Pancreatic cancer surveillance is recommended for selected individuals at increased hereditary risk, including those with pathogenic germline variants in BRCA1, BRCA2, ATM, and PALB2. Endoscopic ultrasound (EUS) and magnetic resonance imaging (MRI) are commonly used for surveillance, but interpreting pancreatic abnormalities detected during imaging remains a clinical challenge.
Although surveillance recommendations have expanded, prospective evidence describing the frequency and evolution of pancreatic imaging findings in these individuals remains limited. Understanding which abnormalities are commonly detected and how they change over time is important for clinical interpretation and patient counseling.
On November 3, 2026, JCO Precision Oncology published a prospective analysis from the Pancreatic Cancer Early Detection (PRECEDE) Consortium titled: “Longitudinal Pancreatic Imaging Findings in BRCA1, BRCA2, ATM, and PALB2 Heterozygotes Undergoing Pancreatic Cancer Surveillance: A Prospective PRECEDE Consortium Study.”
The article was published in conjunction with the 2026 Annual Meeting of the Collaborative Group of the Americas on Inherited Gastrointestinal Cancer (CGA-IGC).
The study was authored by Samantha Kasselman, Jessica M. Long, Tiffiney R. Hartman, Kieran Hawthorne, John J. Graff, Fay Kastrinos, Aimee L. Lucas, and Bryson W. Katona, on behalf of the PRECEDE Consortium.
Study Design and Population
PRECEDE is an ongoing international, multicenter, prospective observational cohort study established to improve pancreatic cancer early detection in individuals at increased risk.
This analysis included participants enrolled between April 2020 and April 2025 who carried a heterozygous pathogenic germline variant in BRCA1, BRCA2, ATM, or PALB2 and underwent at least one pancreatic surveillance imaging examination by July 2025.
Of 8,956 individuals who provided consent across 57 international institutions, 2,085 participants from 45 institutions met the eligibility criteria for this analysis. The distribution of pathogenic germline variants was:
- BRCA2: 1,055 participants (51%)
- BRCA1: 432 participants (21%)
- ATM: 403 participants (19%)
- PALB2: 195 participants (9%)
The median age was 60 years, 72% of participants were female, and 65% reported a family history of pancreatic cancer in a first- or second-degree relative.
Surveillance was performed primarily using EUS and MRI with magnetic resonance cholangiopancreatography (MRI/MRCP). During the study period, 68% underwent at least one EUS examination, 66% underwent MRI/MRCP, and 5% underwent CT imaging. Some participants received more than one imaging modality.
The median number of imaging studies per participant was two, with a median interval of 360 days between examinations among those with multiple studies. The cohort contributed 2,166 person-years of surveillance. Investigators evaluated the prevalence of pancreatic imaging abnormalities, their evolution during follow-up, differences across genetic subgroups, and clinical characteristics associated with their detection.
Pancreatic Imaging Findings Were Common, but Most Were Low-Risk
At least one pancreatic imaging finding was reported in 1,118 of 2,085 participants (54%) during surveillance. The findings included pancreatic cysts, parenchymal changes, main pancreatic duct abnormalities, and solid masses. Pancreatic cysts were detected in 729 participants (35%), making them one of the most frequent abnormalities. Among individuals with cysts, 53% had multiple cysts, while 19% had at least one cyst measuring 1 cm or larger.
Among participants with pancreatic cysts, 13% (92/729) had at least one worrisome cyst feature, including a diameter of 3 cm or greater, pancreatic duct dilation, thickened cyst walls, or a mural nodule. Main pancreatic duct abnormalities were identified in 97 participants (5%), including 43 (2% of the total cohort) with duct dilation of at least 5 mm. Solid pancreatic masses, encompassing both malignant and nonmalignant lesions, were detected in 74 participants (4%).
Overall, 190 participants (9.1%) had at least one concerning imaging finding, defined as a cyst with worrisome features, main pancreatic duct dilation of at least 5 mm, or a solid mass. Parenchymal changes were also frequently observed on EUS, affecting 48% of participants who underwent the procedure. These changes commonly included hyperechoic foci or stranding.
Despite differences in imaging modality use, the overall prevalence of pancreatic findings did not significantly differ across the four genetic subgroups (P = .3). However, individuals with a family history of pancreatic cancer had a modestly higher prevalence of imaging findings than those without such a history (56% vs 50%; P = .01), primarily because of more frequent pancreatic cyst detection.
Most Imaging Findings Remained Stable During Follow-Up
The longitudinal analysis also provided information on how pancreatic abnormalities changed during follow-up. Among 428 individuals without a pancreatic cyst or solid mass at baseline who underwent at least two subsequent imaging examinations, 304 (71%) remained free of these findings throughout the available follow-up.
