[Computerized tomography of pancreatic tumors].

Few pancreatic carcinomas (5-22%) are resectable at the time of diagnosis because this lesion is seldom diagnosed in an early stage. A considerable improvement in the rate of survival is described only for resectable tumors: it is extremely necessary to find an imaging technique for early diagnosis and for accurate staging of pancreatic carcinoma to discern operable from inoperable cancer. The sensitivity of CT in predicting that pancreatic carcinoma is unresectable has been described as approaching 100%. However the reverse is not true. More than one third of the tumors revealed with CT and interpreted as resectable cannot be excised. The major reason for errors with CT are failure to detected liver metastases, peritoneal implants, lymph node involvement and encasement of the great vessels by tumor. Significant progress has recently been made to improve the detection of these details with the recent introduction of helical CT with infusion of a bolus of contrast material and thin section collimation. Traditionally, when a single sequence of images was acquired during abdominal CT, the time of the acquisition was dominated by the requirement to scan during maximal hepatic enhancement, which unfortunately may not be optimal for evaluation of the pancreas. With the advent of helical CT, the acquisition of two sets of images after infusion of contrast material is now possible; the first one takes place during the arterial enhancement; it is useful to detect tumor vascular encasement and the maximum difference of tissue attenuation between normal greater pancreatic enhancement and hypodense pancreatic mass, less vascularizated. It appears that the peak parenchymal enhancement achieved with helical CT may improve the sensitivity of CT scanning in detecting pancreatic carcinoma, especially small tumors confined within the organ. The second phase takes place during the venous or portal enhancement and provides useful information about venous encasement and hepatic metastasis. Extraglandular extension with invasion of adjacent major arterial (celiac axis or its branches, superior mesenteric artery) and venous (portal, splenic, superior mesenteric) appear as soft-tissue attenuation thickening obscuring the perivascular fat, with deformity, thrombosis or occlusion of the vessels. In cases of venous occlusion, collateral vein can be identified. Dilatation of the small veins that surround the head of the pancreas might be used as an additional criterion of extrapancreatic extension of neoplasia. With the features of spiral CT (contrast material optimization and continuous scanning), the detection of small lesions in the liver and peritoneal implants has been increased. Helical CT seems not to detect anything else about lymph node involvement than conventional CT, limited by the same morphologic criteria. The only CT means of detection of node involvement by pancreatic carcinoma is the pathologic enlargement of lymph nodes without specificity for neoplastic or not neoplastic ones. In many cases 2D, 3D and MIP imaging are helpful to evaluate vasculature encasement, especially for visualization of vessels which lie in oblique, coronal or sagittal plane. Consequently helical CT has the potential to become an alternative angiographic technique. Many studies have been done to evaluate spiral CT potential impact and to compare the value of this technique with other ones in the initial diagnosis and staging of pancreatic carcinoma. One of these studies compares dual-phase helical CT and endoscopic endo-sonography. The Authors observe that the two techniques do not differ significant statistically in detecting pancreatic carcinoma, except endoscopic sonography is more sensitive than helical CT for tumors smaller than 15-20 mm. They found the accuracy to predict unresectable carcinoma is 100% for dual-phase helical CT and less for endoscopic endosonography (86%). (ABSTRACT TRUNCATED)