Familial Pancreatic Cancer

Case reports of families in which multiple family members have been diagnosed with pancreatic cancer provided the first evidence that pancreatic cancer can aggregate in families. The first of these was published by MacDermott and Kramer1 in 1973 and described a family in which 4 of 6 siblings had pancreatic cancer. Many additional case reports followed.2–7 In 1990, Lynch et al8 reported a series of 18 kindreds with familial pancreatic cancer from a large familial cancer registry.

These initial studies were followed by more rigorous observational case-control and cohort studies, which are outlined in Table 1. Ghadirian et al9 found that 7.8% of all patients with pancreatic cancer and only 0.6% of controls had a family history of pancreatic cancer, a 13-fold difference, with no differences in environmental risk factors between the 2 groups. A population-based, US study (Atlanta, Detroit, and New Jersey) also reported that individuals with a first-degree relative with a pancreatic cancer had an increased risk of developing pancreatic cancer, with an odds ratio (OR) of 3.2 (95% confidence interval [CI], 1.8– 5.6). This increase in risk was higher among individuals with a first-degree relative with pancreatic cancer who had also smoked for more than 20 years (OR of 5.3 [95% CI, 2.1–13.4]) than for individuals who did not smoke or had smoked for less than 20 years (OR of 2.2 [95% CI, 1.0– 7.9]).10 While these results could represent a synergistic effect between smoking and a pancreatic cancer susceptibility gene, a portion of this increased risk may also be due to the clustering of cigarette smoking, a known risk factor of pancreatic cancer, in families.

In addition to these case-control studies, prospective cohort studies, which are not subject to recall biases, have also demonstrated an increased risk of pancreatic cancer among persons with a family history of pancreatic cancer. As part of the American Cancer Society's Cancer Prevention Study 2, Coughlin et al11 reported an increased risk of developing pancreatic cancer for individuals who reported a positive family history of pancreatic cancer at baseline, with an RR of 1.5 (95% CI, 1.1–2.1) after adjusting for age. Furthermore, a population-based cohort study demonstrated that the risk of pancreatic cancer increased 1.72-fold (95% CI, 1.13–2.54) for individuals with a parent with pancreatic cancer. The risk was not elevated when a more distant relative had been diagnosed with pancreatic cancer. Thus, both case-control and cohort studies strongly support the hypothesis that familial aggregation and genetic susceptibility play an important role in the development of pancreatic cancer. However, the relative contribution of genetic risk factors and environmental risk factors to pancreatic cancer risk that cluster within the families (ie, smoking) remains unclear.

The initial case reports and early population-based studies were followed by the establishment of family cancer registries, including the National Familial Pancreas Tumor Registry (NFPTR). The NFPTR was founded in 1994 at The Johns Hopkins Hospital (Baltimore, Maryland) as a resource for advancing our understanding of familial pancreatic cancer and for facilitating risk assessment and the early detection of neoplasms in these families. As of June 24, 2008, 2877 families have enrolled in the NFPTR. Of these, 1005 meet the established definition of familial pancreatic cancer (a parent-offspring pair, or pair of siblings with pancreatic cancer in the kindred). More detailed information about the NFPTR can be found at http://pathology.jhu.edu/pancreas/PartNFPTR.php (accessed September 11, 2008).

To determine if the aggregation of pancreatic cancer in some families is due to shared genetic effects or to shared environmental effects, we conducted complex segregation analyses. Segregation analysis is a statistical methodology aimed at determining if a major gene or genes could cause the observed familial aggregation of a disease by comparing the fit of both genetic and nongenetic models to family data. Family data from 287 patients treated at Johns Hopkins between January 1, 1994, and December 30, 1999, were included in these analyses. The results of the segregation analyses suggest an autosomal dominant genetic model of inheritance of pancreatic cancer with reduced penetrance (32% by age 85 years) and the analyses estimate a gene carrier frequency of 0.7%.12 These results suggest that the aggregation of pancreatic cancer in families is due, in part, to a yet-to-be identified gene.

