Accurate reconstruction of 3D cardiac geometry from coarsely-sliced MRI

Jordan Ringenberg, Makarand Deo, Vijay Devabhaktuni, Omer Berenfeld, Brett Snyder, Pamela Boyers, Jeffrey Gold

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

We present a comprehensive validation analysis to assess the geometric impact of using coarsely-sliced short-axis images to reconstruct patient-specific cardiac geometry. The methods utilize high-resolution diffusion tensor MRI (DTMRI) datasets as reference geometries from which synthesized coarsely-sliced datasets simulating in vivo MRI were produced. 3D models are reconstructed from the coarse data using variational implicit surfaces through a commonly used modeling tool, CardioViz3D. The resulting geometries were then compared to the reference DTMRI models from which they were derived to analyze how well the synthesized geometries approximate the reference anatomy. Averaged over seven hearts, 95% spatial overlap, less than 3% volume variability, and normal-to-surface distance of 0.32. mm was observed between the synthesized myocardial geometries reconstructed from 8. mm sliced images and the reference data. The results provide strong supportive evidence to validate the hypothesis that coarsely-sliced MRI may be used to accurately reconstruct geometric ventricular models. Furthermore, the use of DTMRI for validation of in vivo MRI presents a novel benchmark procedure for studies which aim to substantiate their modeling and simulation methods using coarsely-sliced cardiac data. In addition, the paper outlines a suggested original procedure for deriving image-based ventricular models using the CardioViz3D software.

Original languageEnglish (US)
Pages (from-to)483-493
Number of pages11
JournalComputer Methods and Programs in Biomedicine
Volume113
Issue number2
DOIs
StatePublished - Feb 1 2014

Fingerprint

Magnetic resonance imaging
Diffusion Magnetic Resonance Imaging
Geometry
Tensors
Benchmarking
Anatomy
Software
Datasets

Keywords

  • 3D modeling
  • Cardiac MRI
  • Patient-specific heart models
  • Variational implicit surfaces
  • Ventricular geometry

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Health Informatics

Cite this

Accurate reconstruction of 3D cardiac geometry from coarsely-sliced MRI. / Ringenberg, Jordan; Deo, Makarand; Devabhaktuni, Vijay; Berenfeld, Omer; Snyder, Brett; Boyers, Pamela; Gold, Jeffrey.

In: Computer Methods and Programs in Biomedicine, Vol. 113, No. 2, 01.02.2014, p. 483-493.

Research output: Contribution to journalArticle

Ringenberg, Jordan ; Deo, Makarand ; Devabhaktuni, Vijay ; Berenfeld, Omer ; Snyder, Brett ; Boyers, Pamela ; Gold, Jeffrey. / Accurate reconstruction of 3D cardiac geometry from coarsely-sliced MRI. In: Computer Methods and Programs in Biomedicine. 2014 ; Vol. 113, No. 2. pp. 483-493.
@article{a9e952e4a07b45a1aad84c32ecd13977,
title = "Accurate reconstruction of 3D cardiac geometry from coarsely-sliced MRI",
abstract = "We present a comprehensive validation analysis to assess the geometric impact of using coarsely-sliced short-axis images to reconstruct patient-specific cardiac geometry. The methods utilize high-resolution diffusion tensor MRI (DTMRI) datasets as reference geometries from which synthesized coarsely-sliced datasets simulating in vivo MRI were produced. 3D models are reconstructed from the coarse data using variational implicit surfaces through a commonly used modeling tool, CardioViz3D. The resulting geometries were then compared to the reference DTMRI models from which they were derived to analyze how well the synthesized geometries approximate the reference anatomy. Averaged over seven hearts, 95{\%} spatial overlap, less than 3{\%} volume variability, and normal-to-surface distance of 0.32. mm was observed between the synthesized myocardial geometries reconstructed from 8. mm sliced images and the reference data. The results provide strong supportive evidence to validate the hypothesis that coarsely-sliced MRI may be used to accurately reconstruct geometric ventricular models. Furthermore, the use of DTMRI for validation of in vivo MRI presents a novel benchmark procedure for studies which aim to substantiate their modeling and simulation methods using coarsely-sliced cardiac data. In addition, the paper outlines a suggested original procedure for deriving image-based ventricular models using the CardioViz3D software.",
keywords = "3D modeling, Cardiac MRI, Patient-specific heart models, Variational implicit surfaces, Ventricular geometry",
author = "Jordan Ringenberg and Makarand Deo and Vijay Devabhaktuni and Omer Berenfeld and Brett Snyder and Pamela Boyers and Jeffrey Gold",
year = "2014",
month = "2",
day = "1",
doi = "10.1016/j.cmpb.2013.11.013",
language = "English (US)",
volume = "113",
pages = "483--493",
journal = "Computer Methods and Programs in Biomedicine",
issn = "0169-2607",
publisher = "Elsevier Ireland Ltd",
number = "2",

