Pancreatic ductal adenocarcinoma (PDAC), commonly known as PC, is currently the third leading cause of cancer-related deaths in the United States and has a 5-year relative survival rate of 8–9%, the lowest of any malignancy [1
]. In 2018, approximately 55,440 new cases of PC will be diagnosed and 44,330 patients will die from PC [1
]. Unfortunately, 80–85% of patients are diagnosed with late stage, inoperable disease because symptoms did not appear until the disease had metastasized [5
]. Symptoms include jaundice, abdominal pain, and weight loss, whereas earlier stages are typically asymptomatic [6
]. To improve PC survival, there is a dire need to develop screening approaches to detect early stage, operable malignancies. Given that the lifetime risk for developing PC is only 1.5% [7
], screening the general population is impractical; however, selectively screening individuals at increased risk for PC can enable detection of early-stage malignancies and even pre-malignant lesions and prolong survival [8
1.1. Risk Factors for PC
It is estimated that approximately 25% of PDAC cases can be attributed to environmental and lifestyle risk factors [9
]. In addition to age, the most established risk factors for PC are tobacco exposure, heavy alcohol use (>60 mL ethanol/day), and a personal history of obesity, pancreatitis, and/or diabetes [9
]. Approximately 10% of PC cases develop because of a familial or hereditary predisposition, which place them at a 1.8- to 132-fold higher risk than individuals in the general population [11
]. Thus, genetically high-risk individuals (HRI) are a prime population for early detection efforts [11
]. HRI can be defined as having “familial PC” when they have two or more first degree relatives (FDR) with the disease or have one FDR and at least two affected second degree relatives (SDR) [7
]. The lifetime risk for developing PC increases with the number of FDR with the disease, and spans from 3% with one affected FDR, 8–12% with two affected FDR, and up to a 40% lifetime risk with three affected FDR [7
]. As summarized in Table 1
, deleterious mutations in genes associated with a hereditary cancer syndrome or chronic inflammation of the pancreas also pose significant risks for PC development [16
]. Of PDACs associated with hereditary PC syndromes, most (around 5–19%) are attributed to mutations in BRCA2
] which confer a 3.5–10-fold increase in risk [20
1.2. Pre-Malignant PC Precursors Exist and Can Be Detected via Imaging
It is now established that PC can develop from three main precursor lesions: pancreatic intraepithelial neoplasia (PanIN), intraductal papillary mucinous neoplasms (IPMN), and mucinous cystic neoplasms (MCN) [32
]. While PanINs are microscopic and are typically only viewed pathologically, IPMNs and MCNs are macroscopic lesions that can be detected via radiologic imaging with endoscopic ultrasound (EUS), magnetic resonance imaging (MRI), or computed tomography (CT) [33
]. Most IPMNs and MCNs are detected incidentally among individuals in the general population who undergo cross-sectional imaging for reasons unrelated to their pancreas and are often thought to be ‘sporadic’ or non-familial. However, PC precursors are also commonly detected in familial kindreds and as part of hereditary cancer syndromes (Table 1
]. For example, in a small Israeli study involving 51 patients with IPMNs who underwent genetic testing, 25% of cases with a family history of PC were found to have BRCA2
]. In a study of 79 P16
carriers who underwent MRIs in the Netherlands, 11% (n
= 7) primarily had side branch duct IPMNs detected [25
]. Most IPMNs with side branch duct involvement have a lower risk of malignant transformation than main pancreatic duct IPMNs, with mean frequencies of high-grade disease or invasive carcinoma of 31% and 62%, respectively [40
]. Those with main pancreatic duct involvement are typically recommended for surgical removal because of their significant malignant potential, whereas side-branch duct IPMNs are typically observed unless they occur with concerning radiologic features [14
Based on recommendations from the International Cancer of the Pancreas Screening (CAPS) consortium, HRI with a 5–10% lifetime risk for developing PC should be offered screening for pancreatic masses as part of a research study using EUS, MRI, or both. CT is not opportune for screening unaffected individuals because of exposure to radiation [7