Discovery of the genetic basis of inherited pancreatic cancer is an active area of research. In 2001, a multicenter linkage consortium, PACGENE, was formed to conduct linkage studies aimed at the localization and identification of pancreatic cancer susceptibility genes.13 Other groups have used linkage studies to suggest that the palladin gene (PALD) on chromosome 4q32 predisposes to pancreatic cancer14; however, this finding has not been validated in subsequent studies.15–19 

To define accurately the risk of pancreatic cancer in families, we prospectively observed families in the NFPTR. More than 838 kindreds were followed up and the prospective risk of developing pancreatic cancer was calculated by comparing the number of observed new cases of pancreatic cancer with the expected number of cases based on the US population-based Surveillance, Epidemiology, and End Results (SEER) Program data. We found that the risk of pancreatic cancer was not significantly elevated for kindreds with sporadic pancreatic cancer. The risk was, however, significantly elevated for kindreds with familial pancreatic cancer. First-degree relatives of a patient with pancreatic cancer had a 9-fold increased risk of developing pancreatic cancer themselves (standardized incidence ratio [SIR] of 9.0 [95% CI, 4.5–16.1]). The risk in kindreds with familial pancreatic cancer was elevated for individuals with 3 (SIR of 32.0 [95% CI, 10.2–74.7]), 2 (SIR of 6.4 [95% CI, 1.8–16.4]), or 1 (SIR of 4.6 [95% CI, 0.5– 16.4]) first-degree relatives with pancreatic cancer.20 Risk was not increased among 369 spouses and other genetically unrelated relatives.

The complex nature of pedigree data makes it difficult to accurately assess risk based upon the simple counting of the number of affected family members, as it does not account for family size, current age or age of onset of pancreatic cancer, and the exact relationship between affected family members. Computer-based, risk-assessment tools have been developed to integrate this complex risk factor and pedigree data into risk assessment. These models can provide more precise risk assessment than guidelines or models that rely on counts of affected family members, such as the Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (HNPCC) or myriad tables for hereditary breast and ovarian cancer.21,22 In April 2007, the first risk prediction tool for pancreatic cancer, PancPRO, was released.23 This model provides accurate risk assessment for kindreds with familial pancreatic cancer,23 and a user-friendly version is freely available as part of the CancerGene package (UTSW Medical Center, Dallas, Texas and The BayesMendel Group, Baltimore, Maryland) at http://www4.utsouthwestern.edu/breasthealth/cagene (accessed September 11, 2008). Additionally, the developer's version is also freely available at http://astor.som.jhmi.edu/BayesMendel/ (accessed September 11, 2008).

Although the genetic basis for most instances of aggregation of pancreatic cancer in families is unknown, the genes responsible for a small portion of familial pancreatic cancer are known (see Table 2). Germline mutations in the BRCA2, p16/CDKN2A, STK11, and PRSS1 genes have all been shown to increase the risk of pancreatic cancer.24–27 Additionally, some studies have described pancreatic cancers developing among individuals with HNPCC; however, the association between HNPCC syndromes and pancreatic cancer is not as well defined as it is for some of the other syndromes.28,29 

While most patients with sporadic and familial pancreatic cancer have classic infiltrating ductal (tubular) adenocarcinoma (Figure 1, A and B), some inherited syndromes are associated with a specific histologic type. Although rare, these cases provide a unique opportunity to correlate genetics with histology. For example, many of the pancreatic cancers that develop in patients with HNPCC syndrome have a medullary phenotype,30–32 and individuals with the Peutz-Jeghers syndrome appear to be predisposed to intraductal papillary mucinous neoplasms (IPMNs).33–35 These associations between phenotype and genotype are important because tumor phenotype can be used to identify at-risk families.

