Measurement of Abdominal Aortic Aneurysm Strain Using MR Deformable Image Registration: Accuracy and Relationship to Recent Aneurysm Progression.


Journal

Investigative radiology
ISSN: 1536-0210
Titre abrégé: Invest Radiol
Pays: United States
ID NLM: 0045377

Informations de publication

Date de publication:
19 Oct 2023
Historique:
pmc-release: 19 04 2025
medline: 19 10 2023
pubmed: 19 10 2023
entrez: 19 10 2023
Statut: aheadofprint

Résumé

Management of asymptomatic abdominal aortic aneurysm (AAA) based on maximum aneurysm diameter and growth rate fails to preempt many ruptures. Assessment of aortic wall biomechanical properties may improve assessment of progression and rupture risk. This study aimed to assess the accuracy of AAA wall strain measured by cine magnetic resonance imaging (MRI) deformable image registration (MR strain) and investigate its relationship with recent AAA progression. The MR strain accuracy was evaluated in silico against ground truth strain in 54 synthetic MRIs generated from a finite element model simulation of an AAA patient's abdomen for different aortic pulse pressures, tissue motions, signal intensity variations, and image noise. Evaluation included bias with 95% confidence interval (CI) and correlation analysis. Association of MR strain with AAA growth rate was assessed in 25 consecutive patients with >6 months of prior surveillance, for whom cine balanced steady-state free-precession imaging was acquired at the level of the AAA as well as the proximal, normal-caliber aorta. Univariate and multivariate regressions were used to associate growth rate with clinical variables, maximum AAA diameter (Dmax), and peak circumferential MR strain through the cardiac cycle. The MR strain interoperator variability was assessed using bias with 95% CI, intraclass correlation coefficient, and coefficient of variation. In silico experiments revealed an MR strain bias of 0.48% ± 0.42% and a slope of correlation to ground truth strain of 0.963. In vivo, AAA MR strain (1.2% ± 0.6%) was highly reproducible (bias ± 95% CI, 0.03% ± 0.31%; intraclass correlation coefficient, 97.8%; coefficient of variation, 7.14%) and was lower than in the nonaneurysmal aorta (2.4% ± 1.7%). Dmax (β = 0.087) and MR strain (β= -1.563) were both associated with AAA growth rate. The MR strain remained an independent factor associated with growth rate (β= -0.904) after controlling for Dmax. Deformable image registration analysis can accurately measure the circumferential strain of the AAA wall from standard cine MRI and may offer patient-specific insight regarding AAA progression.

Sections du résumé

BACKGROUND BACKGROUND
Management of asymptomatic abdominal aortic aneurysm (AAA) based on maximum aneurysm diameter and growth rate fails to preempt many ruptures. Assessment of aortic wall biomechanical properties may improve assessment of progression and rupture risk. This study aimed to assess the accuracy of AAA wall strain measured by cine magnetic resonance imaging (MRI) deformable image registration (MR strain) and investigate its relationship with recent AAA progression.
METHODS METHODS
The MR strain accuracy was evaluated in silico against ground truth strain in 54 synthetic MRIs generated from a finite element model simulation of an AAA patient's abdomen for different aortic pulse pressures, tissue motions, signal intensity variations, and image noise. Evaluation included bias with 95% confidence interval (CI) and correlation analysis. Association of MR strain with AAA growth rate was assessed in 25 consecutive patients with >6 months of prior surveillance, for whom cine balanced steady-state free-precession imaging was acquired at the level of the AAA as well as the proximal, normal-caliber aorta. Univariate and multivariate regressions were used to associate growth rate with clinical variables, maximum AAA diameter (Dmax), and peak circumferential MR strain through the cardiac cycle. The MR strain interoperator variability was assessed using bias with 95% CI, intraclass correlation coefficient, and coefficient of variation.
RESULTS RESULTS
In silico experiments revealed an MR strain bias of 0.48% ± 0.42% and a slope of correlation to ground truth strain of 0.963. In vivo, AAA MR strain (1.2% ± 0.6%) was highly reproducible (bias ± 95% CI, 0.03% ± 0.31%; intraclass correlation coefficient, 97.8%; coefficient of variation, 7.14%) and was lower than in the nonaneurysmal aorta (2.4% ± 1.7%). Dmax (β = 0.087) and MR strain (β= -1.563) were both associated with AAA growth rate. The MR strain remained an independent factor associated with growth rate (β= -0.904) after controlling for Dmax.
CONCLUSIONS CONCLUSIONS
Deformable image registration analysis can accurately measure the circumferential strain of the AAA wall from standard cine MRI and may offer patient-specific insight regarding AAA progression.

