Vascular invasion in pancreatic cancer: evaluation of endoscopic ultrasonography, computed tomography, ultrasonography, and angiography.

Abstract

Current methods for detecting vascular invasion in pancreatic cancer can be inaccurate, invasive, and expensive. The aim of this study is to assess the value of current imaging modalities in determining vascular invasion by pancreatic cancer. The results of Endoscopic Ultrasonography (EUS), Computed Tomography (CT), Ultrasonography (US), and Angiography performed in 170 patients, suffering from pancreatic cancer, were retrospectively studied and correlated with intra-operative findings and surgical anatomopathological diagnosis after resection. We assessed sensitivity, specificity, positive and negative predictive values, and accuracy for detecting vascular invasion. EUS turned out to be the most reliable imaging technique for detecting vascular invasion in pancreatic cancer, with a sensitivity of 55%, specificity of 90%, positive predictive value of 61.1%, negative predictive value of 87.5%, and accuracy of 82.2%. CT results were 39.4%, 90%, 52%, 84.4%, and 79.1% for the respective categories, with however, better results with multislice CT. The US results were 3.7% for the sensitivity, 96.3% for the specificity, 25% for the positive predictive value, 75.2% for the negative predictive value, and 73.4% for the accuracy. For angiography, the sensitivity, the specificity, the positive predictive value, the negative predictive value, and the accuracy were 52.6%, 72.3%, 43.5%, 79.1%, and 66.7% respectively. In this study, EUS was the most valuable imaging modality in assessing vascular invasion (especially for venous invasion) for pancreatic cancer, with an accuracy of more than 80%. A further prospective study should be carried out to evaluate the combination of imaging modalities for the detection of vascular involvement, especially with multi-slice CT which almost reached the performances obtained by EUS.