Poster 20 - Anatomical Reconstruction of the Aorta from an Echocardiogram

Saturday, October 26, 2024

10:00 AM - 12:00 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Jared D. Rosen (he/him/his)

Student
Massachusetts Academy of Math and Science at WPI
Southborough, Massachusetts, United States

Co-Author(s)

BS

Biao Si

Research Professor
Worcester Polytechnic Institute, United States

Primary Investigator(s)

Zhenglun Alan Wei, PhD

Assistant Professor
Worcester Polytechnic Institute, United States

Introduction: Cardiovascular flow simulation involves using computational and mathematical models to replicate blood flow within the cardiovascular system. While the precision of these simulations is continually being refined, the accuracy of reconstructing three-dimensional models of blood vessels is equally vital. Although significant progress has been made in developing models for the aorta and similar cardiovascular structures, there remains a critical need for accurate, patient-specific models, especially in cases where the patient has an abnormal anatomical shape. This study introduces a method that allows medical professionals to efficiently and accurately create a personalized model of a patient's aorta and extending arteries using a two-dimensional echocardiogram scan.

Materials and

Methods: Our method begins by developing a Python script that utilizes OpenCV, an artificial intelligence image recognition library, to detect the aorta's contours against the background of the image. These contour images, along with the original echocardiogram scan as a reference, are then used by a medical professional in 3D Slicer, an open-access, multi-purpose software for three-dimensional medical image modeling. The professional uses 3D Slicer to identify a two-dimensional segment that matches the shape of the aorta in the MRI scan.

With this segment, the medical professional can extract the coordinates and radii values at each point along the aorta's centerline using the VMTK (Vascular Modeling Toolkit) extension within 3D Slicer. These coordinates and radii values are then exported as a CSV file. This file is subsequently re-imported into our Python script, which generates a three-dimensional representation of the aorta by simulating it as a series of circles with linearly varying radii, forming a sequence of obliquely truncated cones replicating the aorta's shape. The resulting three-dimensional image is then re-imported into 3D Slicer to produce a virtual model as a stereolithographic (STL) file.

Results, Conclusions, and Discussions: This process was applied to echocardiogram images from two healthy subjects. As the procedure was carried out on these images, the entire process was documented and compiled into a step-by-step tutorial as a presentation. To evaluate the method's accuracy, the resulting models were compared to a reference model created by a medical expert. The high degree of similarity between the models provided strong evidence that this approach successfully reconstructs an accurate model of a patient’s cardiovascular anatomy from an echocardiogram.

Although the process was tested on the patients’ aortae, its intended and most likely application is to model multiple vessels in addition to the aorta. This would allow for the creation of a complete model of the blood vessels comprising a patient’s central cardiovascular system, which could then be used in various simulated tests available through the wide range of cardiovascular and anatomical simulation software.

In developing this process, an innovative code was created that converts any line represented as a list of coordinates with corresponding radii into a cylinder-like structure. This software has potential applications beyond cardiovascular flow simulation, where the anatomical reconstruction of cylindrical-shaped systems, tissues, and organs is valuable. It could also be useful in other fields such as mechanical, material, and aerospace engineering, where precision in modeling is crucial for system design.

Overall, this process represents an important advancement in personalized treatment and medicine, as technology continues to enhance the ability of medical professionals to understand and model biological processes more accurately.

Acknowledgements (Optional): The authors would like to acknowledge OpenCV, 3D Slicer, and the VMTK extension within 3D Slicer as free and open-source software utilized for the purposes of this study.

References (Optional): Bradski, G. (2000). The OpenCV library. Dr. Dobb's Journal of Software Tools. https://opencv.org/

Fedorov A., Beichel R., Kalpathy-Cramer J., Finet J., Fillion-Robin J-C., Pujol S., Bauer C., Jennings D., Fennessy F.M., Sonka M., Buatti J., Aylward S.R., Miller J.V., Pieper S., Kikinis R. 3D Slicer as an Image Computing Platform for the Quantitative Imaging Network. Magnetic Resonance Imaging. 2012 Nov;30(9):1323-41. PMID: 22770690. PMCID: PMC3466397.

Green, B. (2002) Canny Edge Detection in OpenCV (Version 4.10.0 dev) [Tutorial code]. https://docs.opencv.org/4.x/da/d22/tutorial_py_canny.html

PerkLab Research. (2020, June 6) Vessel segmentation and centerline extraction using 3D Slicer and VMTK [Video]. YouTube. https://www.youtube.com/watch?v=caEuwJ7pCWs

Pujol, S., & Pieper, S. (2020, April 14). Slicer programming tutorial. Slicerprogrammingtutorial. https://spujol.github.io/SlicerProgrammingTutorial/