Imaging-based surveillance in patients with initially detected pancreatic cystic lesions

To determine the long-term outcomes of pancreatic cystic lesions (PCLs) undergoing surveillance, and evaluate the associations between imaging monitoring plans and detection of malignant progression.

This retrospective observational study included patients with incidentally detected PCLs (0.5-4.0 cm) on MRI between January 2015 and December 2016. Imaging modalities, time intervals, cumulative number of imaging examinations, and occurrence of malignant progression were recorded. The impact of cumulative CT/MRI examinations on detection of malignant progression at 1, 2, and 3-years was assessed using Cox-regression analyses.

A total of 911 patients (mean age, 68.8-years; 445 with intraductal papillary mucinous neoplasm [IPMN]) underwent an average of five examinations during follow-up (median, 45.0-months). All 20 cases of malignant progression occurred in patients with IPMN, increasing steadily over the maximum 9-year follow-up. The cumulative number of MRI examinations was significantly associated with detection of malignant progression at 1-year (hazard ratio [HR] = 1.40; P = 0.035), and 2-year (HR = 1.38; P = 0.047) landmark points, while CT showed no such association. The cumulative number of MRI examinations was associated with malignant progression at both 1-year (HR = 1.63; P = 0.007) and 2-year (HR = 1.59; P = 0.013) landmark points in 350 patients with IPMN without worrisome features, but not in 95 patients with IPMN with worrisome features.

Imaging surveillance of incidentally detected PCLs for detecting malignant progression should prioritize IPMNs. MRI-based surveillance would be beneficial in the initial 2-years, especially for patients with IPMN lacking worrisome features.