|Publication type:||Conference other|
|Type of review:||No review|
|Title:||Aneurysm shape as a diagnostic tool : a machine learning approach|
|Conference details:||International Neurovascular Exploratory Workshop (iNEW'2018), Zürich, 7-9 February 2018|
|Subject (DDC):||006: Special computer methods |
616: Internal medicine and diseases
|Abstract:||Recent studies have found supporting evidence that the shape of an intracranial aneurysm can be used as a proxy for disease status. Although the shape, as seen in 3D imaging data, already plays a role in the clinical assessment of aneurysms today, tools to quantify and compare aneurysm morphology in a generic, standardized way are still lacking. Here, we present a machine learning approach based on a broad spectrum of shape descriptors to predict the aneurysm rupture status. Results are based on a dataset consisting of over 400 segmented aneurysm models. We extended our analysis by including human ratings of aneurysm shape. A correlation analysis of these ratings with quantifiable morphological parameters allowed us to identify shape descriptors mimicking the human assessment. Preliminary results based on 134 geometric aneurysm models and 15 assessments of human raters show that human assessment of irregular shape correlates well with curvature metrics, spread of the writhe number distribution and non-sphericity index.|
|Fulltext version:||Published version|
|License (according to publishing contract):||Licence according to publishing contract|
|Departement:||Life Sciences and Facility Management|
|Organisational Unit:||Institute of Computational Life Sciences (ICLS)|
|Published as part of the ZHAW project:||AneuX|
|Appears in collections:||Publikationen Life Sciences und Facility Management|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.