Hereditary breast and ovarian cancer syndrome is an autosomal, dominantly inherited disease characterized by early-onset breast and/or ovarian cancers. Germline mutations in BRCA1 and BRCA2 are responsible for the breast and ovarian cancer syndrome in most families.36 Point mutations account for most germline BRCA1 and BRCA2 mutations in these families, but germline deletions of these genes also occur.37–39 Germline deletions can be missed with standard sequencing techniques and may explain a portion of false-negative genetic tests.38 

Germline BRCA2 mutations have been clearly associated with an increased risk of pancreatic cancer. Analysis of a large series of BRCA2 mutation–positive families, ascertained for young age at onset of breast and/or ovarian cancer, demonstrated a 3.5-fold (95% CI, 1.87–6.58) increased risk of pancreatic cancer in mutation carriers. Furthermore, the probability that a patient with pancreatic cancer has a germline mutation in BRCA2 increases as the number of family members with pancreatic cancer increases. Goggins et al40 demonstrated that 7% of the patients with apparently sporadic pancreatic cancer at The Johns Hopkins Hospital had germline BRCA2 gene mutations. The probability of a germline BRCA2 mutation increases to between 6% and 12% in patients with pancreatic cancer who have at least 1 first-degree relative with pancreatic cancer,37,41 and Murphy and colleagues38 reported that germline BRCA2 gene mutations were present in 5 of 29 (17.2%) families with familial pancreatic cancer with 3 or more relatives having pancreatic cancer.

A founder mutation in BRCA2, 6174delT, carried by approximately 1.53% of individuals of Ashkenazi Jewish descent, has been reported in many families with pancreatic cancer.42 Therefore, BRCA2 gene mutations should be considered in patients with pancreatic cancer of Ashkenazi Jewish heritage, especially, but not limited to, those with a family history of early-onset breast and/or ovarian cancer.39,41 To date, mutations in the BRCA2 genes are considered the most common known genetic mutations associated with pancreatic cancer.

Germline BRCA2 mutations do not appear to be associated with a specific type of pancreatic cancer,41 as most pancreatic cancers that develop in BRCA2 carriers are traditional ductal adenocarcinomas. There are, however, significant clinical differences between BRCA2-deficient and BRCA2-intact pancreatic cancers. The BRCA2 gene is a member of the Fanconi anemia gene family and the gene product of BRCA2 functions in the repair of DNA interstrand cross-links and double-strand breaks.43 Pancreatic cancer cells with mutations in the Fanconi anemia/BRCA2 pathway are hypersensitive to DNA-interstrand cross-linking agents, such as mitomycin C, cisplatin, chlorambucil, and melphalan,44 as well as to inhibitors of poly(ADP-ribose) polymerase.45,46 Therefore, the BRCA2 gene could be a potential target for a genotype-based anticancer therapy.

Large studies of BRCA1 mutation–positive families, ascertained for young age of onset of breast and/or ovarian cancers, suggest that BRCA1 gene mutation carriers have a 2-fold increased risk of pancreatic cancer.47,48 ,BRCA1 gene mutations, however, appear to be substantially less common in families with pancreatic cancer without a significant breast cancer history,49 such that the possibility that adenocarcinoma of the pancreas is an incidental finding in BRCA1 mutation carriers cannot be ruled out.

Peutz-Jeghers syndrome (PJS) is an autosomal, dominantly inherited disease characterized by hamartomatous polyps of the gastrointestinal tract and pigmented macules of the lips and buccal mucosa.50 A variety of cancers have been associated with PJS, including gastrointestinal, gynecologic, lung, breast, and pancreatic cancer.50–54 Inherited mutations in the STK11/LKB1 gene are responsible for most cases of PJS, and as many as 80% of patients with PJS have a germline STK11/LKB1 mutation.50 

The hamartomatous polyps found in patients with PJS most commonly occur in the small intestine; however, they can also involve the stomach, colon, and rectum.50 These polyps range in size from several millimeters to more than 5 cm in diameter. Grossly, the polyps have a long stalk and the larger polyps are usually lobulated. Microscopically, Peutz-Jeghers polyps have a “Christmas tree” appearance at low power, with prominent arborizing smooth muscle (Figure 2).

