Growth model for abdominal aortic aneurysms using longitudinal CT images

Date

2020

Author

Akkoyun, Emrah

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

1
views

1
downloads

An Abdominal Aortic Aneurysm (AAA) is diagnosed by an enlargement of the abdominal aorta. The rupture of an AAA, associated with high mortality, is eventually observed if no surgical intervention is performed. Aneurysm repair prior to rupture is thus vital. The decision to intervene is made primarily based on the AAA size measured by a maximum diameter or its growth rate. However, 10 – 24% of aneurysms below the intervention threshold experience rupture in some series. There are many complex interactions involved, from the hemodynamics and geometric properties of the aorta to patient demographic information, affecting the aneurysms’ expansion. Furthermore, the follow-up diameters can be predictable if a patient follows the common growth model of the population. However, a rapid expansion of AAA, often associated with higher rupture risk, might be observed. This study aims to build enhanced Bayesian inference methods to predict maximum aneurysm diameter using 106 CT scans. The utility of master curves and their prediction capabilities in terms of different geometrical parameters were examined. Among all the parameters, the master curve of spherical diameter performed best, predicting the diameter within 0.42 mm in 95% of all scans. Furthermore, a two-step approach based on Bayesian calibration was used and the aneurysm growth model was specified according to individual patient characteristics. Using the enhanced prediction model, 86% of scans were correctly predicted. Thus, the prediction of a measurement at any time-point can be made, along with an associated uncertainty to provide a clinically helpful tool for surgical planning and patient management.

Subject Keywords

Aortic aneurysms., Abdominal aortic aneurysm, clinical decision making, aneurysm growth, probabilistic programming, patient-oriented growth modeling

URI

http://etd.lib.metu.edu.tr/upload/12625366/index.pdf
https://hdl.handle.net/11511/45373

Collections

Graduate School of Informatics, Thesis