]. Current CAPS guidelines recommend screening those with at least two affected FDRs; patients with Peutz-Jeghers syndrome (PJS); and P16
and hereditary non-polyposis colorectal cancer (HNPCC) mutation carriers with ≥1 affected FDR [7
]. The proper age for screening, however, remains a topic of debate. Although experts from CAPS do not make specific recommendations regarding the proper age to initiate screening, some studies have recommended screening at 55 years of age or 10 years younger than the closest relative with PDAC [7
], while others have shown diagnostic yield is highest in individuals >65 years [7
]. It is noteworthy that HRI with PJS are prone to developing PDAC at a much younger age and require earlier screening [16
The objective of the current study was to describe characteristics and clinical outcomes of cohorts of HRI identified at H. Lee Moffitt Cancer Center and Research Institute (Moffitt), the only National Cancer Institute-designated Comprehensive Cancer Center based in the state of Florida. Specifically, we set out to determine the uptake of screening, the prevalence and characteristics of solid and cystic lesions detected via screening or as incidental findings, and the age at which lesions are detected.
Advancements in the early detection and prevention of pancreatic cancer requires strategies to detect and treat pre-malignant lesions and early invasive disease among individuals at high risk to develop this malignancy. We performed a descriptive analysis of the characteristics and clinical outcomes of three groups of high risk individuals (HRI) identified at our cancer center in order to better understand the uptake of screening, the prevalence and characteristics of solid and cystic pancreatic lesions detected via screening or as incidental findings, and the age at which lesions were detected.
As expected, individuals meeting the 5–10% lifetime risk of PC recommended by CAPS had the highest uptake of screening, primarily with a previously-described EUS-only High-risk Surveillance protocol [6
]. Nearly 40% of the patients meeting CAPS criteria with abdominal imaging on file were found to have pancreatic lesions, primarily unclassifiable cysts due to their small size, which is consistent with findings from other studies [46
]. This suggests that proactive screening efforts for high risk individuals are effective in identifying a high prevalence of smaller pancreatic lesions that should be monitored. Of the 22 individuals for whom we do not have a report of abdominal imaging on file, 18 were at least 50 years old and therefore should be recommended for a screening protocol. Moreover, 11 of these individuals without a report of pancreas screening on file harbored a deleterious BRCA1
mutation and a family history of pancreatic cancer, including one individual with six second degree relatives with the disease. It is possible that these HRI received imaging elsewhere, but our observation highlights the importance of continuity of care between cancer genetics and gastrointestinal oncology clinics so that HRI are made aware of opportunities for screening studies.
Among those at increased moderate genetic risk for PC who did not meet the CAPS guidelines threshold for screening and therefore were not part of a surveillance study at our institution, we identified 10 individuals who had abdominal imaging on file and for whom cystic and solid pancreatic lesions were detected as part of a diagnostic work-up. This is likely an underestimate since most of these patients did not have abdominal imaging on file. We observed an earlier mean age of lesion diagnosis (60.7 years) in this group of HRI compared to those that met CAPs guidelines (62.61 years), and also observed a significantly larger lesion size (1.95 cm, p < 0.0001). Although this cohort is small, the earlier age, the larger size of lesions detected, and the detection of one invasive tumor and several high-grade IPMNs seems relevant when considering expanding current screening guidelines to include patients at an increased risk for disease that are on the cusp of reaching current CAPS guidelines. Moreover, a personalized approach that also examines individual patient characteristics such as hereditary cancer disorders, medical conditions such as pancreatitis, diabetes and obesity, and environmental factors such as smoking, and drinking histories can help to determine whether these people are at a 5% lifetime risk and would benefit from screening.
