PURPOSE: To determine the optimal timing for enhancement of the pancreatic parenchyma and peripancreatic vessels for the detection and staging of pancreatic adenocarcinoma using multi-detector CT.
METHOD AND MATERIALS: Thirty-seven patients with known or suspected pancreatic adenocarcinoma were imaged using multi-detector CT (Lightspeed, General Electric) following administration of 150cc of omnipaque 300 at 4cc/second. Scan delay after initiation of intravenous contrast injection was 20 seconds for arterial phase (AP), 40 seconds for parenchymal phase (PP), and 60 seconds for portal venous phase (PVP). Arterial phase was obtained using 1.25mm collimation and parenchymal and portal venous phase using 2.5mm collimation. Using regions of interest, attenuation values were obtained of the normal pancreas, pancreatic tumor, the celiac axis, superior mesenteric artery, portal vein and superior mesenteric vein in all phases of imaging.
RESULTS: The mean enhancement of the normal pancreas was 66HU (AP), 107.2HU (PP), and 101.5HU (PVP). The maximum mean arterial enhancement was observed in the PP (celiac axis=274.4 HU, SMA=282.7HU), as compared with the AP (celiac axis=224.6HU and SMA=240.5HU), and the PVP (celiac axis=148.3HU and SMA=155.7HU). The maximum mean venous enhancement was observed in the PVP (portal vein=175.2HU, SMV=164.5HU), as compared with the PP (portal vein=141HU, SMV=135.4HU), and the AP (portal vein=52.6HU, SMV=53HU). In 11 of the 37 patients, a pancreatic mass subsequently pathologically proven to represent pancreatic adenocarcinoma was detected on CT. The maximum tumor to pancreatic parenchymal contrast difference was observed in the portal venous phase in 7/11 patients (range of contrast difference=27-74), and in the parenchymal phase in 3/11 patients (range of contrast difference=36-68). In one patient, due to severe pancreatic atrophy, the attenuation measurements were inaccurate.
CONCLUSIONS: Mean enhancement of the pancreatic parenchyma was slightly superior in the parenchymal phase. Masses were best detected in the PVP. The maximum arterial enhancement was observed in the PP, and the maximum venous enhancement in the PVP. These findings suggest that when using the faster multi-detector scanners, the previously used AP may be unneccessary, and for optimal pancreatic parenchymal enhancement, scan delays of greater than 40 seconds should be used. (IRF received research support from Coulter Pharmaceuticals and Berlex. JFP received research support from Berlex and GE Medical Systems).
