Quantitative analysis of dynamic and thoracic computed tomography for the non-invasive diagnosis of pulmonary hypertension.

Abstract

Pulmonary hypertension (PH) is a chronic disorder of the pulmonary circulation marked by elevated pressure and vascular resistance. This results in functional limitations, increased load on the right heart and ultimately leads to right-sided heart failure. Currently, the diagnosis of PH requires invasive right heart catheterisation. While a number of quantitative readouts determined from various imaging modalities have been proposed for non-invasive diagnosis of PH, so far they have not been adopted in clinical practice. This is mainly because of their limited diagnostic power and their need for often extensive user interaction. Computed tomography (CT) is typically applied early in the patient’s diagnostic work-up and is therefore a promising modality for early recognition of PH. This work deals with the analysis of novel quantitative readouts for the diagnosis of PH, determined with little or no user interaction from CT scans, and the development of methods to achieve them. Specifically, we examine whether there exist differences in the passage of a contrast material bolus through the pulmonary vasculature, which can be determined from dynamic CT images and enable a distinction between patients with and without PH. This data also allows to determine the prognostically relevant parameter cardiac output. Further, the lung vessel morphology can be analysed with fully-automatic algorithms to yield information on the presence of PH. Application of state-of-the-art algorithms can distinguish between arteries and veins with a high accuracy and thus further refine the readouts of vessel morphology.

Date of Award	2021
Original language	English
Awarding Institution	Graz University of Technology