For individuals with pancreatic cysts, the findings were similarly reassuring in the short term. Among 273 participants with a pancreatic cyst and follow-up imaging using the same modality, 237 (87%) did not demonstrate worrisome cyst growth, defined according to the Kyoto guidelines as an increase of at least 2.5 mm per year.
Only 36 participants (13%) met this growth threshold. Investigators also observed that increases in cyst size were detected more frequently between sequential EUS examinations than between MRI examinations. They suggested that differences in measurement variability or patient selection could contribute to this observation. These findings suggest that most pancreatic cysts did not demonstrate worrisome growth during short-term follow-up, although continued surveillance remains important.
Pancreatic Cancer Detection During Surveillance
During the study period, 12 participants (0.6%) were diagnosed with incident pancreatic cancer, corresponding to an incidence rate of 5.54 cases per 1,000 person-years of surveillance. Of these 12 individuals, 11 (92%) were diagnosed with stage I or II pancreatic cancer.
The remaining patient was diagnosed with stage IV disease six months after baseline imaging, which had identified a pancreatic cyst without worrisome features. Two additional participants had advanced pancreatic cancer at baseline imaging, representing prevalent rather than incident cases. Four participants (0.2%) were also diagnosed with pancreatic neuroendocrine tumors.
Most incident pancreatic cancers were diagnosed at stage I or II, suggesting that surveillance may help detect disease at an earlier stage. However, the study was observational and did not establish whether surveillance improved overall survival or pancreatic cancer-specific mortality.
Factors Associated With Pancreatic Imaging Abnormalities
The investigators also examined clinical characteristics associated with pancreatic imaging findings. Increasing age at baseline imaging was significantly associated with pancreatic cysts, main pancreatic duct abnormalities, and parenchymal changes.
New-onset diabetes, defined as diabetes diagnosed within three years before enrollment, was associated with higher odds of pancreatic cyst detection (OR 2.48; P = .01).
A family history of pancreatic cancer was also independently associated with pancreatic cyst detection (OR 1.51; P < .001). Additionally, a history of pancreatitis was associated with main pancreatic duct abnormalities (OR 4.33; P = .002), potentially reflecting inflammation-related changes in pancreatic duct morphology.
Notably, the absence of a pancreatic cancer family history did not exclude imaging abnormalities or pancreatic cancer diagnoses. Half of participants without a known family history had pancreatic imaging findings, and four developed pancreatic cancer during the study.
These observations are relevant as surveillance recommendations continue to evolve for carriers of pathogenic germline variants, including individuals without a family history of pancreatic cancer. Nevertheless, the association between new-onset diabetes and cyst detection should be interpreted cautiously because only 36 participants reported new-onset diabetes before enrollment.
Clinical Implications and Study Limitations
The PRECEDE findings provide prospective evidence that can help clinicians counsel individuals undergoing pancreatic surveillance because of inherited genetic susceptibility.
Detection of a pancreatic abnormality during surveillance does not necessarily indicate malignancy or require immediate intervention. Most imaging findings were considered benign or low-risk, and the majority of cysts with available longitudinal assessment did not demonstrate worrisome growth. The study also provides information about imaging outcomes across BRCA1, BRCA2, ATM, and PALB2 carriers, showing broadly similar frequencies of pancreatic findings despite differences in surveillance modality use.
At the same time, several limitations should be considered. The available longitudinal follow-up was relatively short, limiting conclusions about the long-term evolution and malignant potential of pancreatic lesions. Differences in imaging interpretation and modality use across participating centers may also have influenced the reported findings.
The cohort was predominantly female (72%) and White (89%), which limits the generalizability of these results to more diverse populations. Furthermore, although most incident pancreatic cancers were detected at an early stage, the study was not designed to demonstrate a survival benefit from surveillance or to establish the optimal imaging strategy for each genetic subgroup.
Conclusion
The prospective PRECEDE Consortium study provides important evidence on pancreatic imaging outcomes in individuals carrying BRCA1, BRCA2, ATM, and PALB2 pathogenic germline variants. While pancreatic abnormalities were detected in more than half of participants, most were benign or low-risk, and the majority of cysts evaluated longitudinally did not show worrisome growth during short-term follow-up.
The detection of 11 of 12 incident pancreatic cancers at stage I or II also highlights the potential of surveillance to identify pancreatic cancer earlier, although longer follow-up and additional evidence are needed to determine its impact on survival.
For clinical practice, these findings offer a stronger basis for counseling genetically predisposed individuals about what to expect during pancreatic surveillance, how frequently abnormalities may be encountered, and why continued follow-up remains important even when initial findings appear reassuring.