We also reported on 66 individuals not known to have a familial or hereditary predisposition to PC who presented with incidental findings of pre-malignant pancreatic cysts. This group of “sporadic” HRI had a later age at diagnosis, which may have led to larger lesions than the other two groups, in line with other studies [47
]. Moreover, this group had the highest percentage (45.5%) of patients requiring surgical resection who had significant pathology. The National Comprehensive Cancer Network (version 1.2018) recently started recommending that genetic counseling and germline testing should be considered for patients with a personal or family history of PC or those with a clinical suspicion of inherited susceptibility. Data from a recent case-control study support this recommendation, as germline mutations in one of six genes (CDKN2A
, and BRCA1
) were identified in 5.5% of all PC patients, including 7.9% of patients with a family history of PC and 5.2% of patients without a family history of PC [49
]. Taken together, this group with sporadic findings of pre-malignant pancreatic cysts may also represent a population for whom germline genetic testing for cancer predisposition genes should be offered.
Our study is limited because it is based on a retrospective cohort of individuals from a single institution. Despite this, our cohort is unique because it includes patients from two protocols that recruited participants based on their family history and/or presence of a known germline mutation and another protocol that recruited participants unselected for family history or genetic profile. In doing so, we included a diverse cohort of HRI from our comprehensive cancer center. Additionally, it is very possible that a subset of HRI from this cohort has been followed for their medical management or received genetic counseling and testing at other centers and that our information is incomplete. Thus, further follow-up of this cohort is warranted to update their medical and screening history. Moreover, results of this follow-up could inform a prospective study that could serve to verify this retrospective review. Despite these limitations, our study identifies a number of areas in which we can work to optimize guidelines for who is offered genetic counseling and testing and screening. Large multi-center trials are needed to validate our findings and better define the optimal screening regimen for various groups of HRI, with an overarching goal of promoting the detection and treatment of pre-malignant lesions and early, operable disease.
Data was gathered from three IRB-approved protocols based at Moffitt Cancer Center: the Inherited Cancer Registry (ICARE) Initiative (MCC 12347, PI: C. Laronga), the High Risk Surveillance Protocol (MCC 14882, PI: J. Klapman) [6
], and the Florida Pancreas Collaborative (MCC 18336, PI: J. Permuth) [50
]. The ICARE initiative aims to establish a registry of individuals interested in participating in studies of the genetic causes of cancer and seeks to evaluate the roles of genetic and environmental risk factors in the development of tumors or related conditions. Eligible individuals joined ICARE between July 2010 and March 2017 at Moffitt Cancer Center. The High Risk Surveillance Protocol provides annual EUS (with fine needle aspiration if possible) for individuals meeting CAPs guidelines and took place between June 2007 and December 2017 [7
]. To be eligible, individuals must have two or more relatives with PDAC with at least one FDR affected, and be at least 40 years old or 10 years younger than the youngest affected family member, have Peutz-Jeghers Syndrome and be at least 30 years old, have hereditary pancreatitis, have familial atypical multiple mole melanoma syndrome, or have a BRCA2
mutation and at least one FDR or SDR with documented PC [6
]. The Florida Pancreas Collaborative is an ongoing biorepository that was established in September of 2015 and includes individuals newly-diagnosed with a breadth of pancreatic conditions ranging from early and late-stage PDAC; benign, pre-malignant, and malignant pancreatic cysts; and pancreatitis [50
Eligible HRI were identified from the aforementioned protocols and data was collected from the electronic medical record on variables including: age, gender, race/ethnicity, abdominal imaging modalities used and at what age, type of lesion detected, age at which lesion was detected, surgical pathology, and vital status. Data was also collected on personal medical history, body mass index, and history of tobacco exposure and alcohol use. When recording the type of lesion detected, the diagnosis was abstracted from the first imaging report on which the lesion was found or an ambulatory care note. Comparisons were made across different at-risk groups using chi-squared test and Fisher’s exact test (where the cell size was <5) for categorical variables and ANOVA for continuous variables.