Tissue differentiation using optical emission spectroscopy for gastric mucosal devitalisation

Abstract

Argon plasma coagulation is a promising new approach to combat obesity. During the procedure, a large area of the gastric lamina mucosae and the parietal cells that are located in this layer are ablated. This reduces the secretion of the hormone ghreline, which is known to play a major role in inducing appetite. One hypothesis is that this procedure partly suppresses the feeling of hunger which lowers the food-intake and body weight effectively. To rule out side effects, the treatment of other structures of the gastric wall that are much more sensitive to thermal injury, such as the tela submucosa and the lamina muscularis propria, needs to be avoided. The possibility of layer specific tissue differentiation is explored in this work using optical emission spectroscopy on ex vivo human stomach tissue. The optical emission generated during the plasma treatment of each layer is measured with two different spectrometers. The three main layers of the gastric wall can be differentiated by specific emissions lines of electrolytic ions and trace elements. In order to evaluate the diagnostic quality of this method a linear support vector classifier is used to differentiate the three layers against each other on the basis of these emission lines of these elements. This differentiation between ‘mucosa’ and ‘submucosa’ results in a sensitivity of up to 82% and specificity of up to 92%. However, a lower sensitivity and specificity is found for a classification between ‘submucosa’ and ‘muscularis’ which implies that a classification between ‘mucosa’ and ‘not mucosa’ would be a much more suitable classification for a later clinical application.