Monitoring Cavitation Activity And Detecting Tissue Abnormal Signals Based On Signal Processing

High-intensity focused ultrasound(HIFU)is widely used as a noninvasive method in ultrasonic diagnosis and therapy applications.Because of the high energy required for HIFU,cavitation and thermal effects are always considered to be important mechanisms in HIFU-induced treatments.Pulsed HIFU(pHIFU)has been more commonly used in recent years for applications such as shockwave treatment,drug delivery and treatment of painful conditions.In despite of the impressing therapeutic effects by employing cavitation,there is non-ignorable evidence that it causes tissue injury meanwhile.Besides bubbles formed in the propagation path scatter and absorb much of the acoustical energy before it reaches the target area.Not only primary cavitation excited by pHIFU is crucial to the treatment,the residual daughter bubbles accompanying the collapse of cavitation bubbles also play an important role.These residual daughter bubbles generated during the implosion of the primary bubbles at one pulse can persist for seconds until the next pulse.On one side,they will attenuate the energy and delay the time for reaching the target,which weakens therapeutic efficiency.On the other side,they may also work as nuclei which may be excited by subsequent waves,which means the residual bubbles may improve the theraphy,or lead to further damages.How to take the advantages of ultrasound cavitation while avoid the damage that it can cause is an urgent problem to be solved.Monitoring cavitation process is the foundation for the further researches.In this work,the cavitation event was monitored based on digital signal processing.The cavitation dose calculated from the cumulative broad-band acoustic noise signals was proposed to present the cavitation activity.Taking the benefit of high sensitivity,the influences of various factors(e.g.,driving power,gas content and temperature)were studied for this system.Furthermore,the monitoring and anticipation of inertial cavitation activity under pHIFU was explored by in vitro experiments.The inertial cavitation thresholds and the inertial cavitation doses of various ultrasound contrast agent solutions under different physical conditions were measured by passive cavitation detection systerm.A model based on a support vector machine was then proposed to detect the occurrence of the inertial cavitation events.Additionally,a regression model for predicting the inertial cavitation dose was also given.The experimental results proved that the combination of the passive cavitation detection system with the developed support vector machine framework can assist in both reducing the time required to calculate the inertial cavitation threshold to enable the detection of inertial cavitation events,and anticipating the required inertial cavitation dose before use of pHIFU.The two models will be helpful in controlling and optimizing inertial cavitation activity during ultrasound therapy.The behaviour of the residual daughter bubbles was also investigated.The B-mode imaging technique was applied to monitor the cavitation bubbles dissolution behaviors during in vivo shockwave treatments.Besides,the integrated scattering cross-section of dissolving gas bubbles was simulated by employing gas bubble dissolution equations and Gaussian bubble size distribution.The results showed that B-mode imaging technology is an effective tool for monitoring the temporal evolution of bubble dissolution process,which is crucial for evaluating and controlling the cavitation activity generated by residual daughter bubbles during a treatment.Furthermore,the bubble characters,including the bubble size distribution and gas diffusion,can be evaluated by simulating the experimental data properly.This work may provide a possible method to achieve an optimal shockwave treatment of minimizing injuries without compromising the curative effect by adjusting the dissolving procedure of residual daughter bubbles between two sequent shockwaves.The researches on detecting tissue abnormal signals in ultrasonography were also explored.Taking the benefit of convenience and high sensitivity,ultrasonography is the general choice for detecting tissue abnormalities.Due to the complexity of lesions and uncertainties in medical knowledge,ultrasound diagnosis still needs to be improved although it has been applied for years.A new approach of discriminating tissue abnormlity was developed based on statistical analysis.The microstructure changes of the lesion was quatified as relative P-value by comparing its distribution with that of surrounding normal part using only one frame of raw RF data.Meanwhile,fractal dimension which reflects the texture properties was also extracted from the corresponding B-mode images.Both the two features show promising result in differentiating thyroid nodules and can assist the ultrasound diagnose.This study lays a theoretical foundation for controlling and optimizing ultrasound cavitation in its applications,and proposes methods for discriminating tissue abnormalities.