A new method for evaluation of cavitation near mechanical heart valves.

Evaluation of cavitation in vivo is often based on recordings of high-pass filtered random high-frequency pressure fluctuations. We hypothesized that cavitation signal components are more appropriately assessed by a new method for extraction of random signal components of the pressure signals. We investigated three different valve types and found a high correlation between the two methods (r2: 0.8806-0.9887). The new method showed that the cavitation signal could be extracted without a priori knowledge needed for setting the high-pass filter cut off frequency, nor did it introduce bandwidth limitation of the cavitation signal.