Pancreatic cysts stable for five years carried a low later cancer risk

Design

Retrospective single-centre cohort using LLM-extracted report features

Population

13,269 patients with pancreatic cystic lesions, 2000–2025

Primary outcome

Five-year malignancy prediction; cancer after 5 years of stability

Effect

AUROC 0.655 Kyoto vs 0.639 ACR vs 0.623 European; 8-year incidence after stability 0.71% (0.46% meeting Kyoto stop criteria)

This single-centre retrospective study, published in JACR on 28 September, used a validated large language model to extract cyst features from abdominal imaging reports between 2000 and 2025, and assigned each patient a risk category under the ACR, Kyoto and European guidelines. It included 13,269 patients; 209 (1.6%) developed pancreatic cancer or high-grade dysplasia within five years.

All three guidelines predicted cancer only modestly (AUROC 0.66 Kyoto, 0.64 ACR, 0.62 European). They agreed completely on high-risk cysts but differed widely on worrisome features (308 to 507 patients), with similar five-year cancer rates of 10% to 12% in those groups. In 1,830 patients whose cysts were stable for five years, eight-year cancer incidence was 0.71%, 0.46% if they met Kyoto criteria to stop surveillance, and 1.57% if they did not.

Most cyst surveillance never finds cancer, and the question is when to stop. These data suggest that after five years of stability, a low-risk cyst meeting stopping criteria carries a small ongoing risk, supporting a discussion about ending surveillance. It is one centre's retrospective data, and guidelines differ on stopping.

• In reports on long-followed cysts, state explicitly whether the cyst has been stable for five years.
• Use the stability record to support a discussion about stopping surveillance in low-risk cysts.
• Keep surveillance where Kyoto stopping criteria are not met; risk was about three times higher.
• Expect the guidelines to agree on high-risk features and to differ on worrisome ones.
• Recommend follow-up consistent with the guideline your service has adopted.

Why it matters

Years of scans for cysts that rarely turn malignant may be safely shortened for many patients.

Don't overread it

Retrospective, single-centre data; guidelines differ on when to stop surveillance.

The statistics, in plain English

An AUROC of about 0.65 means the guidelines ranked a future cancer above a non-cancer only about two times in three, modest discrimination. An eight-year incidence of 0.46% means about 5 in 1,000 low-risk patients developed cancer after stopping criteria were met.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.