This dissertation investigates novel markers for cardiovascular diseases (CVD) and Alzheimer's disease (AD). Both CVD and AD are among the large diseases counted in morbidity and mortality in the western world, which makes them huge and increasing problems. By investigating and learning characteristics of the specific disease, and how it evolves, one can design novel markers for the disease. These markers could, in the long run, aid in clinical screening programs and in drug trials which in the end could make a great difference in the life quality of the affected patients and their surroundings. The first part of this dissertation studies the growth patterns of atherosclerotic calcified deposits in the lumbar aorta based on x-ray images over an 8-year time period. We have been able to find simple growth patterns that explain how the calcifications evolve. The calcifications grew on average 41 % (p < 0.001) in the direction along the aorta wall and 21 $% (p < 0.001) in the radial direction. The center of mass of the calcification moved in average 0.60 mm (p < 0.01) downstream the aorta. We have made a prediction model for the atherosclerotic growth that is significantly better to predicting the growth of the calcifications than simpler prediction models. The second half of this dissertation investigate different proximity markers and their ability of classifying normal (NC), mild cognitive impaired (MCI), and Alzheimer subjects (AD) using a linear discrimination model. First, we have investigated four different proximity markers which lead to significantly improved marker values between NC and AD after correction for whole brain and hippocampus volume. Based on the different proximity markers we have chosen the surface connectivity marker that gave the best separation to investigate further on a larger data set. We were able to classify the three classes NC-AD, NC-MCI, and MCI-AD. The surface connectivity separated well by looking at the functional regions in the brain resulting in an AUC score at 0.877 (p< 0.001), 0.785 (p < 0.001), and 0.766 (p < 0.001) for NC-AD, NC-MCI, and MCI-AD, respectively. We have also been able to significantly distinguish between MCI-converters and MCI-non-converters with an AUC at 0.599 (p< 0.01) over a 1-year period based on the surface connectivity marker. Based on the findings in this dissertation, we can conclude that it has been possible to develop novel diagnostic and prognostic markers for CVD and AD, but there is still a long way from verifying good markers to implementing them into clinical settings.