CONTRAST BOLUS DYNAMIC COMPUTED-TOMOGRAPHY FOR THE MEASUREMENT OF SOLID-ORGAN PERFUSION

RATIONALE AND OBJECTIVES. The authors have investigated the aortic res ponses to various intravenous bolus injections of nonionic and ionic c ontrast media and have presented data illustrating the potential of ul trafast computed tomography (CT) to quantify perfusion in the kidney, liver, and spleen. METHODS. Bolus Dynamics Study: Performed in 3 healt hy dogs (weight: 35kg to 36 kg). In 2 dogs, 15 mL of the nonionic agen t iohexol and the ionic agent sodium-meglumine diatrizoate were inject ed at 5, 10 and 20 mL/sec via a venous catheter placed in the superior vena cava; the order of injection was alternated between the 2 dogs. In the third dog, 25 mL of iohexol 300 mg I/mL was compared with diatr izoate 370 mg I/mL with injection rates of 10 and 20 mL/sec. Computed tomography scanning at the level of the midabdominal aorta was perform ed using an ultrafast CT scanner. Time-density curves were drawn for r egions of interest over the aorta, and gamma-variate fits performed. P erfusion Studies: Dynamic perfusion scans of the upper abdomen were pe rformed in more than 50 patients. A dose of 25 mL of iohexol 300 mg I/ mL was injected at 10 mL/sec via an intravenous cannula in the antecub ital fossa, followed immediately by 25 mL of saline, at the same rate. Scanning was performed at a single level using an ultrafast CT scanne r. Regions of interest were drawn and gamma-variate fits were applied to the vascular time-density curves. RESULTS. Bolus Dynamics: Excellen t curve fits for aortic time-density curves were obtained. A 10-mL/sec versus a 5-mL/sec bolus produced an 8% higher peak density. Nonionic contrast increased the peak density by a mean of 6%, increased the are a under the corrected time-density curve by a mean of 22%, and lengthe ned the increase time by a mean of 21%. Perfusion Studies: Values obta ined were reproducible and correlated well with values predicted from inert gas wash-out techniques. CONCLUSIONS. Changes in the CT number i n a region after an intravenous injection of contrast medium may be us ed to calculate blood flow per unit volume of tissue. Ultrafast CT off ers sufficient data points for accurate calculation. The quality of th e aortic bolus is of great importance. Nonionic media offer several im portant advantages: hemodynamic perturbation is minimized, and they ar e better tolerated at the high injection rates needed. Low-osmolality nonionic agents produce ''better'' curves than conventional high-osmol ality ionic agents, all other factors being equal. The resulting data are relevant to intravenous digital subtraction angiography and indire ct portography as well as to perfusion measurement. The technique of q uantitative dynamic CT is theoretically applicable to any cross-sectio nal modality, notably magnetic resonance.