The design of an aortic heart valve prosthesis. Analysis of 100 target population patients.

The work presented focuses on the aortic valve and on the anatomical district surrounding it, the aortic root. The aim of this thesis is the study of a 3D model of the aortic root representative of the real anatomy in order to carry out experimental tests able to provide results comparable with the patient’s physiological conditions and useful to optimize the valve prosthesis design. An overview of the anatomy and physiology of the heart will be presented. The study of the anatomy of the aortic root is fundamental for the realization of the model, and for this reason the attention will be focused on it and all its aspects, evaluating which are main geometric characteristics. An analysis on patients through the visualization of pre-operative CT scans was carried out by using a specific program, Mimics, a processing software dedicated to DICOM images produced by medical diagnostic apparatus. By analyzing the CT scans, from a data-base of 100 patients, was defined in order to extrapolate the geometrical parameters useful for defining the design of a new valve. Next step a statistical analysis of each parameters (perimeter, area and eccentricity of the annulus, distance of the annulus plane from the plane of the sinuses of Valsalva and from the tubular sinus junction, and other geometrical quantities) obtained from different operators, to demonstrate the procedure repeatability, and the correlation between the data, to explain which and how data correlate each other. Also, the analysis of variance (ANOVA) was performed. For this analysis, the Minitab software was used, and it was possible to compare data groups obtained from different operators and also to understand information of interests in terms of statistical equalities or differences. The analysis carried out identified parametric and Gaussian distributions, having the patient populations a random variability between one patient and the other, even though they are patients with an aortic pathology that could affect the data distributions. Once the normality of the distributions was established, a parametric analysis was performed. For the normality test, the P-value was considered, verifying a normal distribution, for the variable under examination, only for P - values above 0.05, ie corresponding to a 95% confidence level. The corresponding boxplots were also made of each variable. In order to understand how the parameters influence each other, regressions were also performed.