Liver perfusion studied with ultrafast CT

Our goal was to quantify absolute hepatic arterial and portal venous perfusion noninvasively in patients with and without liver disease using ultrafast CT.

A single slice through the porta hepatis was repeatedly scanned after bolus injection of 25 ml of iohexol 300 mg I/ml, followed by a 25 ml saline "chaser" intravenously at 10 ml/s. Thirty-nine controls, 7 cirrhotic patients, and 5 patients with known metastases on the slice plane were studied; hepatic arterial perfusion was determined in 41 patients and portal venous perfusion in 24. Time-attenuation curves from regions of interest drawn over the liver, spleen, aorta, and portal vein were analysed. Hepatic arterial perfusion was calculated by dividing the peak gradient of the liver time-attenuation curve prior to the time of peak splenic attenuation by the peak aortic CT number increase. Splenic perfusion was calculated by dividing the peak gradient of the splenic time-attenuation curve by the peak aortic CT number increase. Portal perfusion was derived by scaling the splenic time-attenuation curve by the ratio of hepatic arterial/splenic perfusion. This scaled curve was subtracted from the liver time-attenuation curve to give a portal curve. The peak up-slope of this curve was divided by the peak rise in splenic or portal vein density.

Hepatic arterial perfusion averaged 0.19 ml/min/ml (n = 31) in controls and was raised in cirrhosis to 0.25 ml/min/ml (n = 6) and metastases 0.43 ml/min/ml (n = 4). Portal venous perfusion was 0.93 ml/min/ml (n = 19) in controls and 0.43 ml/min/ml (n = 4) in cirrhosis. Reproducibility has been confirmed.

Dynamic ultrafast CT shows potential in quantifying arterial and portal hepatic perfusion. The technique may be adaptable to dynamic bolus MRI.