Feasibility of Measuring Human Pancreatic Perfusion In Vivo Using Imaging Techniques

OBJECTIVE: The objective of this study was to demonstrate the feasibility of pancreatic perfusion computed tomography (CT) and review pancreatic perfusion measurements by various imaging modalities. METHODS: Dynamic CT data from 8 patients (4 men; mean age, 64.8 [SD, 12.1] years; range, 40-80 years) with normal pancreas were analyzed using 2 analytical models: the maximum-slope and compartment-model methods. Literature search was also performed. RESULTS: Although the perfusion value estimated by the maximum-slope method (88.1 [SD, 42.1] mL/min per 100 mL) was significantly smaller than that of the compartment-model method (127.0 [SD, 70.5]; P < 0.001), there was a linear correlation between them (r = 0.97, P < 0.001). In the literature review, 15 studies that reported the absolute values of normal pancreatic perfusion, by using perfusion CT, dynamic magnetic resonance imaging, hydrogen gas clearance method, and 15O-H2O-positron emission tomography were found. The reported mean values of normal pancreatic perfusion ranged from 38.4 to 356 mL/min per 100 mL, and there was a great deal of individual variation. CONCLUSIONS: Perfusion CT may provide reliable perfusion measurements of the pancreas, and the normal value was estimated at around 100 mL/min per 100 mL with a great deal of individual variation. The maximum-slope method may provide a lower perfusion value compared with the compartment-model method.