Patients with PJS have a greater than 132-fold increased risk of developing pancreatic cancer.24 Of interest, these cancers may progress through an intraductal papillary mucinous neoplasm (IPMN) precursor pathway. Recently, Sato et al35 reported the findings from 2 patients with PJS who had a noninvasive IPMN of the pancreas and noted that one of the IPMNs showed bialleilic inactivation of the STK11/LKB1 gene. In addition, STK11/LKB1 gene inactivation is more frequently seen in sporadic IPMNs than it is in conventional ductal adenocarcinoma.55,56 The association of PJS with IPMN precursor lesions has significant ramifications for screening because most IPMNs are detectable with currently available imaging technologies. Indeed, Canto and colleagues57 screened asymptomatic patients with PJS for early pancreatic neoplasia by using a combination of computed tomography and endoscopic ultrasonography (EUS), and an asymptomatic IPMN was detected in 1 of the patients screened. This patient underwent surgery, and pathologic examination of the resected pancreas confirmed the presence of an IPMN with high-grade dysplasia (carcinoma in situ). These findings suggest that screening for early, curable, pancreatic neoplasia may be achievable in patients with PJS.

Hereditary pancreatitis is a rare inherited form of chronic pancreatitis characterized by repeated attacks of acute pancreatitis, usually starting early in childhood, and leading to long-term exocrine and endocrine failure.58 Germline mutations in the cationic trypsinogen gene (PRSS1) have been associated with an autosomal dominant form of hereditary pancreatitis, while germline mutations in the serine protease inhibitor gene (SPINK1) have been associated with an autosomal recessive form of hereditary pancreatitis.59 ,PRSS1 gene mutations in hereditary pancreatitis have been extensively studied. Multiple mutation sites have been identified, most of which cluster in the N-terminal half of the molecule encoded by exons 2 and 3. The most common mutations are R122H and N29I.60 

Some PRSS1 gene mutations appear to increase the stability of the trypsin protein by eliminating a trypsin autodegradation site while other PRSS1 gene mutations appear to enhance trypsinogen autoactivation, both of which eventually result in chronic pancreatitis.60–63 

Klöppel et al64,65 have carefully examined pancreatic specimens from 6 patients with hereditary pancreatitis, and they hypothesized that hereditary pancreatitis begins with necrosis of the duct-lining cells and periductal tissue, and gradually progresses to dilatation of the involved ducts, periductal fibrosis, and in advanced cases, intralobular fibrosis. Microscopically, in the early stages, the involved ducts are characterized by epithelial injury and/or necrosis and inflammatory cell infiltration. Periductal fibrosis is more prominent than intralobular fibrosis, and the pancreatic parenchyma away from the involved ducts is relatively well preserved. In the advanced stages of hereditary pancreatitis, there is extensive periductal as well as intralobular fibrosis, and the lobular parenchyma is eventually completely replaced by sclerotic tissue containing metaplastic acini and aggregates of islets of Langerhans (Figure 3). The ducts can be dilated or very irregular in shape, and some ducts contain protein plugs and calculi.

Individuals with hereditary pancreatitis have an approximately 53-fold increased risk for pancreatic cancer after the age of 50 years compared with the general population.66 Cumulative rates of pancreatic adenocarcinoma in patients with hereditary pancreatitis reach 30% to 40% by the age of 70 years.66,67 Smoking, early-onset of pancreatitis, and diabetes mellitus are associated risk factors for the development of pancreatic cancer in these patients,68 and smokers tend to develop disease 20 years before nonsmokers.67 No specific histopathologic phenotype of pancreatic cancer has been associated with hereditary pancreatitis. Instead, most patients have a classic tubular type of infiltrating ductal adenocarcinoma.

