A diverse array of highly penetrant hereditary cancer syndromes predispose individuals to increased risks of pancreatic cancer, including BRCA1
- and BRCA2
-associated hereditary breast/ovarian cancer, Lynch syndrome (caused by germline DNA mismatch repair [MMR] gene mutations), Peutz–Jeghers syndrome (caused by germline STK11
mutations), Li–Fraumeni syndrome (caused by germline TP53
mutations), and familial atypical multiple mole/melanoma syndrome (caused by germline CDKN2A
alterations have also been linked to risks of pancreatic cancer, although the magnitude of risk remains poorly understood.1
Because these syndromes confer elevated risks of various cancers beyond pancreatic cancer, many of which have evidence-based methods for screening and risk reduction, identifying families with inherited mutations in these genes can be a powerful means of facilitating genetically-driven cancer prevention. In addition, there may soon be significant implications of identifying such mutations for pancreatic cancer probands themselves, in light of promising early-phase data with poly (ADP-ribose) polymerase (PARP) inhibitors in small cohorts of patients with pancreatic cancer with germline mutations in BRCA1
Research to date, however, has provided insufficient guidance on which patients with pancreatic cancer should undergo hereditary cancer risk assessment. Recent data4-10
have shown that germline mutations in these well-described cancer susceptibility genes can be found in patients with pancreatic cancer who lack obvious clinical features of inherited cancer risk (eg, young age at diagnosis, classic family history patterns), suggesting that a number of mutation carriers may go undiagnosed in current practice.
Furthermore, existing guidelines from the National Comprehensive Cancer Network11
only provide recommendations on which patients with pancreatic cancer warrant germline testing in the context of suspected hereditary breast/ovarian cancer. There are no guidelines to help clinicians determine which patients with pancreatic cancer may benefit from other forms of germline evaluation, including multigene testing.
In the article accompanying this editorial, Shindo et al12
provide essential insight into the prevalence and spectrum of inherited factors in pancreatic cancer. They performed germline sequencing of 32 genes in 854 individuals who underwent pancreatic cancer resection at Johns Hopkins Hospital. In total, they found pathogenic or likely pathogenic germline alterations in 3.9% of patients with pancreatic cancer in their cohort.12
Of the 33 patients with a germline mutation in their study, 29 (88% of carriers; 3.4% of the overall cohort)12
carried alterations in genes that are theoretically targetable with agents such as PARP inhibitors (BRCA1
, and PALB2
) or immune checkpoint inhibitors (MLH1
Although it was a single-institution study12
limited to individuals undergoing pancreatic resection, this represents by far the largest analysis to date examining systematic germline testing in patients with pancreatic cancer who, importantly, were not preselected by age, personal/family cancer history, or other high-risk features. Interestingly, the 3.9% mutation prevalence identified in this study12
is on the low end of the 3.8% to 11.3% range observed in prior, smaller studies4,7,10
of multigene germline testing in patients with pancreatic cancer who were not preselected by age or clinical history. This wide range may be due in part to differences in the specific panel of genes assessed in these studies, as well as the identification of founder mutations, the frequency of which may vary markedly5
between populations from differing geographic and ethnic backgrounds.
Despite these differing prevalence estimates, the data from Shindo et al12
confirm and expand on a number of critically important issues related to hereditary cancer risk assessment in pancreatic cancer. As has been suggested by earlier studies that examined more limited panels of cancer susceptibility genes,4,5,7,9,10
found that nearly all (30 of 33; 91%) of the germline alterations identified in patients with pancreatic cancer were in genes (BRCA1
, MMR genes, and TP53
) for which current guidelines1,11,16
recommend specific forms of enhanced screening and other interventions to reduce the risk of associated cancers in healthy carriers. For example, salpingo-oophorectomy is associated with improved overall survival in women with germline BRCA1
whereas early and frequent colonoscopic screening significantly reduces colorectal cancer–associated mortality in individuals with Lynch syndrome.18
Thus, identifying patients with pancreatic cancer with these mutations has tremendous potential to facilitate cancer prevention in their at-risk family members. Shindo et al12
also found that a significant majority of patients with pancreatic cancer with germline mutations lacked personal and family histories suggestive of inherited risk, and only 12% of carriers were diagnosed with pancreatic cancer before the age of 50 years, corroborating similar findings from prior, smaller studies.4-6,8-10
illustrate both the diverse spectrum of potentially actionable germline mutations that occur in patients with pancreatic cancer, as well as the substantial limitations that exist in actually identifying such mutation carriers in contemporary practice. In particular, Shindo et al12
have demonstrated that syndrome-specific germline testing is inadequate at evaluating patients with pancreatic cancer for inherited mutations. This finding thereby supports their assertion that multigene testing (including BRCA1
, MMR genes, ATM
, and CDKN2A
) should instead be the standard approach when performing germline evaluation in patients with pancreatic cancer. Furthermore, given the inability of current clinical guidelines11
to reliably identify patients with pancreatic cancer with these mutations, Shindo et al conclude12
that all individuals with pancreatic cancer should be offered germline testing, regardless of their personal or family history of cancer and age at diagnosis.
