

Rulyak SJ, Brentnall TA, Lynch HT, Austin MA Characterization of the neoplastic type in the familial atypical multiple-mole melanoma-pancreatic carcinoma syndrome

pheno-Cancer 2003;98(4):798-804.

Schenk M, Schwartz AG, O’Neal E, et al Familial risk of pancreatic cancer J Natl Cancer Inst 2001;93(8):640-644.

Schutte M, Hruban RH, Geradts J, et al Abrogation of the Rb/p16 tumor-suppressive

path-way in virtually all pancreatic carcinomas Cancer Res 1997;57:3126-3130.

Tersmette AC, Petersen GM, Offerhaus GJ, et al Increased risk of incident pancreatic cancer

among first-degree relatives of patients with familial pancreatic cancer Clin Cancer Res.

2001;7:738-744

van der Heijden MS, Yeo CJ, Hruban RH, Kern SE Fanconi anemia gene mutations in

young-onset pancreatic cancer Cancer Res 2003;63(10):2585-2588.

van Heek NT, Meeker AK, Kern SE, et al Telomere shortening is nearly universal in

pancre-atic intraepithelial neoplasia Am J Pathol 2002;161(5):1541-1547.

Identification and Management of Familial Pancreatic Cancer 85

I NTRODUCTIONCarcinoma of the exocrine pancreas is a significant health problem in the UnitedStates and other Western nations It is one of the most lethal malignancies with an over-all 5-year survival rate of approximately 3% and an incident rate nearly equivalent to itsmortality rate Although a minority of patients are resectable at the time of diagnosis,surgical resection offers the only potentially curative treatment, and surgical treatmentclearly provides the best long-term results For patients in whom curative resection isnot possible, surgical intervention remains important as a means of palliation, particu-larly in patients with locally advanced disease.

This chapter reviews the various roles of surgical intervention in the curative andpalliative treatment of patients with pancreatic cancer, focusing on patients with carci-noma of the exocrine pancreas Preoperative evaluation, with an emphasis on the con-troversial and evolving approaches to preoperative staging and determination ofresectability, is discussed

P ANCREATIC N EOPLASIAPancreatic neoplasia can be subdivided into 3 clinical entities: 1) pancreatic duct celladenocarcinoma—frequently referred to as adenocarcinoma of the pancreas; 2) non-pancreatic periampullary cancer, which includes distal bile duct tumors, tumors origi-nating in the ampulla of Vater, and periampullary duodenal tumors; and 3) rare pan-creatic neoplasms, which include acinus cell carcinoma (a tumor that occurs more fre-quently in the tail of the pancreas), primary pancreatic lymphoma, cystic neoplasms ofthe pancreas including intraductal papillary mucinous tumors (IPMT), solid and pap-illary pancreatic neoplasms, and pancreatic neuroendocrine tumors The clinical pres-entation of these various entities is frequently similar Since the decision for resection isoften made without a tissue diagnosis, the surgical approach to these different neo-plasms is generally the same, although the prognosis and role for adjuvant therapy variesdepending on the histology

Surgical Approach to Ampullary and Pancreatic

Neoplasia

Kristoffel R Dumon, MD; Robert J Canter, MD;

and Noel N Williams, MD

Patients who undergo surgical resection for localized nonmetastatic adenocarcinoma

of the pancreatic head have a long-term survival rate of approximately 20% and a

medi-an survival of 12 to 20 months (Table 6-1) Disease recurrence following a potentiallycurative pancreaticoduodenectomy remains common In patients who undergo surgeryalone, local recurrence occurs in up to 86% of patients, peritoneal recurrence in 25%,and liver metastases in 50% When surgery and chemoradiation are used to maximizelocal regional tumor control, liver metastases become the dominant form of tumorrecurrence and occur in 25% to 53% of patients

