Computed Tomography Perfusion Analysis Of Pancreatic Adenocarcinoma Using Deconvolution, Maximum Slope, And Patlak Methods - Evaluation Of Diagnostic Accuracy And Interchangeability Of Cut-Off Values

Purpose The goal of this study was to evaluate the diagnostic accuracy of perfusion computed tomography (CT) parameters obtained by different mathematical-kinetic methods for distinguishing pancreatic adenocarcinoma from normal tissue. To determine cut-off values and to assess the inter-changeability of cut-off values, which were determined by different methods.Materials and Methods Perfusion CT imaging of the pancreas was prospectively performed in 23 patients. 19 patients with histopathologically confirmed pancreatic adenocarcinoma were included in the study. Blood flow (BF), blood volume (BV) and permeability-surface area product (PS) were measured in pancreatic adenocarcinoma and normal tissue with the deconvolution (BF, BV, PS), maximum slope (BF), and Patlak methods (BV, PS). The interchangeability of cut-off values was examined by assessing agreement between BF, BV, and PS measured with different mathematical-kinetic methods.Results Bland-Altman analysis demonstrated poor agreement between perfusion parameters, measured with different mathematical-kinetic methods. According to receiver operating characteristic (ROC) analysis, PS measured with the Patlak method had the significantly lowest diagnostic accuracy (area under ROC curve = 0.748). All other parameters were of high diagnostic accuracy (area under ROC curve = 0.940-0.997), although differences in diagnostic accuracy were not statistically different. Cut-off values for BF of <= 91.83 ml/100ml/min and for BV of <= 5.36 ml/100ml, both measured with the deconvolution method, appear to be the most appropriate cut-off values to distinguish pancreatic adenocarcinoma from normal tissue.Conclusion Perfusion parameters obtained by different methods are not interchangeable. Therefore, cut-off values, which were determined using different methods, are not interchangeable either. Perfusion parameters can help to distinguish pancreatic adenocarcinoma from normal tissue with high diagnostic accuracy, except for PS measured with the Patlak method.Key Points:Perfusion CT parameters showed high diagnostic accuracy in differentiating between pancreatic adenocarcinoma and normal tissue.Only PS measured with the Patlak method showed a significantly lower diagnostic accuracy.Perfusion parameters measured with different mathematical-kinetic methods are not interchangeable.A specific cut-off value must be determined for each method and each perfusion parameter.