pancreatic cysts and familial pancreatic cancer surveillance
First question for an incidental cyst: would this patient have pancreatic surgery? Not fit → no further work-up, whatever the size (ACG strong). Fit → MRI/MRCP to characterise; pseudocysts and serous cystadenomas need nothing; presumed IPMN or MCN → size-based MRI surveillance. Refer (EUS ± FNA, pancreas multidisciplinary team) for jaundice or pancreatitis from the cyst, raised CA 19-9, a mural nodule or solid component, main duct >5 mm, cyst ≥3 cm, or high-grade dysplasia on cytology; short-interval imaging for growth ≥3 mm/year or new diabetes. Stop surveillance when the patient is no longer a surgical candidate; reconsider it after 75. Familial/genetic risk (CAPS): annual imaging (MRI/MRCP and EUS at baseline, then alternating) from ~50 (40 for CDKN2A and Peutz–Jeghers) in those meeting family-history or gene criteria, plus glucose/HbA1c; consider germline testing.
incidental cysts: most are harmless, a few aren’t
Cysts are common on cross-sectional imaging and become more common with age; most are side-branch IPMNs that never progress. The guideline goal is to find the few with high-grade dysplasia or early cancer without operating on the many. Evidence for every recommendation is low or very low certainty; size and imaging features are imperfect surrogates.
| cyst type | who | imaging / fluid | cancer risk | what to do |
|---|---|---|---|---|
| pseudocyst | prior acute or chronic pancreatitis | fluid ± debris; brown fluid, high amylase, low CEA | none | nothing if asymptomatic; drain endoscopically if symptomatic → acute pancreatitis |
| serous cystadenoma | women (75%), 50s | microcystic “honeycomb”, central scar (<30%); low CEA, low amylase | ~0.1% | nothing if asymptomatic |
| side-branch IPMN | men = women, 60s–70s | communicates with the duct; often multiple; high CEA, high amylase | low | surveillance by size |
| main-duct / mixed IPMN | as above | dilated main duct (may be segmental), patulous mucin-extruding papilla | high (38–68% HGD or cancer at resection) | refer |
| mucinous cystic neoplasm | almost only women, middle age, body/tail | no duct communication; high CEA | ~10% HGD/cancer in resected series; none when <3 cm, normal CA 19-9, no concerning features | surveillance by size |
| solid pseudopapillary neoplasm | young women | solid-cystic | malignant potential | refer to MDT for resection (ACG strong) |
- a cyst found with an episode of pancreatitis may be a neoplasm that caused it, not a pseudocyst
- don’t attribute vague abdominal symptoms to a small cyst (ACG): most are asymptomatic
characterise, then decide
- MRI/MRCP is the test of choice: no radiation, best at showing duct communication. Pancreas-protocol CT or EUS if MRI isn’t possible
- imaging is only ~40–50% accurate for cyst type and 55–76% for benign vs malignant: hence the reliance on “presumed IPMN/MCN”
- EUS-FNA only when the answer would change management: cyst-fluid CEA separates mucinous from non-mucinous cysts (very high favours mucinous, very low against), amylase <250 IU/L practically excludes a pseudocyst; cytology for high-grade dysplasia or cancer. CEA does not detect malignancy within a mucinous cyst
- not fit for surgery → stop here (ACG strong): no imaging, no EUS
concerning features → EUS ± FNA and/or pancreas MDT
| features | |
|---|---|
| clinical | jaundice from the cyst, acute pancreatitis from the cyst, markedly raised CA 19-9 (>37 U/mL) |
| imaging | mural nodule or solid component (in the cyst or parenchyma), main pancreatic duct >5 mm, focal duct dilation or obstructing lesion, mucinous cyst ≥3 cm |
| cytology | high-grade dysplasia or cancer |
| prompt a short-interval MRI or EUS | new-onset or worsening diabetes, growth ≥3 mm/year |
surveillance of presumed IPMN or MCN (ACG 2018)
| largest diameter | surveillance |
|---|---|
| <1 cm | MRI every 2 years × 4 years; lengthen if stable |
| 1–2 cm | MRI yearly × 3 years, then every 2 years × 4 years if stable |
| 2–3 cm | MRI or EUS every 6–12 months × 3 years, then MRI yearly × 4 years if stable |
| >3 cm | refer to MDT; MRI alternating with EUS every 6 months × 3 years, then yearly × 4 years if stable |
- same modality each time so measurements compare (inter-observer variability can be ~4 mm)
- growth ≥3 mm/year: cysts <2 cm → MRI or EUS ± FNA within 6 months (then back to size-based intervals if stable); cysts ≥2 cm → MDT referral and EUS ± FNA
- stop when the patient is no longer a surgical candidate (ACG strong); review the value of surveillance at >75, individualise at 76–85. No evidence to stop after 5 stable years in a fit patient
- after resection: serous cystadenoma, pseudocyst or MCN without cancer → no follow-up; any resected IPMN → ongoing surveillance (6-monthly if high-grade dysplasia was resected); solid pseudopapillary neoplasm → yearly for ≥5 years
Guidelines from other societies differ on size cut-offs and stopping rules (ACG acknowledges contradictory recommendations); the shared principles are surgical fitness first, worrisome features drive referral, size drives intervals.
familial and hereditary pancreatic cancer (CAPS 2020)
Average lifetime risk (~1 in 64) is too low for population screening. Surveillance aims to find stage I cancer or high-grade dysplasia (PanIN or IPMN) in high-risk people; programmes report downstaging (most cancers found at stage I–IIB), but there are no randomised data.
who
| group | criteria |
|---|---|
| Peutz–Jeghers (STK11) | all carriers |
| CDKN2A (FAMMM) | all carriers |
| BRCA2, BRCA1, PALB2, ATM, Lynch (MLH1/MSH2/MSH6) | carrier + ≥1 affected first-degree relative (BRCA2: or ≥2 affected relatives of any degree) |
| familial, no known variant | ≥2 affected blood relatives on the same side, of whom ≥1 is a first-degree relative of the person (e.g. a parent and a sibling, or two siblings) |
| hereditary pancreatitis | most experts: from age 40 or 20 years after the first attack (no full consensus) |
- known variants explain only ~10–20% of familial clustering; ~5–10% of apparently sporadic pancreatic cancers carry one → consider germline testing for anyone eligible for surveillance
- two affected first-degree relatives ≈ 8% lifetime risk
- a single affected relative with an incidental cyst: manage the cyst as above
when and how
- start: familial without a variant → 50 or 55, or 10 years younger than the youngest affected relative; BRCA2/ATM/PALB2/BRCA1/Lynch → 45–50 (or 10 years younger than youngest affected); CDKN2A and Peutz–Jeghers → 40. No consensus on when to stop
- baseline: MRI/MRCP + EUS + fasting glucose and/or HbA1c. Then alternate MRI and EUS (no consensus on how)
- every 12 months if normal or only low-risk findings (cyst without worrisome features, small non-functioning NET <1 cm); 3–6 months for concerning findings not yet needing surgery
- new-onset diabetes in a high-risk person → investigate now, and start surveillance regardless of age
- CA 19-9 only when imaging shows a worrisome feature; CT reserved for solid lesions or unexplained duct strictures
- EUS-FNA only for solid lesions ≥5 mm, cysts with worrisome features or asymptomatic duct strictures
- surgery for positive FNA or high suspicion on imaging, at a high-volume centre. Interval cancers still happen despite annual imaging (stage I can reach stage IV within a year)
Pancreatic cancer work-up and treatment → oncology.
Prepared 2026-11-19. Not yet verified — check every number against the sources.