Among patients with periampullary adenocarcinoma treated by denectomy, those with duodenal adenocarcinoma are most likely to be long-term sur-vivors Five-year survival is less likely for patients with ampullary, distal bile duct, andpancreatic primaries, in declining order Resection margin status, resected lymph nodestatus, and degree of tumor differentiation also significantly influence long-term out-come Particularly for patients with pancreatic adenocarcinoma, 5-year survival is notequated with cure, because many patients die of recurrent disease >5 years after resec-tion (Table 6-2)

pancreaticoduo-Pancreatic periampullary cancer is felt to have an improved prognosis compared toadenocarcinoma of the exocrine pancreas This may be secondary to both earlier diag-nosis and more favorable tumor biology Neoplasms of the ampullary region cause bil-iary obstruction when the primary tumor is relatively small Patients who undergo pan-creaticoduodenectomy for localized periampullary adenocarcinoma of nonpancreaticorigin have an improved survival duration compared with similarly treated patients whohave adenocarcinoma of pancreatic origin

I NTRADUCTAL P APILLARY M UCINOUS T UMORCystic neoplasms of the exocrine pancreas comprise a small fraction of pancreatictumors Within this group, awareness of intraductal papillary mucinous tumor (IPMT),

a disease heretofore commonly confused with chronic pancreatitis, has increased sincethe WHO classified these tumors as a separate group in 1996 This disease is character-ized by dilation of the main pancreatic duct or branch ducts associated with mucin over-production There may be peripheral lesions consisting of ectatic branch ducts connect-

ed to the main duct or cysts that do not connect with the main duct Either of these can

88 Chapter 6

Table 6-1

P ANCREATIC C ANCER S TAGE AND O VERALL P ROGNOSIS AT

THE T IME OF D IAGNOSIS

Patients Survival Survival

mimic so-called mucinous cystic neoplasm (MCN) Because the incidence of invasivecancer at surgery is 25 to 50%, it is important to distinguish this entity from chronicpancreatitis This is usually done on the basis of typical changes evident on CT andERCP The ERCP examination may reveal mucus exuding from the papilla or charac-teristic intraductal filling defects The lesion, even if it does not contain invasive cancer,

is premalignant, and benign lesions contain several genetic mutations associated withpancreatic cancer Because of their favorable prognosis, an extensive diagnostic workupfor IPMTs should be performed in patients presenting with cystic lesions of the pan-creas Surgical resection is the therapy of choice for IPMTs The type of resectiondepends upon the extent of the ductal involvement Total pancreatectomy is currentlythe recommended treatment for an IPMT that comprises the entire main duct

P REOPERATIVE E VALUATIONAccurate preoperative assessment of resectability is the most critical aspect of thediagnostic and treatment sequence for patients with pancreatic cancer Imaging includesnonoperative techniques, such as abdominal ultrasonography, CT scan, MRI, angiogra-phy, and a variety of invasive techniques, such as laparoscopy and laparoscopic or intra-operative ultrasonography

The studies for the diagnosis and staging of pancreatic cancer differ considerablyfrom center to center The challenge to this process is that a high percentage of patientswho currently undergo surgery with curative intent have unresectable disease The goal

of preoperative imaging is to spare patients a nontherapeutic laparotomy in those withunresectable pancreatic cancer However, this must be taken in the context of possiblepalliative surgery

HISTORY ANDPHYSICAL EXAMINATION

Virtually all patients with periampullary neoplasms and approximately 50% ofpatients with pancreatic cancer have jaundice due to extrahepatic biliary obstruction atthe time of diagnosis The classical presentation is painless jaundice in association with

Surgical Approach to Ampullary and Pancreatic Neoplasia 89

Table 6-2

S URGICAL R ESULTS FOR P ATIENTS U NDERGOING

P ANCREATICODUODENECTOMY FOR P ANCREATIC C ARCINOMA

Author Yeo Sohn Conlon Richter Billingsley

a palpable gallbladder (Courvoisier’s sign) Tumors arising in the ampulla of Vater orwithin the intrapancreatic portion of the common bile duct typically cause biliaryobstruction early in the course of disease and therefore may be associated with a betterprognosis Small tumors of the pancreatic head located near the intrapancreatic portion

