Measuring portal venous perfusion with contrast-enhanced CT: comparison of direct and indirect methods

RATIONALE AND OBJECTIVES: Two algorithms can be used to measure portal venous perfusion (PVP) with contrast material-enhanced single-level liver computed tomography. The "direct" and "indirect" algorithms use data from the portal vein and aorta, respectively. This study compared PVP values obtained with direct and with indirect algorithms in a series of patients. MATERIALS AND METHODS: Both techniques were applied in 27 patients with cirrhosis (14 men and 13 women; mean age, 56.1 years +/- 9.4) and 18 control patients (seven men and 11 women; 52.8 years +/- 12.3). A single section through the liver was scanned after intravenous injection of ioversol (40-mL bolus; 320 mg of iodine per milliliter). RESULTS: Both techniques showed reduced PVP in patients with cirrhosis (0.63 for direct and 0.17 for indirect method) compared with control patients (1.06 and 0.26, respectively), but only the direct method agreed with physiologic expectations based on animal and human studies. In separating cirrhotic and control patients, the area under the receiver operating characteristic curve was significantly greater for the direct method (0.91 vs 0.78; P = .03). CONCLUSION: Both direct and indirect methods are feasible and distinguish well between cirrhotic and control patients, but the direct method is more physiologic and is preferable if portal venous data are available.