Research On Non-invasive Detection Of Esophageal Tumor Based On Body Surface Vibration Signal

Esophageal neoplasm is a malignant tumor that originates in the esophageal epithelium.The therapeutic effect of early esophageal cancer is the best.Detection and diagnosis of early esophageal cancer is of great significance.Currently,most of the diagnostic methods for esophageal cancer used in clinical are invasive,which makes the patient less tolerated and causes inconvenience to the prevention and treatment of esophageal cancer.In order to promote the prevention and treatment of esophageal cancer and realize early detection of esophageal cancer,this paper studies and implements a non-invasive detection method for esophageal cancer based on surface vibration signals.This method collects the surface vibration signals of human body,which are generated by swallowing a certain amount of water under the instruction of the operator.The signals are processed and analyzed to obtain vibration characteristics of the healthy person and the esophageal cancer patient,respectively,which are served as the basis for diagnosing.The main research contents of this paper are as follows:First,simulate the peristaltic movement characteristics of the esophagus during swallowing.Establish a hydromechanical simulation geometric model and mathematical model of the esophagus,and simulate the esophageal peristalsis to analyze the radial acceleration.The simulation results show that the frequency of the peristaltic wave is mainly concentrated at 0-30 Hz,and the signal amplitude is highest near the first stenosis,so in the experiments,the signal collection point is chosenat the center of the ventral cricoid cartilage of the neck.The shear modulus of the esophagus wall affects the amplitude of the vibration signals and the energy distribution of the frequency.In the analysis part,we use the wavelet packet coefficient energy spectrum of the vibration signal as the eigenvector.The viscosity of the bolus swallowed had no significant effect on the peristaltic movement of the esophagus.The subjects should swallow the water in combination with the actual study.Second,establish a body surface vibration signal acquisition system.Design and build a signal acquisition unit consisting of a vibration signal acquisition module and an audio signal acquisition module,a signal pre-processing unit composed of a filter noise reduction module,an amplification module,an ADC module,and a controller unit composed of an embedded system and a power supply module.The signal acquisition unit,the pre-processing unit and the controller unit constitute a body surface vibration signal acquisition systemThird,acquire vibration signal data used for feature extraction and detect vibration signals of esophageal cancer patients.Acquire the body surface vibration signals from healthy subjects and esophageal cancer patients when swallowing water by the body surface vibration signal acquisition system.The signals are divided into training group and test group respectively.The signals of the training group are decomposed and reconstructed by wavelet packet,and the energy spectrum of wavelet coefficients is used as the eigenvector to establish the classifier model.Then the signals of the test group were tested by the classifier model to evaluate the body surface vibration signals of esophageal cancer patients.The signal recognition rate of esophageal cancer patients reaches76.7%,the signal recognition rate of healthy subjects reaches 83.3%.