of the bile duct may also obstruct the bile duct and cause the patient to seek medicalattention when the tumor is still localized and potentially resectable In contrast, ade-nocarcinomas arising in the pancreas that do not cause obstruction of the intrapancre-atic portion of the bile duct are often not diagnosed until locally advanced or metastat-

ic In the absence of extrahepatic biliary obstruction, few pancreatic cancer patients ent with potentially resectable disease

pres-If jaundice is not present, patient complaints are often nonspecific, as are clinicalsigns on physical examination The pain typical of locally advanced pancreatic cancer is

a dull, fairly constant pain of visceral origin localized to the region of the middle andupper back Some patients have vague, intermittent epigastric pain Fatigue, weight loss,and anorexia are common even in the absence of mechanical gastric outlet obstruction.Pancreatic exocrine insufficiency due to obstruction of the pancreatic duct may result inmalabsorption and steatorrhea Although malabsorption and mild changes in stool fre-quency are common, diarrhea is uncommon Onset of diabetes mellitus may herald theappearance of pancreatic cancer, particularly if the diabetes occurs during or beyond thesixth decade Diabetes mellitus is present in 60% to 81% of patients with pancreaticcancer, and the majority of patients receive the diagnosis within 2 years of recognition

of pancreatic cancer

There is a postulated increased association of thromboembolic disorder and atic carcinoma (Trousseau’s syndrome) as well as an association with cutaneous manifes-tations such as pancreatic panniculitis (Weber-Christian disease)

Since 7% to 8% of patients with pancreatic cancer have a family history of atic cancer (first-degree relative), family history is a key element in the detection of riskgroups

pancre-The importance of the clinical examination cannot be overstated This has 2 tives: to assess fitness for operation and to detect evidence of metastatic disease Clinicalfactors such as the patient’s performance status and cardiopulmonary function can indi-cate that the perioperative risk is prohibitive The presence of left supraclavicularadenopathy, umbilical nodes, peritoneal carcinomatosis, or ascites can provide evidence

objec-of diffuse malignant disease Clinically apparent disseminated disease is an obvious traindication to attempted pancreatic resection

con-LABORATORYTESTS/TUMOR MARKERS

Several tumor markers have been evaluated for pancreatic cancer These includeserum mucins (such as CA 19.9), pancreatic amyloid (IAPP), and genetic markers (K-

ras, p53, p15/MTS2) Because of the limited sensitivity and specificity, tumor markers

are currently not widely used in clinical practice but some studies suggest that tumormarkers may have a useful role in monitoring clinical response after surgical resection orchemotherapy

The most widely used marker is the serum concentration of cancer-associated gen 19-9 (CA 19-9), which was found to have the greatest sensitivity (70%) and speci-ficity (87%) for diagnosis of pancreatic cancer with a cutoff value of 70 U/mL With alower cutoff of 37 U/mL, sensitivity was somewhat higher (86%) and specificity wasidentical (87%) However, biliary tract obstruction with cholangitis caused by a lesionother than cancer causes high levels of CA 19-9

anti-90 Chapter 6

Genetic markers may be used to detect pancreatic cancer, but the clinical value ofthese markers remains unproven The most common gene abnormality (90%) described

in pancreatic cancer is a K-ras mutation Mutations of the p53 tumor cell suppressor

gene are found in 50% to 70% of pancreatic cancers, and approximately 50% havereduced expression of the DCC gene A number of other gene deletions are less frequent

in pancreatic cancer, including alterations of tumor suppressor genes FHIT, 16/MTS1and p15/MTS2

K-ras mutation is the most widely studied genetic marker K-ras in pancreatic tions may be an early marker for pancreatic cancer, but whether K-ras mutations found

secre-in duodenal or pancreatic juice or stools of patients with chronic pancreatitis herald creatic cancer is not clear

pan-ULTRASOUND

Abdominal ultrasonography (US) is the initial screening technique because of its lowcost and easy availability US will confirm the presence of gallstones, assess the liver formetastatic deposits, detect abdominal ascites, and identify the level of biliary obstruc-tion In certain patients, US can enable one to manage the patient without any furtherinvestigation Patients with malignant ascites and liver deposits require confirmatorybiopsy and no further investigation

