CT liver perfusion in patients with hepatocellular carcinoma: can we modify acquisition protocol to reduce patient exposure?

Objectives To investigate the potential of decreasing the number of scans and associated radiation exposure involved in CT liver perfusion (CTLP) dynamic studies for hepatocellular carcinoma (HCC) assessment. Methods Twenty-four CTLP image datasets of patients with HCC were retrospectively analyzed. All examinations were performed on a modern CT system using a standard acquisition protocol involving 35 scans with 1.7 s interval. A deconvolution-based or a standard algorithm was employed to compute ten perfusion parametric maps. 3D ROIs were positioned on 33 confirmed HCCs and non-malignant parenchyma. Analysis was repeated for two subsampled datasets generated from the original dataset by including only the (a) 18 odd-numbered scans with 3.4 s interval and (b) 18 first scans with 1.7 s interval. Standard and modified datasets were compared regarding the (a) accuracy of calculated perfusion parameters, (b) power of parametric maps to discriminate HCCs from liver parenchyma, and (c) associated radiation exposure. Results When the time interval between successive scans was doubled, perfusion parameters of HCCs were found unaffected ( p > 0.05) and the discriminating efficiency of parametric maps was preserved ( p