Incidental pancreatic cystic lesions: retrospective analysis of natural history and efficacy of imaging surveillance guidelines

Objectives

To assess the natural history of incidental pancreatic cystic lesions (PCLs) and the sensitivity of surveillance guidelines to detect malignancy.

Materials and methods

We performed a single-center retrospective analysis of incidental PCLs discovered between 2012 and 2017. Patients were identified by searching radiology reports for relevant keywords, then search results were manually reviewed to exclude patients with < 5 years of follow-up, non-incidental PCLs, known pancreatic cancer, or pancreatitis. Baseline and follow-up imaging was reviewed to record size, growth (as defined by the American College of Radiology (ACR)), high-risk stigmata, and incidence of malignancy. Four major guidelines (International Consensus, European, ACR, and Canadian Association of Radiologists (CAR)) were retrospectively applied to assess sensitivity to detect invasive disease. Benign and malignant PCL outcomes were compared (Fisher’s exact test), and PCL growth was modeled (mixed-effects regression).

Results

Four hundred and forty-nine patients (67 ± 12 years, 284 women) with 556 PCLs were included. PCL natural growth was 0.33 mm/year (0.26–0.40 mm 95% CI), which overlaps with significant growth thresholds for ACR and CAR guidelines. Nine patients over 21,996 patient-years of follow-up developed pancreatic cancer (4.1 cases/10,000 patient-years). Malignancy was not associated with significant growth (4/9 vs 107/547, p = 0.084) but was with high-risk stigmata (3/9 vs 3/547, p < 0.001). Guideline sensitivities ranged from 33.3 to 66.7% (2–4/6 cases).

Conclusion

PCLs are common, overwhelmingly benign, and grow slowly. Development of high-risk stigmata is associated with malignancy but not “significant growth,” which can overlap with natural PCL growth. Despite differences in recommended duration and frequency of follow-up, surveillance guidelines still miss 1/3–2/3 of cancers.

Key Points

Question There is limited evidence on the natural growth and malignant potential of incidental pancreatic cystic lesions.

Findings Incidental pancreatic cystic lesions grew slowly (0.3 mm/year) and were uncommonly malignant (2%). Retrospective application of surveillance guidelines missed 1/3–2/3 of cancers.

Clinical relevance Radiologists and imaging surveillance guidelines should account for the natural growth of incidental pancreatic cystic lesions. Given the low rate of malignancy, high proportion of missed cancers, and risk of overtreatment, the effectiveness of imaging surveillance guidelines remains unclear.

Graphical Abstract