Back to Annual Meeting Posters

Introduction: Pancreatic cysts remain a challenge with respect to diagnosis and management. Pancreatic cysts are detected with increased frequency because of the widespread use of advanced imaging modalities. The aim of the study was to perform a clinical comparison of patients with pancreatic cysts who underwent resection versus imaging surveillance and differentiate between clinical, imaging, biochemical and pathologic characteristics.
Methods: This was a retrospective review of adult patients with cystic lesions of the pancreas treated in a single health system between January 2002 and September 2009. Endoscopic ultrasound (EUS) was required for inclusion. An encounter was defined as a clinical visit in which pancreatic imaging was performed. Imaging modalities included were EUS, with or without fine needle aspiration (FNA), computed tomography (CT), magnetic resonance (MR) and endoscopic retrograde cholangiopancreatography (ERCP). Symptoms were defined as weight loss or abdominal pain. Demographics, clinical characteristics, imaging features, biochemical analysis, procedure, pathology and follow up intervals were analyzed. Only encounters prior to resection were included in the analysis for the resection group.
Results: The final analysis included 262 patients, 58 in the resection group (22.1%) and 204 in the surveillance group (77.8%). Demographics, clinical features and follow up analysis are included in Table 1; p values and interquartile range (IQR) have been included. The median time to resection from first encounter was 85.5 days; 52 patients (89.6%) underwent resection within one year of first encounter. The most common cyst locations were the pancreatic body and tail (53%). The median number of cysts by CT was one in both the resection and surveillance groups. The median initial cyst size by CT in millimeters was 31 in the resection group and 21.5 in the surveillance group (p<0.001). The presence of septation (p=0.216) or a solid component (p=0.957) were not significantly different between groups. Median cyst fluid carcinoembryonic antigen (CEA) level (ng/ml) was 293.9 in the resection group and 19.8 in the surveillance group (p=0.006). Median cyst fluid amylase (U/L) was 72.5 in the resection group and 5096 in the surveillance group (p=0.007). Pathologic analysis of resected specimens demonstrated 29% of lesions were malignant, 26% had dysplasia and 45% were benign. In the resection group the most common diagnoses were pseudocyst (22.4%) and adenocarcinoma (18.9%) and the most frequent procedure was distal pancreatectomy (50%).
Conclusion: The data suggests patients who require resection can be identified early. This is based primarily on cyst size and elevated cyst fluid CEA. Patients who underwent resection had fewer encounters at shorter intervals.
Table 1: Demographic and Clinical Characteristics of Patients who underwent Resection versus Surveillance

Characteristic	Total N=262	Resection N=58	Surveillance N=204	P Value
Female (%)	158 (60.31%)	31 (53.45%)	127 (62.25%)	0.226
Deceased (%)	70 (26.72%)	17 (29.31%)	53 (25.98%)	0.613
Symptomatic (%)	169 (64.5%)	41 (70.69%)	128 (62.75%)	0.264
Median age at diagnosis in years (IQR)	66 (55, 75)	61 (49, 72)	67.5 (55.5, 76)	0.027
Median number of encounters (range 1-15)(IQR)	3 (2, 5)	2 (1, 3)	3 (2, 5)	<0.001
Median interval between encounters in months (IQR)	4.13 (1.73, 10.5)	1.87 (0.83, 4.73)	4.63 (1.93, 11.69)	<0.001
Median follow up in months(IQR)	10 (1.1, 30.8)	1.17 (0, 6.9)	16 (3, 35.5)	<0.001