Week 3: Dog in the PET scan?

Week three started off a bit slower compared to week 2. I learned more about what texture analysis is and the different methods for quantifying the “texture” of an image. One way to think about texture is to look at the frequency of the same pixel throughout the image, rather than the pixel intensities that are present in a scan. For example, in a picture of a tomato and an apple, both the fruits might have similar coloring, and the shape might be different, but the pattering of the apple skin will be different than that of the tomato. Although the human eye is good at identifying the difference in texture between these two fruits, it is not able to quantify that difference. However, with texture analysis, the difference in the patterns can be quantified, by looking at the standard deviation, range, spread or relation to the nearest neighbor of the segmented pixel intensities. These are just a few examples of the different types of texture analyses that are possible. The applications of texture analysis are in identifying diseased tissue from healthy tissue and finding similarities in output information from different biomedical imaging modalities. This has the potential for increasing the efficiency of the diagnosis process, by allowing the patient to take only one set of scans, rather than multiple ones.

I also learned about the physical theory behind how a positron emission tomography (PET) scan works. A radioactive isotope, germanium-68, is injected into the patient, which gives off gamma rays. When the gamma rays interact with the electrons in the patient’s body, they undergo a process called Compton scattering. During this scattering, part of the energy of the gamma ray is absorbed by a valence electron, emitting a lower energy electron and gamma ray. The electron is absorbed in the scintillator, which converts the electrical energy into photons. The scintillator is then connected to a photomultiplier tube, which absorbs the photons and progressively increased the electron output causing a high output signal. Although there are more complex procedures that take place, this is my understanding of the imaging modality so far. Next week, I hope to have more data so that I can test whether our hypothesis will be disproved or not.