Parametric analysis of cerebral ischemia with contrast enhanced ultrasound and comparision with histology in an experimental stroke model

Background: Modern ultrasound imaging modes allow to image brain perfusion with the help of echocontrast agents. The mode of continuous infusion uses the high-pulse destruction of echo contrast agents and monitors the replenishment into cerebral capillaries. We compared the value of a planimetric and a parametric analysing mode with an automated pixel analysis programme (MCE 2.9.3, Jabko Inc.) to detect perfusion defects in an experimental stroke model. Methods: In 6 rats cerebral perfusion defects were measured after craniectomy with the pulse inversion imaging mode 24 hours after induction of cerebral ischemia by the filament model. The replenishment of cerebral capillaries was measured with increasing pulsing intervals from 0,15 to 15 seconds. The region lacking any intact microcirculation was determined by planimetry of the area lacking opacification at the longest pulsing intervals. Secondly, automated pixel intensity threshold analysis (PITA) was used which displays only pixels that do not exceed a priori defined ratios of the lowest intensity on parametric flow pictures. Histology was performed with 2,3,5 triphenyltetrazoliumchlorid to stain for infarcted areas on the sonographic level. Results: Increasing the threshold from 1% to 10% of minimal intensity resulted in an enlargement of the computer-determined region of relative hypoperfusion. The selection of a minimum pixel threshold of 5% of minimum resulted in a mean area of 18.3%±5.7%. There was also a close correlation between PITA using a 5% threshold and the region determined by subjective planimetry from averaged frames with the longest PI with a mean size of 18.3%±6.2%. Intra- and interobserver variability were low (SEE=0.02 for all error measurements, r values >0.97). Conclusion: The extent of cerebral ischemia measured by the planimetric approach corresponds well with the the observed-independent PITA analysis of parametric pictures. Automated analysis of ultrasound perfusion maps enable to give a good prediction of cerebral infarction in focal cerebral ischemia.