Hereditary nonpolyposis colorectal cancer syndrome (HNPCC) is an autosomal dominant hereditary disease characterized by early onset of colon cancer with a predilection for the right colon.69 Patients with HNPCC have germline mutations in genes coding for proteins associated with DNA mismatch repair. These genes include hMSH2, hMLH1, hPMS1, hPMS2, and hMSH6/GTBP.69 Adenocarcinomas of the colon in patients with HNPCC show microsatellite instability (MSI⁺) and a distinct medullary histopathology.70 Women who carry mutations in these genes are also at a very high risk for endometrial cancer, greater than 50% by age 70 years.71 In addition, patients with HNPCC are at increased risk for a spectrum of extracolonic neoplasms, including carcinomas of the endometrium, ovary, stomach, bile duct, kidney, bladder, ureter, and skin.69 

While some studies have suggested individuals with HNPCC may also have an increased risk for pancreatic cancer,30,72 additional studies are needed to accurately quantify this risk. Lynch et al72 first reported pancreatic carcinoma in kindreds with HNPCC. Further evidence linking HNPCC and pancreatic cancer comes from a study of medullary carcinomas of the pancreas by Wilentz et al.30 Three of the 18 patients with a medullary cancer of the pancreas reported colon cancer in a first-degree relative.30 In addition, 1 of the patients in this study had a synchronous MSI⁺ pancreatic and colonic cancer. This study was supported by a report of medullary carcinoma of the pancreas that developed in an individual with a MSI⁺ tumor who had a germline mutation in the MSH2 gene.31 

The pancreatic cancers that arise in patients with HNPCC often have a distinctive medullary appearance. Medullary carcinoma of the pancreas is a rare variant of pancreatic adenocarcinoma. As with medullary carcinoma of the colon, it is associated with a better prognosis than conventional ductal adenocarcinoma.29,30 Grossly, medullary carcinomas tend to form well-circumscribed soft masses. Microscopically, medullary carcinomas are poorly differentiated, and they have a syncytial growth pattern with pushing borders (Figure 4, A and B). The infiltration of lymphocytes into the carcinoma can be very prominent. Therefore, the morphology of pancreatic medullary carcinoma is very similar to that of medullary carcinoma of the colon.

Unlike conventional ductal adenocarcinoma of the pancreas, most medullary carcinomas do not harbor KRAS2 gene mutations. Instead, medullary carcinomas of the pancreas often harbor BRAF gene mutations and are MSI⁺.30,32,73 As one would expect in a neoplasm with genetic inactivation of a DNA mismatch repair gene, medullary carcinomas of the pancreas often show loss of expression of one of the DNA mismatch repair proteins (Mlh1 and Msh2). For example, a medullary carcinoma of the pancreas has been recently reported for a patient with HNPCC due to a mutation of the hMSH2 mismatch repair gene.31 

The presence of medullary phenotype in a pancreatic cancer may suggest inherited susceptibility to HNPCC. Indeed, patients with medullary carcinoma of the pancreas are more likely to have a family history of cancer in first-degree relatives.30 

Taken together, these observations suggest a paradigm for the evaluation of patients with a medullary carcinoma of the pancreas. If clinically appropriate, medullary carcinomas of the pancreas can be tested for microsatellite instability. If the carcinoma is MSI⁺, and after appropriate genetic counseling, patients with a microsatellite-unstable pancreatic cancer can then be genetically tested for germline mutations in one of the DNA mismatch repair genes.

The classification of a neoplasm as a medullary carcinoma of the pancreas has at least 3 important clinical ramifications. First, the medullary histology is associated with a better prognosis.29,30,74 In 1 series, patients with surgically resected microsatellite-unstable pancreatic cancers lived a mean of 62 months compared with 10 months for patients with conventional ductal adenocarcinomas.29,30,74 Second, patients with medullary carcinoma of the pancreas are more likely to have a family history of cancer.30 Third, as with medullary carcinoma of the colorectum, medullary carcinoma of the pancreas could serve as a predictor of poor response to certain adjuvant chemotherapies such as 5-fluorouracil.75 

Familial atypical multiple mole melanoma (FAMMM) is an autosomal dominant inherited syndrome with incomplete penetrance. It is characterized by greater-than-normal numbers of melanocytic nevi, multiple atypical melanocytic nevi, and an increased risk of cutaneous malignant melanoma.76,77 Germline mutations in the p16/ CDKN2A gene are responsible for a portion of FAMMM cases.25,77,78 A variety of cancers, other than melanoma, have been documented in kindreds with familial melanoma, including carcinoma of the lung, pancreas, and breast as well as sarcoma.79,80 

