Szczegóły publikacji
Opis bibliograficzny
Preoperative planning for Coronary Artery Disease using 3D segmentation and Mixed Reality / Manahil ZULFIQAR, Maciej STANUCH, Andrzej SKALSKI // W: IST 2023 [Dokument elektroniczny] : IEEE international conference on Imaging Systems & Techniques : 17–19 October 2023, Copenhagen, Denmark : conference proceedings. — Piscataway : IEEE, cop. 2023. — (IEEE International Conference on Imaging Systems and Techniques ; ISSN 2471-6162). — e-ISBN: 979-8-3503-3083-0. — S. [1–6]. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. [5–6], Abstr. — Publikacja dostępna online od: 2023-12-20. — Dod. afiliacja autorów: MedApp S. A.
Autorzy (3)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 151144 |
|---|---|
| Data dodania do BaDAP | 2024-01-16 |
| Tekst źródłowy | URL |
| DOI | 10.1109/IST59124.2023.10355732 |
| Rok publikacji | 2023 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Institute of Electrical and Electronics Engineers (IEEE) |
| Czasopismo/seria | IEEE International Conference on Imaging Systems and Techniques |
Abstract
Coronary arteries are crucial for keeping the function of our heart muscles by delivering the necessary resources like oxygen and nutrients. However, many problems may arise that can lead to potentially serious health issues such as Coronary Artery Disease, Myocardial Infarction, Dissection, Aneurysms and many more. Sometimes, surgery like Percutaneous Coronary Intervention is needed to fix a blocked artery. The surgery needs an intensive planning to prevent perioperative complications. To perform a proper planning the anatomy of coronary arteries needs to be studied. From the technical point of view automatic segmentation of coronaries can provide information related to the 3D anatomy and is essential for the diagnosis and quantification of coronary artery disease. In clinical practice the manual and semi-automatic segmentation of coronary arteries for quantification of narrowing part and treatment planning is still being used. In this study, the 3D Dense-U-Net architecture was proposed for precise segmentation of the coronary arteries. Additionally, Mixed Realty solution was introduced for pathology assessment during planning stage. The dataset consists of 1000 computed tomography angiography of patients. The dataset was divided in three subsets. Training, validation and test sets in ratio 70 : 15 : 15. The segmentation results were evaluated using Dice and Hausdorff metrics. The proposed method achieved 0.8206 and 22.06mm respectively. The segmentation results are combined with the Computed Tomography volumetric rendering in Mixed Reality. The operator wearing the head-mounted display can assess the anatomy of coronary arteries, analyse the potential blockages and evaluate his/her approach for the surgery during preoperative planning and also intraoperatively.