Albeit provocative at first glance, the notion of systematically screening all cases of a given cancer type to identify the minority of individuals with inherited risk has precedent in other tumor types. One particularly relevant parallel is in ovarian cancer where, since 2007, National Comprehensive Cancer Network guidelines11
have recommended germline BRCA1
evaluation for all patients. A key factor underlying this recommendation has been the recognition19,20
that a large fraction of patients with ovarian cancer with germline BRCA1
mutations lack suggestive clinical features (eg, young age, family history of BRCA
-associated cancers), mirroring what has now been found in pancreatic cancer.4-6,8,9,12
Although the 3.9% prevalence of mutations observed by Shindo et al12
in patients with pancreatic cancer is markedly lower than the 15% to 18% BRCA1/2
mutation prevalence in ovarian cancer,19-21
these mutations have profound importance for cancer prevention in at-risk family members and occur with sufficient frequency in pancreatic cancer that novel approaches to comprehensively identify carriers are warranted. Another relevant scenario, in which systematic risk assessment is endorsed despite a low mutation prevalence, is in colorectal cancer, where microsatellite instability and/or MMR deficiency tumor testing are standard for all patients to screen for Lynch syndrome, even though only 3% of cases harbor germline MMR mutations.16,22,23
The argument for universal genetic testing in pancreatic cancer may become all the more compelling if germline status is confirmed to predict for benefit from novel therapeutics (eg, PARP inhibitors) in patients with pancreatic cancer with mutations in BRCA1
, and other genes.2,3,14
The data from Shindo et al12
demonstrate the intriguing potential benefits of systematic germline testing in pancreatic cancer. However, there remains a critical need to rigorously explore the question of how to implement widespread genetic risk assessment in the context of modern clinical oncology,24
particularly for a disease as lethal and medically complex as pancreatic cancer. Even in ovarian cancer, for which national guidelines have recommended universal germline testing for the past decade, < 20% of patients are referred for genetic evaluation in current practice.25-28
These rates are likely to be even lower in patients with pancreatic cancer because of the multitude of symptomatic, endoscopic, and therapeutic needs inherent to the disease. As such, it is unrealistic to expect that many patients with pancreatic cancer would be able to undergo germline testing within the traditional model of genetic medicine, where patients attend dedicated consultations for in-depth pre- and post-test genetic counseling. Furthermore, many cancer genetics practices routinely have wait times of several months when scheduling an initial consultation. This inherently limits access to genetic testing for many patients with pancreatic cancer, whose median survival is < 1 year in the metastatic setting.29,30
Novel care delivery models (such as having genetic counselors embedded within oncology workflows to actively seek out patients during clinical care [eg, during chemotherapy]) have yielded substantially improved (> 85%) rates of genetic testing in single-institution experiences with ovarian cancer.31,32
These are appealing paradigms to consider for systematic risk assessment in pancreatic cancer. The resource-intensive nature of such programs, however, including the reliance on prompt in-person access to genetic counselors, is likely to preclude these strategies from being exportable to care settings outside of large academic centers.
Another important factor in considering widespread germline testing in pancreatic cancer needs to be the cost of such processes, even though genetic sequencing costs have fallen dramatically in recent years. Paralleling experiences from other forms of universal genetic screening (BRCA1/2
testing in ovarian cancer25
; Lynch syndrome tumor testing33
in colorectal cancer), the cost effectiveness of systematic germline testing in pancreatic cancer will probably be driven predominantly by whether probands and their families actually act upon the identification of inherited mutations. In a recent study7
from the Australian Pancreatic Cancer Genome Initiative, 5.1% of 392 patients with pancreatic cancer were found to carry germline alterations, yet only 35% of these carriers (1.8% of the cohort) had family members who actually sought genetic counseling, and only two probands (0.5% of the cohort) received systemic treatment tailored to their germline mutation. To justify the cost, personnel, and infrastructure that would be required to implement universal germline testing in pancreatic cancer, considerable improvements need to be made in maximizing the uptake of cascade genetic testing for probands’ family members.
In conclusion, the data from Shindo et al12
demonstrate the considerable limitations of existing genetic testing strategies in pancreatic cancer and make a persuasive case for the potential benefits of offering germline testing to all such patients. If ongoing studies confirm that PARP inhibitors are indeed an effective therapy for pancreatic cancers in patients with germline alterations in BRCA1
, and other genes, then universal genetic testing will undoubtedly become an essential part of pancreatic cancer care. Translating the theoretical benefits of systematic germline testing into improved outcomes for patients with pancreatic cancer and their families, however, will require careful consideration of the complexities inherent to efficiently delivering genetic medicine on a large scale, within diverse practice settings, to medically challenging patients. The issue of how to effectively implement systematic genetic testing in pancreatic cancer may prove to be even more challenging than the question of whether such universal testing is beneficial in the first place.
Supported by the Dana-Farber/Harvard Cancer Center Specialized Program of Research Excellence (SPORE) in Gastrointestinal Cancer (P50 CA127003) Developmental Research Project Award and the Dana-Farber Cancer Institute Department of Medical Oncology Translational Research Grant.