Identifiants

pubmed: 37855728
doi: 10.1097/RLI.0000000000001035
pii: 00004424-990000000-00166
pmc: PMC11026303
mid: NIHMS1928845
doi:

Types de publication

Journal Article

Langues

eng

Sous-ensembles de citation

IM

Subventions

Organisme : CSRD VA
ID : I01 CX002071
Pays : United States

Informations de copyright

Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc.

Références

Ann Vasc Surg. 2000 Mar;14(2):152-7
pubmed: 10742430
J Thorac Cardiovasc Surg. 2005 Sep;130(3):677-83
pubmed: 16153912
J Vasc Surg. 1993 Jul;18(1):74-80
pubmed: 8326662
J Ultrasound Med. 2022 May;41(5):1085-1093
pubmed: 34296470
J Cardiovasc Magn Reson. 2019 Sep 16;21(1):59
pubmed: 31522679
Magn Reson Med. 2021 Jan;85(1):357-368
pubmed: 32851707
Radiology. 2020 Mar;294(3):707-713
pubmed: 31990263
PLoS One. 2013 Dec 02;8(12):e81260
pubmed: 24312543
Vasc Med. 2016 Jun;21(3):209-16
pubmed: 26896335
Circulation. 2015 May 19;131(20):1783-95
pubmed: 25904646
J Ultrasound Med. 2021 Dec;40(12):2619-2627
pubmed: 33555036
Quant Imaging Med Surg. 2021 Feb;11(2):823-830
pubmed: 33532280
Eur J Vasc Endovasc Surg. 2017 Nov;54(5):597-603
pubmed: 28882515
Circ Cardiovasc Imaging. 2021 Dec;14(12):1112-1121
pubmed: 34875845
Eur Radiol. 2019 Jul;29(7):3658-3668
pubmed: 30770973
BMC Cardiovasc Disord. 2020 Sep 3;20(1):400
pubmed: 32883201
J Vasc Surg. 2013 May;57(5):1261-8, 1268.e1-5
pubmed: 23384493
Am J Cardiol. 2017 May 15;119(10):1542-1547
pubmed: 28335982
Ann Surg. 1999 Sep;230(3):289-96; discussion 296-7
pubmed: 10493476
J Vasc Surg. 2003 May;37(5):1106-17
pubmed: 12756363
Radiology. 2020 Jun;295(3):730-732
pubmed: 32233921
Ann Vasc Dis. 2014;7(4):393-8
pubmed: 25593624
J Vasc Surg. 2018 Jan;67(1):2-77.e2
pubmed: 29268916
Lancet. 2002 Nov 16;360(9345):1531-9
pubmed: 12443589
Eur J Vasc Endovasc Surg. 2016 Oct;52(4):475-486
pubmed: 27558090
Eur Radiol. 2017 Apr;27(4):1404-1415
pubmed: 27491873
Eur Radiol. 2019 Dec;29(12):6846-6857
pubmed: 31297633
J Magn Reson Imaging. 2018 Aug;48(2):404-414
pubmed: 29283466
Radiology. 2022 Sep;304(3):721-729
pubmed: 35638926
Exp Clin Cardiol. 2011 Spring;16(1):11-5
pubmed: 21523201
Arterioscler Thromb Vasc Biol. 2005 Aug;25(8):1558-66
pubmed: 16055757
Circulation. 1977 Sep;56(3 Suppl):II161-4
pubmed: 884821
J Vasc Surg. 2003 Jan;37(1):112-7
pubmed: 12514586
JAMA. 2019 Dec 10;322(22):2211-2218
pubmed: 31821437
J Cardiovasc Transl Res. 2012 Jun;5(3):264-73
pubmed: 22290157

Auteurs

Huiming Dong (H)

From the Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA (H.D., J.L., E.K., A.Z., C.Z., M.B., Y.J.L., D.S., M.H., D.M.); Vascular Imaging Research Center, San Francisco Veteran Affairs Medical Center, San Francisco, CA (H.D., J.L., E.K., A.Z., C.Z., M.B., D.S., M.H., D.M.); Siemens Healthineers (T.C.); Department of Radiology, University of Washington, Seattle, WA (C.Z.); Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA (F.J.); Department of Surgery, University of California, San Francisco, San Francisco, CA (J.I., W. G.); and Department of Vascular Surgery, San Francisco Veteran Affairs Medical Center, San Francisco, CA (J.I., W.G.).

Classifications MeSH