X-Ray Tomography on Intestine Samples

Sammanfattning: Enteric neuropathy describes a disease of the bowel caused by anomalies in the enteric nervous system. Patients suffering from this disease have symptoms including nausea, vomiting, diarrhea and pain. As enteric neuropathy is not well understood yet, its diagnosis is a non-trivial task. Research on the enteric nervous system is needed in order to gain a better understanding of this disease, making a correct diagnosis and targeted treatment possible. Thus, this thesis presents the investigation of nervous cells in the bowel and different setups that can be used for this purpose. More specifically, this involves the recording of high-resolution 3D images, with the ultimate aim to segment the ganglia, that are made up of neurons, in the intestinal samples to determine their volumes and compare the ganglia volumes of samples from healthy with those of diseased patients. In order to do this, X-ray tomography based on absorption and phase contrast was performed. This master thesis presents the comparison of two experimental setups, namely a lab based micro-CT in Lund, Sweden and the TOMCAT beamline at the Paul-Scherrer-Institut in Villigen, Switzerland, that were used to carry out these studies. It was concluded that the setups could not be compared in terms of contrast-to-noise ratio, as their determination was not successful. Regarding the resolution, it was found that the micro-CT setup is appropriate to perform overview scans of the sample, but its overall resolution is not high enough for the successful segmentation of the ganglia. In contrast, the TOMCAT setup delivered a resolution sufficient for the ganglias’ segmentation. The visibility of the ganglia could be significantly improved by applying phase retrieval, using the Paganin algorithm, to the datasets before their reconstruction due to this procedure enhancing the contrast. A reconstruction of the TOMCAT data with a self-developed script was conducted and compared to the reconstruction that was done directly at the beamline. It was found that the self-developed reconstruction was successful overall, but not optimal when it came to applying the Paganin algorithm, leading to an ambiguous separation between tissue and background. Furthermore, the optimal δ/β ratio for the phase retrieval reconstruction at the micro-CT setup was confirmed to be 217.6. After these preparations to find the optimal conditions for the segmentation of ganglia, the volumes of the ganglia, remaining biological tissue and background were determined to be 0.011 mm3, 0.44 mm3 and 0.067 mm3 in a 0.5 mm3 volume of one sample. The segmentation and subsequent determination of the volumes of further ganglia will pave the way for gaining a deeper knowledge of enteric neuropathy.