}

TY - JOUR

T1 - Accurate reconstruction of 3D cardiac geometry from coarsely-sliced MRI

AU - Ringenberg, Jordan

AU - Deo, Makarand

AU - Devabhaktuni, Vijay

AU - Berenfeld, Omer

AU - Snyder, Brett

AU - Boyers, Pamela

AU - Gold, Jeffrey

PY - 2014/2/1

Y1 - 2014/2/1

N2 - We present a comprehensive validation analysis to assess the geometric impact of using coarsely-sliced short-axis images to reconstruct patient-specific cardiac geometry. The methods utilize high-resolution diffusion tensor MRI (DTMRI) datasets as reference geometries from which synthesized coarsely-sliced datasets simulating in vivo MRI were produced. 3D models are reconstructed from the coarse data using variational implicit surfaces through a commonly used modeling tool, CardioViz3D. The resulting geometries were then compared to the reference DTMRI models from which they were derived to analyze how well the synthesized geometries approximate the reference anatomy. Averaged over seven hearts, 95% spatial overlap, less than 3% volume variability, and normal-to-surface distance of 0.32. mm was observed between the synthesized myocardial geometries reconstructed from 8. mm sliced images and the reference data. The results provide strong supportive evidence to validate the hypothesis that coarsely-sliced MRI may be used to accurately reconstruct geometric ventricular models. Furthermore, the use of DTMRI for validation of in vivo MRI presents a novel benchmark procedure for studies which aim to substantiate their modeling and simulation methods using coarsely-sliced cardiac data. In addition, the paper outlines a suggested original procedure for deriving image-based ventricular models using the CardioViz3D software.

AB - We present a comprehensive validation analysis to assess the geometric impact of using coarsely-sliced short-axis images to reconstruct patient-specific cardiac geometry. The methods utilize high-resolution diffusion tensor MRI (DTMRI) datasets as reference geometries from which synthesized coarsely-sliced datasets simulating in vivo MRI were produced. 3D models are reconstructed from the coarse data using variational implicit surfaces through a commonly used modeling tool, CardioViz3D. The resulting geometries were then compared to the reference DTMRI models from which they were derived to analyze how well the synthesized geometries approximate the reference anatomy. Averaged over seven hearts, 95% spatial overlap, less than 3% volume variability, and normal-to-surface distance of 0.32. mm was observed between the synthesized myocardial geometries reconstructed from 8. mm sliced images and the reference data. The results provide strong supportive evidence to validate the hypothesis that coarsely-sliced MRI may be used to accurately reconstruct geometric ventricular models. Furthermore, the use of DTMRI for validation of in vivo MRI presents a novel benchmark procedure for studies which aim to substantiate their modeling and simulation methods using coarsely-sliced cardiac data. In addition, the paper outlines a suggested original procedure for deriving image-based ventricular models using the CardioViz3D software.

KW - 3D modeling

KW - Cardiac MRI

KW - Patient-specific heart models

KW - Variational implicit surfaces

KW - Ventricular geometry

UR - http://www.scopus.com/inward/record.url?scp=84892809824&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84892809824&partnerID=8YFLogxK

U2 - 10.1016/j.cmpb.2013.11.013

DO - 10.1016/j.cmpb.2013.11.013

M3 - Article

C2 - 24345413

AN - SCOPUS:84892809824

VL - 113

SP - 483

EP - 493

JO - Computer Methods and Programs in Biomedicine

JF - Computer Methods and Programs in Biomedicine

SN - 0169-2607

IS - 2

ER -