ROI for outlining an entire tumor is a reliable approach for quantification of lung cancer tumor vascular parameters using CT perfusion

OBJECTIVE To investigate the effect of position and size of tumor region of interest (ROI) on the estimation of lung cancer vascular parameters using 256-slice computed tomography (CT) perfusion. METHODS After institutional review board approval and written informed consent, 16 men and 11 women with lung cancer were enrolled in this CT perfusion study. Perfusion, blood volume, and peak enhancement were determined for 60 or 120 mm(2) circular ROIs placed at the edge, center, and around (outlining) the visible tumor. Average values were obtained by performing ROI analysis twice by the same observers without any procedural changes. RESULTS Perfusion, blood volume, and peak enhancement measurements were substantially higher at the edge than at the center for both 60 and 120 mm(2) ROIs (all P<0.05). Measurements varied substantially depending on the ROI size. Perfusion, blood volume, and peak enhancement for the ROIs outlining tumor were intermediate between those at the tumor edge and center. There were significant correlations between median values and interquartile ranges as follows; perfusion (12.51 [7.91-28.10] mL⋅min(-1)⋅100 mL(-1)), blood volume (29.31 [21.82-37.65] mL⋅100 g(-1)), peak enhancement (12.93 [2.42-22.50]) for the ROIs outlining the tumor, and microvascular density ([19.43±8.78] vessels/0.74 mm(2)), respectively (r values were 0.732, 0.590, and 0.544 respectively, all P<0.05). CONCLUSION Spatial and size selection of ROI significantly affects CT perfusion analysis. ROI outlining of entire tumor provides efficient and reliable measurements for clinical assessment of lung cancer using CT perfusion.