Transcranial perfusion sonography using a low mechanical index and pulse inversion harmonic imaging: reliability, inter-/intraobserver variability.

ABSTRACT

PURPOSE: Transcranial perfusion sonography (TPS) is an emerging noninvasive bedside method for evaluating brain perfusion. The purpose was to assess the feasibility of a low MI/almost real-time frame rate approach and to test its intra-/interobserver variability. MATERIALS AND METHODS: 10 healthy volunteers were investigated 3 times with TPS at a low MI (1.0) and a high frame rate (8.3 Hz). Investigations were performed by 2 sonographers in a cross-over design: 1.) twofold measurements each with 5 volunteers (intraobserver test), and 2.) single measurements of the other 5 volunteers (interobserver test). From 8 established regions of interest (ROI), time-intensity curves (TIC) with the following parameters were calculated peak intensity (PI), time-to-PI (TTP), area-under-curve (AUC), and cerebral transit time (CTT). The TIC quality was described by the coefficient of determination. TIC parameters were presented descriptively. Intra- and interobserver variability was tested by Spearman's correlation. RESULTS: The overall quality of the TIC was very good (mean r(2) = 0.92, 0.87 - 0.97). TTP (25.7 - 28.1 sec; mean 26.8 sec) and CTT (8.2 - 10.7 sec; mean 9.9 sec) were the most robust parameters. The intraobserver variability was lower with the more experienced sonographer (r = 0.70 vs. r = 0.29). The interobserver reliability was r = 0.34 (p < 0.05). CONCLUSION: Low MI TPS allows for nearly real-time imaging facilitating probe control. Sound sonographer experience allows for a high reliability and makes TPS an interesting tool for the diagnosis and follow-up of perfusion changes, e. g. in stroke or anti-angiogenic brain tumor therapy.