Semi-automatic calcium scoring in CT-scans
Gutiérrez Santamaría, Sofía
MetadataShow full item record
Vascular calcification is the pathological deposition of calcium in vascular structures and it is considered a complication of atherosclerosis. Vascular calcium detection by Computed Tomography (CT) can be used as a clinical marker of atherosclerosis. Nowadays, calcium scoring is mainly based on the Agatston method, although there are others based on the volume or mass of the calcifications. In this study, semi-automatic functionalities are implemented in a new calcium-scoring software application based on MeVisLab with the sole purpose of saving radiologists time in their clinical practice. On the one hand, the interpolation of calcification segmentations have been implemented and on the other hand, the Agatston and volume scores have been computed. For its assessment, the results obtained are compared with those from imageXplorer, the software used so far, which is considered the "gold standard" in this study. For the interpolation of calcification segmentations, Dice coefficients that compare the similarity between the segmentations performed in imageXplorer and MeVisLab were calculated. The values obtained before using the editing modes fall in the range [2.3∙10-4 - 0.7] and those obtained afterwards are in the range [0.8 - 1.0]. As for the computation of the calcium scores, a Student t-test between the calcium scores from imageXplorer and MeVisLab was conducted. For both the Agatston Score (t32 = 0.03, p = 0.97) and the volume score (t32 = 0.04, p = 0.96) not statistically difference was found. Both the interpolation of calcification segmentations and the computation of the calcium scores are validated, although some future improvements are suggested.