Objectives: Pancreatic cysts are being found with increasing frequency. Although symptomatic cysts should be resected, what to do about asymptomatic cysts is less clear. The purpose of this study was to determine threshold values to choose resection over observation for a patient with an asymptomatic pancreatic cyst.
Methods: Decision analysis assesses the consequences of a decision based on occurrence probabilities. This decision analysis assessed 1 decision, to resect or observe a patient with an asymptomatic pancreatic cyst. The consequences for resection are operative mortality/no operative mortality, short-term morbidity of patients surviving the operation, benign/malignant cyst, and life expectancy after resection of malignant lesions. The probabilities are based on the extant literature. The age-specific life expectancy for benign cysts was considered the same as for the general age-specific population from the 2002 United States Life Table. Because the operative mortality, distribution of benign versus malignant cysts, and life expectancy after resection or observation for malignant cysts vary, sensitivity analysis was done to assess the threshold values of these factors when resection becomes favored over observation.
Results: The baseline decision analysis is based on the following assumptions: 30% of lesions are malignant, operative mortality rate is 3%, and the 5-year survival rate of resected malignant cysts is 50%. Varying on age, the risk of the cystic lesion being malignant to favor resection is greater than 11.6% to 15.5%, the operative mortality rate has to be less than 7.4% to 13.8%, and the life expectancy gain by resecting, rather than observing a malignant lesion, has to be greater than 3.49 to 5.38 years.
Conclusions: The decision to resect must be based on the surgeon's operative mortality rate, the predicted operative mortality for the individual patient, the probability that the lesion is malignant, and the survival difference between resecting and observing malignant cystic lesions. Overall, for resection to be recommended, the physician must believe that the risk that the lesion is malignant is approximately 15% or greater, that the life expectancy gained from resection be approximately 5 years or greater, and that the surgeons' operative mortality rate be approximately 8% or less.