A subset of FAMMMs is associated with pancreatic cancer. Kindreds with FAMMM have a 13- to 22-fold increased risk for pancreatic cancer80 and the risk for pancreatic cancer among mutation carriers is 38-fold higher than that of the general population.81 Lynch and colleagues79 studied 159 families with familial pancreatic carcinoma and identified 19 families with FAMMM. DNA testing of the 8 “best” of these kindreds with FAMMM– pancreatic carcinoma (FAMMM-PC) revealed a germline p16/CDKN2A gene mutation in every case. Researchers in a Dutch study conducted mutation analyses of the p16/ CDKN2A gene in 27 families with FAMMM and demonstrated that 19 of the 27 families harbored a 19-bp deletion in exon 2 of the p16/CDKN2A gene (p16-Leiden founder mutation). In this particular subset of families with FAMMM, the estimated cumulative risk for the development of pancreatic cancer in the mutation carriers was 17% by the age of 75 years.82 This suggests a strong link between FAMMM-PC and p16-Leiden.

Although many reports have described the association of pancreatic carcinoma with FAMMM, no unique histopathologic features have been reported for the pancreatic cancers that develop in patients with FAMMM.

The association between FAMMM and pancreatic cancer suggests that a good family history of melanoma should be taken in patients with pancreatic cancer.

Familial adenomatous polyposis (FAP) is an autosomal, dominantly inherited disorder characterized by the development of hundreds to thousands of colonic adenomatous polyps at an early age. Some of the adenomas can progress to invasive adenocarcinoma, and, if untreated, invasive adenocarcinoma of the colon will develop in almost all patients by the age of 40 years.69 Germline mutations in adenomatous polyposis coli (APC) gene, a tumor suppressor gene, are responsible for the development of FAP.83,84 Patients with FAP are at increased risk for other neoplasms, including thyroid tumors, gastric, duodenal, and ampullary adenocarcinoma.

Although the association of pancreatic cancer and FAP is not as strong as the association of FAP with other cancer types, several lines of evidence suggest that patients with FAP are also at increased risk for the development of pancreatic neoplasms. Pancreatic adenocarinoma has been described in individuals with germline APC gene mutations, and patients with FAP may have a 4-fold increase in risk for pancreatic adenocarcinoma.85 Furthermore, in a report of a patient with FAP and an IPMN with high-grade dysplasia (in situ carcinoma), it was noted that the IPMN showed biallelic inactivation of the APC gene, a fact that supports the genetic link between FAP and IPMN for this patient.86 

In addition to the association of FAP with pancreatic adenocarcinoma, Abraham and colleagues28 reported a rare pancreatic neoplasm, pancreatoblastoma, arising in a patient with FAP. Pancreatoblastoma is a malignant epithelial neoplasm with acinar differentiation and squamoid nests (Figure 5). These neoplasms may also have components with ductal, endocrine, and mesenchymal differentiation. They are most commonly seen in infants and young children, although cases in adults have also been reported.87 In contrast to conventional ductal adenocarcinoma of the pancreas, both sporadic and FAP-associated pancreatoblastomas lack KRAS2 and TP53 gene mutations. Instead, most cases harbor alterations in the APC/ β-catenin pathway.28 In addition, Abraham et al have reported biallelic APC gene mutations in this FAP-associated pancreatoblastoma.

Taken together, these results suggest that there is a genetic link between FAP and pancreatoblastoma, and pancreatoblastoma might represent one of the extracolonic manifestations of FAP. It should be noted that the Beckwith-Wiedemann syndrome has also been associated with pancreatoblastoma in newborns.