In addition, patients presenting with biliary obstruction are usually evaluated withabdominal US to confirm the mechanical nature of the obstruction and to determinewhether the site of obstruction is the intrahepatic or extrahepatic portion of the biliarytree Obstruction of the intrapancreatic portion of the bile duct is then evaluated with acombination of CT, ERCP, and EUS/MRCP

However, there remain many limitations to abdominal US in the evaluation ofpatients with pancreatic cancer Bowel gas obscures the image in up to 15% of patients

US is notoriously operator-dependent and not accurate in assessing central abdominaland retroperitoneal structures

CROSS-SECTIONALIMAGING/COMPUTEDTOMOGRAPHYSCAN

CT scan has undergone a rapid evolution over the last 2 decades, with each newdevelopment enhancing the imaging capability of the technique Conventional CT hasbeen superceded by dynamic thin section CT, spiral CT, and multidetector CT(MDCT) These new techniques have dramatically improved the ability of CT to diag-nose and stage pancreatic cancer (Table 6-3) High-quality CT can identify the majori-

ty of pancreatic tumors and accurately define the relationship of the tumor to the celiacaxis and superior mesenteric vessels

The use of standardized, objective radiologic criteria for preoperative tumor staging

is critical for treatment planning The following CT criteria definine a potentiallyresectable pancreatic cancer:

• The absence of extrapancreatic disease

• No evidence of arterial encasement/no direct tumor extension to the celiacaxis or superior mesenteric artery

• A patent superior mesenteric-portal venous (SMPV) confluence, assumingthe technical ability to resect isolated involvement of the superior mesentericvein (SMV) or SMPV confluence

Surgical Approach to Ampullary and Pancreatic Neoplasia 91

A patient is deemed to have locally advanced, unresectable disease when there is clearevidence on CT scans of the following:

• Hepatic and distant metastasis

• Locally advanced disease (eg, peripancreatic extension of tumor to locallycontiguous structures, vascular encasement or invasion of the SMA or celiacaxis)

• Occlusion of the SMPV confluence

Despite the wide application of conventional CT scan in the assessment of ity, it remains inaccurate in assessing local vascular invasion Therefore, most efforts havebeen directed to improving its ability to assess this factor

resectabil-The value of conventional CT in predicting unresectability has been reviewed insome studies In a study by McCarthy et al comparing CT assessment of resectabilitywith surgical findings, the sensitivity and specificity for CT prediction of resectabilitywere 72% and 80%, respectively Positive predictive value and negative predictive valuewere 77% and 76%, respectively (see Table 6-3) Of concern are false-positive and false-negative scans Most studies confirm that CT cannot be relied on entirely because somecriteria of unresectability––such as venous encasement, dorsal extension, and lym-phadenopathy––are too inaccurate with this imaging modality

The advent of dual-phase thin-cut spiral CT scan has offered the potential ofimproved accuracy Spiral CT can generate highly detailed images that enable a greaterappreciation of the relationship between the tumor and surrounding vascular structures.Spiral CT performed with contrast enhancement and a thin-section technique can accu-rately assess the relationship of the low-density tumor to the celiac axis, SMA, and supe-rior mesenteric–portal vein (SMPV) confluence Several authors have shown spiral CT

to be more accurate for diagnosing pancreatic cancer and for predicting resectability

92 Chapter 6

Table 6-3

P REOPERATIVE I MAGING OF P ANCREATIC C ANCER

Study Sensitivity (%) Specificity (%) Accuracy (%)

In the latest advance in spiral CT technology, volume rendering of spiral CT datacan be combined with a 3-dimensional display and is referred to as multi detector CT(MDCT) It allows the user to modify parameters to optimize visualization of structures.The viewing plane can be altered to allow inclusion of key elements of anatomy thatmight have been missed This provides CT angiography, which rivals the imagesobtained by conventional angiography Visualization of the bile and pancreatic ducts isalso optimized Some authors have demonstrated the incredibly accurate images thatMDCT can provide They believe that MDCT accurately assesses the patency of thesuperior mesenteric artery, celiac axis, and portal venous system Although software forthis technology is not currently widely available, it is likely to be in the near future Itsaccuracy and reliability remain to be proven.

