No abstract

No abstract

The principle of microwave ablation (MWA) is to cause irreversible damage (protein coagulation, necrosis, etc.) to tumor cells at a certain temperature by heating, thereby destroying the tumor. We have long used functional near-infrared spectroscopy (fNIRs) to monitor clinical thermal ablation efficacy. After a lot of experimental verification, it can be found that there is a clear correlation between the reduced scattering coefficient and the degree of tissue damage. During the MWA process, the reduced scattering coefficient has a stable change. Therefore, both temperature (T) and reduced scattering coefficient ( μ s ′ ${\mu }_{s}^{\prime }$ ) are related to the thermal damage of the tissue. This paper mainly studies the changing law of T and μ s ′ ${\mu }_{s}^{\prime }$ during MWA and establishes a relationship model. The two-parameter simultaneous acquisition system was designed and used to obtain the T and μ s ′ ${\mu }_{s}^{\prime }$ of the ex vivo porcine liver during MWA. The correlation model between T and μ s ′ ${\mu }_{s}^{\prime }$ is established, enabling the quantitative estimation of μ s ′ ${\mu }_{s}^{\prime }$ of porcine liver based on T. The maximum and the minimum relative errors of μ s ′ ${\mu }_{s}^{\prime }$ are 79.01 and 0.39%, respectively. Through the electromagnetic simulation of the temperature field during MWA, 2D and 3D fields of reduced scattering coefficient can also be obtained using this correlation model. This study contributes to realize the preoperative simulation of the optical parameter field of microwave ablation and provide 2D/3D therapeutic effect for clinic.

This paper puts forward a measurement method of alcohol content in the mixture of water and alcohol based on near-infrared spectroscopy and segmentation regression model. The wavelengths, which are 1580, 2081 and 2309 nm, are picked out as three characteristic absorption wavelengths of alcohol, whereas 1206, 1456 and 1936 nm are picked out as three characteristic absorption wavelengths of water. The characteristic absorbance has a linear relationship to the alcohol concentration in the sample according to the principle of Lambert. Multiple linear regression equations 1 and 2 are built respectively based on the characteristic absorbance of alcohol and water, and the alcohol concentration is calculated using these two equations. However, the representative characteristic of overtone and combination absorption will change as the concentration of composing component changes so that the measuring effects of different regression equations, which are built by characteristic absorption values of different component, are different in the measurement of different concentration samples. In order to improve the prediction accuracy, this paper adopts the segmentation regression model equation to calculate the alcohol concentration in the sample, ie, we use equation 2 to calculate the alcohol concentration for the low alcohol content samples and use equation 1 to calculate the alcohol concentration for the samples of alcohol concentration between 50% and 70%; on the other hand, we use the mean calculation of equations 1 and 2 as prediction for the high alcohol concentration samples. The adoption of the regression model equation can reduce the mean Downloaded from prediction error for the samples in different alcohol concentrations without increasing the complexity of the equation. Nevertheless, we need to obtain the approximate value of alcohol concentration first in order to determine the distribution range of alcohol concentration.