Screening for Pancreatic Cancer

Pancreatic ductal adenocarcinoma is the third leading cause of cancer death in the United States, accounting for approximately 45 750 deaths per year,¹ and is projected to become the second leading cause of cancer death by 2020.² The vast majority of patients with pancreatic cancer are diagnosed at a late stage of disease, when treatments are palliative. Only 10% of patients diagnosed with pancreatic cancer undergo potentially curative surgery,³ and 5-year survival across all stages is approximately 9%.¹ Early diagnosis and detection of precursor lesions is therefore an attractive approach to decrease pancreatic cancer morbidity and mortality.

In this issue of JAMA, the US Preventive Services Task Force (USPSTF) presents a recommendation statement⁴ that reaffirms its 2004 recommendation against screening for pancreatic cancer in asymptomatic individuals.⁵ The updated D recommendation is based on a comprehensive evidence report and systematic review of literature through April 2018.⁶ Included in this review were clinical trials, intervention studies, cohort studies, and studies of diagnostic accuracy; notably, studies that focused exclusively on individuals with known genetic cancer syndromes were excluded. The USPSTF found “no evidence that screening for pancreatic cancer or treatment of screen-detected pancreatic cancer improves disease-specific morbidity or mortality, or all-cause mortality.”⁴ Little to no benefit was thought to be derived from screening, and there was sufficient evidence that individuals undergoing screening could experience at least moderate harm.

Even without this comprehensive evaluation of existing literature, it could be anticipated that screening for pancreatic cancer in asymptomatic, average-risk individuals would generate more harm than good. Approximately 1.6% of individuals in the United States will develop pancreatic cancer during their lifetime.¹ With this relatively low prevalence, even an ideal screening test with 99% sensitivity and 99% specificity would yield 1000 false-positive results if applied to 100 000 patients. These false-positive results would require subsequent diagnostic evaluation and accrue additional complications, costs, and patient distress that would cause the risks of screening to outweigh any potential benefit.

However, certain high-risk individuals with greater than 5% lifetime risk of pancreatic cancer, or a 5-fold increased relative risk, may derive benefit from surveillance. These include individuals from familial pancreatic cancer (FPC) kindreds (individuals with 2 or more blood relatives affected with pancreatic cancer, with at least 1 affected first-degree relative), whose risk increases with a greater number of affected family members or if those relatives developed cancer at a younger age.^7-9 Germline mutations in ATM, BRCA1, BRCA2, CDKN2A, PALB2, PRSS1, STK11, TP53, and the Lynch syndrome mismatch repair genes are also associated with an increased risk of pancreatic cancer.^10,11 Another population of at-risk individuals are those with a recent diagnosis of diabetes. Emerging epidemiologic data demonstrate that individuals with new-onset diabetes have up to an 8-fold greater risk of pancreatic cancer. Diabetes in this setting is distinct from type 1 and type 2 diabetes and is pancreatic in origin, which may result from a paraneoplastic process related to the pancreatic tumor.¹²

Additional considerations can help further define the ideal timing of initiation and subsequent intervals of surveillance in those individuals at heightened risk for pancreatic cancer. For example, even in families with multiple affected blood relatives, the mean patient age at the time of pancreatic cancer diagnosis is 68 years,⁸ which may inform the starting age for surveillance of at-risk individuals. Recent studies suggest that new-onset diabetes can predate the diagnosis of pancreatic cancer by up to 36 months, a finding that may define the interval in which this population may benefit from intensive surveillance.¹² Furthermore, even among high-risk individuals, competing all-cause mortality can exceed pancreatic cancer mortality as age and comorbidities increase. Clear parameters for initiation and completion of surveillance are critical to avoid false-positive results and patient harm and warrant additional investigation.

The reaffirmation of the D recommendation should by no means mitigate the significant advances made over the last decade in identifying those individuals at sufficient risk of developing pancreatic cancer to possibly benefit from surveillance. A recent study of a US screening cohort of 354 high-risk individuals (defined by genetic factors or family history) reported a survival benefit in those individuals with pancreatic cancers and high-grade precursor neoplasms detected during surveillance.¹³ The median survival time for the 20 high-risk individuals diagnosed with pancreatic neoplasia during surveillance was 5.3 years (interquartile range, 1.2-11.1 years), compared with 1.4 years (interquartile range, 0.4-3.5 years) for those who did not undergo the recommended surveillance. In addition, the overall 3-year survival among patients with pancreatic cancer was 85% for those undergoing surveillance, compared with 25% for those with cancer detected outside surveillance. Additional data have emerged from recent international collaborations on the outcomes of high-risk individuals who have undergone surveillance and surgery for both pancreatic cancer and precursor lesions.^14,15 These studies were published after the USPSTF review and extend beyond the task force’s population of interest, which includes “adults with screen-detected, asymptomatic, or incidentally detected pancreatic adenocarcinoma”³ and not those with high-risk precursor lesions, which is the stated goal of existing surveillance programs in individuals at high risk.^7,9,16

Accurate assessment of an asymptomatic individual’s risk for pancreatic cancer is of the utmost importance in deciding whether surveillance is warranted. At present, physicians must obtain an extensive family history to determine the patient’s likelihood of harboring a germline mutation associated with pancreatic cancer or whether the patient belongs to an FPC kindred. Advances in next-generation sequencing technologies have increased the use of multigene panels for germline testing at markedly decreased costs, which will ultimately allow FPC kindreds to be reclassified as having a clear hereditary basis for cancer, thereby reducing some of the heterogeneity that has been evident in cohort studies to date.

Recent guidelines have additionally recommended germline genetic testing for all patients with pancreatic cancer, given the improved treatment outcomes in those with identified defects in DNA damage repair genes such as BRCA1 and BRCA2.^16,17 When a patient is found to carry a germline cancer susceptibility gene, at-risk family members (many of whom may be unaffected by cancer) can be tested for the familial gene. If such a gene is present, those patients may be eligible for pancreatic cancer surveillance. This process, known as cascade testing, will undoubtedly increase the pool of newly identified high-risk individuals, as many may have been previously misclassified as at average or moderately increased risk for pancreatic cancer. The development and implementation of systems to routinely and precisely evaluate familial risk of pancreatic cancer will enable identification of asymptomatic individuals who stand to benefit from inclusion in a pancreatic cancer surveillance program.

In conclusion, screening for pancreatic cancer in asymptomatic, average-risk individuals should be avoided. Enrichment of the surveillance population with individuals who carry germline mutations associated with pancreatic cancer, FPC, and potentially new-onset diabetes would alter screening test characteristics in favor of surveillance in specific populations. Recent studies have begun to provide support for the hypothesis that treatment of screen-detected pancreatic cancer and precursor lesions improves pancreatic cancer mortality. Further study is needed to fully define the population who should be screened and by what optimal strategy to improve outcomes and minimize harms resulting from pancreatic cancer surveillance in individuals at increased risk.

Corresponding Author: Aimee L. Lucas, MD, MS, Henry D. Janowitz Division of Gastroenterology, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, PO Box 1069, New York, NY 10029 (aimee.lucas@mssm.edu).

Conflict of Interest Disclosures: None reported.