MAGNETICRESONANCEIMAGING

Magnetic resonance imaging in the evaluation of pancreatic neoplasia was initiallylimited by image artifacts from respiration, aortic pulsation, bowel peristalsis, and a lack

of suitable contrast material for the gut lumen The introduction of ultrafast MRI(UMRI) has increased the usefulness of the technique (see Table 6-3) UMRI is moreaccurate than both US and CT in predicting resectability, with sensitivity, specificity,and overall accuracy of 95.7%, 93.5%, and 80.4%, respectively Some authors maintainthat UMRI may be superior to other imaging modalities, avoiding endoscopy, vascularcannulation, allergic reactions, and ionizing radiation The use of MRI to visualize thebiliary system without the administration of intravenous contrast is known as MRCP.MRCP had a sensitivity and specificity of 84% and 97%, respectively, for the diagnosis

of pancreatic carcinoma Although some authors believe that MRCP is not as accurate

as spiral CT for assessment of respectability, the ability of MRI to combine raphy and angiography in one sitting without exposure to ionizing radiation makes itattractive as a potential replacement to spiral CT scan as the imaging modality of choicefor pancreatic cancer

pancreatog-ENDOSCOPICRETROGRADE CHOLANGIOPANCREATOGRAPHY

If a low-density mass is not seen on CT scans, patients with extrahepatic biliaryobstruction undergo diagnostic and therapeutic ERCP A malignant obstruction of theintrapancreatic portion of the common bile duct is characterized by the double-duct sign,which represents proximal obstruction of the common bile and pancreatic ducts A malig-nant obstruction can often be accurately differentiated from choledocholithiasis and thelong, smooth tapering bile duct stricture seen with chronic pancreatitis Conversely, anormal pancreatogram does not entirely exclude malignancy Potentially “silent areas” onERCP are the uncinate process, the accessory duct, and the tail Small tumors just belowthe papilla will dilate the entire duct and are often missed on ERCP It can also be diffi-cult to differentiate between chronic pancreatitis and pancreatic cancer on ERCP ERCP also permits the sampling of pancreatic fluid for cytologic examination Thiscan help establish a definitive preoperative diagnosis However, a benign or nondiag-nostic result by ERCP-guided aspiration should not be interpreted as an indication tonot operate

Surgical Approach to Ampullary and Pancreatic Neoplasia 93

Positron emission tomography (PET) provides an alternative in tumors less than 2

cm in diameter In lesions of this size, it is often difficult to distinguish focal tis from tumor Preliminary studies indicate that PET has a positive predictive value fortumor of 92% compared with 91% for CT PET may also be able to differentiatebetween otherwise unclear pancreatic masses and may be superior to CT scan in distin-guishing benign from malignant cystic neoplasms of the pancreas Nevertheless, the truediagnostic value for this image modality remains unclear and results from larger studiesare needed

pancreati-ENDOSCOPICULTRASONOGRAPHY

EUS allows an ultrasound probe to be placed in close proximity to the pancreas Thiseliminates interference from overlying bowel gas and allows higher frequencies to beused, resulting in markedly improved resolution of images of the pancreas and sur-rounding structures EUS is not as widely available as spiral CT and requires extensiveexperience and knowledge of pancreatic ultrasonographic anatomy Although EUS givessuperior results compared with conventional CT scan, it appears to be equivalent to spi-ral CT scan in detecting tumors >3 cm and in assessing venous invasion and lymph nodeinvolvement However, its assessment of arterial invasion is limited Although it allowsthe pancreas to be assessed for small tumors (<3 cm) that may be missed on CT scan,EUS is an invasive procedure that requires sedation and is highly operator dependent.Moreover, it is the most reliable method to obtain accurate tissue for histologic diagno-sis