Screening the general population for early pancreatic neoplasia may not be practical, as even a test with very high sensitivity and specificity will have a low positive predictive value. This is because the incidence of pancreatic cancer is low in the general population. Approximately 9 in 100 000 Americans develop pancreatic cancer each year.88 Screening of high-risk populations, such as those with a strong family history of pancreatic cancer or a known genetic syndrome may, however, be possible. A screening program for high-risk individuals was recently established at The Johns Hopkins Hospital.57,89 Individuals with a strong family history of pancreatic cancer (at least 3 close relatives with pancreatic cancer) or with the Peutz-Jeghers syndrome were screened by using a combination of EUS and computed tomography, and in some cases, endoscopic retrograde cholangiopancreatography. One hundred sixteen patients were screened as part of a study, called “Cancer of the Pancreatic Screening Study (CAPS).” Surgery was recommended for patients with a significant focal lesion detected by imaging studies and/or atypical epithelial cells in EUS-guided fine-needle aspirates from the pancreas.

Ten of the individuals in this study underwent surgical resection at The Johns Hopkins Hospital. These resected pancreata provide a unique opportunity to study early familial pancreatic neoplasia.90 Three features dramatically stood out in these familial cases compared with sporadic cancers. First, most of the pancreata harbored multifocal precursor lesions, involving as many as 27% of the duct profiles (Figure 6, A through D). The multifocality of the early lesions in these pancreata was confirmed by using genetic analyses for KRAS2 gene mutations. Second, in some cases these lesions were so numerous that they could even be appreciated grossly. Figure 7 is a representative example showing multifocal lesions grossly. It shows 2 dilated cystic lesions filled with abundant thick mucin, which are 2 IPMNs, as well as multiple thickened small ducts with surrounding white firm areas, which are pancreatic intraepithelial neoplasia (PanIN) with lobulocentric atrophy and fibrosis. The third feature that stands out in these cases is that the precursor lesions are often associated with lobulocentric atrophy (Figure 8). Lobulocentric atrophy, as the name suggests, is characterized by loss of acinar parenchyma in a lobular pattern, fibrosis, and acinar to ductal metaplasia (Figure 6, B and C). In some cases, this periductal fibrosis associated with PanINs and IPMNs can be appreciated grossly, as demonstrated in Figure 9. The degree of the parenchymal changes associated with PanINs is variable, ranging from partial acinar atrophy with focal acinar ductal metaplasia to a complete loss of acinar cells and complete replacement of the lobular unit by acinar ductal metaplasia, fibrotic stroma, and aggregates of islets of Langerhans.

The finding of multifocal PanINs in patients with a strong family history of pancreatic cancer provides a rational basis for screening at-risk patients. The multifocal intraductal precursor lesions produce a pattern of parenchymal atrophy and fibrosis in a background of normal pancreas. This heterogeneity can be detected by EUS as chronic pancreatitis-like changes.57,89,90 In fact, Brune et al90 observed that an increased degree of heterogeneity of the pancreatic parenchyma observed on EUS was significantly correlated with an increase in percentage of ducts involved with PanIN lesions.

A portion of pancreatic cancer has an inherited genetic basis. Most inherited genetic changes responsible for the aggregation of pancreatic cancer are unknown. Some genetic causes of familial pancreatic cancer are known, and some of these genetic changes are associated with pancreatic cancers with a distinctive microscopic appearance, such that the diagnostic findings at the microscopic level may increase the index of suspicion for a genetic syndrome. In addition, careful study of surgically resected pancreata from patients with a strong family history of pancreatic cancer has shown that multifocal PanINs with lobulocentric atrophy develop in many of these patients. These changes give the pancreas a heterogenous appearance that can be detected by EUS, suggesting that at-risk individuals can be screened and early curable neoplasms detected before they progress to invasive cancer. Further elucidation of the inherited genetic basis may provide greater insight into the microscopic differences between familial and sporadic pancreatic cancer cancers, as precancerous changes in the pancreas. This in turn will aid early-detection screening efforts.

This work was supported by Cigarette Restitution Fund of Maryland, The Sol Goldman Pancreatic Cancer Research Center, and an NCI SPORE grant in Gastrointestinal Cancer CA62924.