Perfusion CT of head and neck cancer: effect of arterial input selection

The purpose of this study is to evaluate the effect of arterial input selection on perfusion CT parameters of head and neck tumors.

Perfusion calculations were done for 50 cases using deconvolution-based software. Peak enhancement values of the ipsilateral internal carotid artery (ICA) and external carotid artery (ECA) were recorded. Blood flow (BF), blood volume (BV), mean transit time (MTT), and permeability surface area product were calculated using ipsilateral ECA, ipsilateral ICA, and contralateral ICA as input arteries. Values were compared using Wilcoxon's matched pair test and Pearson's correlation coefficients (r).

A highly significant correlation was observed between peak enhancement values of the ICA and ECA (r = 0.97; p < 0.0001). A high correlation was observed between perfusion calculations obtained using ipsilateral ICA and ECA (BF, r = 0.98; BV, r = 0.92; MTT, r = 0.91; and permeability surface area product, r = 0.89), ipsilateral and contralateral ICA (BF, r = 0.97; BV, r = 0.95; MTT, r = 0.93; and permeability surface area product, r = 0.89), as well as left and right ICA (BF, r = 0.97; BV, r = 0.95; MTT, r = 0.94; and permeability surface area product, r = 0.88). All correlations were statistically significant (p < 0.0001). No significant differences were observed between perfusion calculations obtained using ipsilateral ICA versus ECA, ipsilateral versus contralateral ICA, or left versus right ICA (p > 0.05).

Arterial input selection has no significant effect on perfusion CT calculation of head and neck cancer. For standardization and simplification of postprocessing, we recommend the use of the ICA instead of the ECA as the arterial input because of its better visualization, perpendicular course, and larger caliber, all of which decrease partial volume effects.