ENDOSCOPICULTRASOUND-GUIDEDFINENEEDLE ASPIRATION AND

PREOPERATIVETISSUEDIAGNOSIS

In most patients, accurate staging and biliary decompression are achieved with CTand ERCP, but these modalities cannot reliably distinguish nodal metastases frominflammatory lymphadenopathy or differentiate focal pancreatitis from tumor Toobtain a cytologic diagnosis of malignancy, EUS-guided fine-needle aspiration (FNA) iscurrently the procedure of choice If neoadjuvant chemotherapy or radiotherapy is pro-posed, a firm tissue diagnosis is imperative Since a negative FNA result does notabsolutely exclude a malignancy and since some studies have raised the possibility ofperitoneal seeding of malignant cells along the needle tract during biopsy, many centerswith an experienced surgical team forego a preoperative biopsy in potentially resectabletumors

In a patient who presents with extrahepatic biliary obstruction, a ing stricture of the intrapancreatic portion of the common bile duct, and no history ofrecurrent pancreatitis or alcohol abuse, the absence of a mass on CT or EUS should notrule out the possibility of a carcinoma of the pancreas or bile duct Similarly, negativeresults of EUS-guided FNA should not be considered definitive proof that a malignan-

malignant-appear-cy does not exist The results of EUS, with or without FNA, should be considered in thecontext of the clinical picture and as a complement to CT and ERCP findings

94 Chapter 6

of the 194 patients thought to have resectable tumors were found to have occult tic disease at laparotomy This indicates that only a small group of patients would ben-efit from laparoscopy.

metasta-The extent to which laparoscopy should be used remains controversial Laparoscopy

is reasonable to consider prior to laparotomy (during the same anesthesia induction) inpatients with biopsy-proven or suspected potentially resectable pancreatic cancer inwhom a decision has been made to proceed with pancreaticoduodenectomy However,

we do not recommend the routine use of laparoscopy as a staging procedure under a arate anesthesia induction

sep-DIAGNOSTIC APPROACH TO ASSESSSURGICALMANAGEMENT

There seems to be an overall trend that spiral CT is the most useful initial tive tool in the assessment and staging of pancreatic cancer (Figure 6-1) Although spi-ral CT can reliably predict nonresectability, there are limitations

investiga-If no mass is seen on spiral CT or in a patient who presents with extrahepatic biliaryobstruction, ERCP with biliary decompression is the second diagnostic step With amalignant-appearing stricture of the intrapancreatic portion of the common bile ductand no history of recurrent pancreatitis or alcohol abuse, the absence of a mass on CT

or EUS should not rule out the possibility of a carcinoma of the pancreas or bile duct.Concern remains about the potential for false-positive diagnosis of unresectability andthe inappropriate denial of potentially curative surgery Therefore, many experiencedpancreatic surgeons feel comfortable proceeding to resection without histology This alsoallows the use of an operative imaging modality such as laparoscopic ultrasonography inquestionable cases Local expertise and resources will determine the optimal policy

T ECHNIQUES OF P RIMARY R ESECTIONThe fundamental principle in surgery for pancreatic cancer is an en bloc resection ofthe primary tumor with adequate margins However, given its highly vascular network

as well as its close proximity to the major mesenteric and hepatic vessels, pancreaticresection is a technically challenging surgical procedure In fact, Dr Whipple’s originalseries of pancreaticoduodenectomy published in the 1930s was complicated by a mor-tality rate approaching 50% Even during the 1980s, perioperative mortality followingthe Whipple procedure remained near 20%, which, in light of the poor long-term sur-vival for patients with pancreatic cancer, prompted some prominent clinicians to adoptthe nihilistic view that the risk of pancreatic resection outweighed the potential benefits.With the improvements in perioperative anesthetic management and postoperativeintensive care, contemporary series and large epidemiologic data report acceptable mor-tality rates in the 1% to 4% range (see Table 6-2) Nevertheless, pancreatic resectionSurgical Approach to Ampullary and Pancreatic Neoplasia 95