Clinical validation of an AI-based automatic quantification tool for lung lobes in SPECT/CT.

AI-based segmentation Lobar quantification Perfusion SPECT/CT Ventilation SPECT/CT

Journal

EJNMMI physics
ISSN: 2197-7364
Titre abrégé: EJNMMI Phys
Pays: Germany
ID NLM: 101658952

Informations de publication

Date de publication:
21 Sep 2023
Historique:
received: 05 12 2022
accepted: 05 09 2023
medline: 21 9 2023
pubmed: 21 9 2023
entrez: 21 9 2023
Statut: epublish

Résumé

Lung lobar ventilation and perfusion (V/Q) quantification is generally obtained by generating planar scintigraphy images and then imposing three equally sized regions of interest on the data of each lung. This method is fast but not as accurate as SPECT/CT imaging, which provides three-dimensional data and therefore allows more precise lobar quantification. However, the manual delineation of each lobe is time-consuming, which makes SPECT/CT incompatible with the clinical workflow for V/Q estimation. An alternative may be to use artificial intelligence-based auto-segmentation tools such as AutoLung3D (Siemens Healthineers, Knoxville, USA), which automatically delineate the lung lobes on the CT data acquired with the SPECT data. The present study assessed the clinical validity of this approach relative to planar scintigraphy and manual quantification in SPECT/CT. The Autolung3D software was tested on the retrospective SPECT/CT data of 43 patients who underwent V/Q scintigraphy with The three methods provided similar V/Q estimates for the left lung lobes and total lungs. However, compared to the manual SPECT/CT method, planar scintigraphy yielded significantly higher estimates for the middle right lobe and significantly lower estimates for the superior and inferior right lobes. The estimates of the manual and automated SPECT/CT methods were similar. However, the post-processing time in the automated method was approximately 5 min compared to 2 h for the manual method. Moreover, the automated method associated with a drastic reduction in interobserver variability: Its maximal relative standard deviation was only 5%, compared to 23% for planar scintigraphy and 19% for the manual SPECT/CT method. This study validated the AutoLung3D software for general clinical use since it rapidly provides accurate lobar quantification in V/Q scans with markedly less interobserver variability than planar scintigraphy or the manual SPECT/CT method.

Sections du résumé

BACKGROUND BACKGROUND
Lung lobar ventilation and perfusion (V/Q) quantification is generally obtained by generating planar scintigraphy images and then imposing three equally sized regions of interest on the data of each lung. This method is fast but not as accurate as SPECT/CT imaging, which provides three-dimensional data and therefore allows more precise lobar quantification. However, the manual delineation of each lobe is time-consuming, which makes SPECT/CT incompatible with the clinical workflow for V/Q estimation. An alternative may be to use artificial intelligence-based auto-segmentation tools such as AutoLung3D (Siemens Healthineers, Knoxville, USA), which automatically delineate the lung lobes on the CT data acquired with the SPECT data. The present study assessed the clinical validity of this approach relative to planar scintigraphy and manual quantification in SPECT/CT.
METHODS METHODS
The Autolung3D software was tested on the retrospective SPECT/CT data of 43 patients who underwent V/Q scintigraphy with
RESULTS RESULTS
The three methods provided similar V/Q estimates for the left lung lobes and total lungs. However, compared to the manual SPECT/CT method, planar scintigraphy yielded significantly higher estimates for the middle right lobe and significantly lower estimates for the superior and inferior right lobes. The estimates of the manual and automated SPECT/CT methods were similar. However, the post-processing time in the automated method was approximately 5 min compared to 2 h for the manual method. Moreover, the automated method associated with a drastic reduction in interobserver variability: Its maximal relative standard deviation was only 5%, compared to 23% for planar scintigraphy and 19% for the manual SPECT/CT method.
CONCLUSIONS CONCLUSIONS
This study validated the AutoLung3D software for general clinical use since it rapidly provides accurate lobar quantification in V/Q scans with markedly less interobserver variability than planar scintigraphy or the manual SPECT/CT method.

Identifiants

pubmed: 37733103
doi: 10.1186/s40658-023-00578-z
pii: 10.1186/s40658-023-00578-z
pmc: PMC10513978
doi:

Types de publication

Journal Article

Langues

eng

Pagination

57

Informations de copyright

© 2023. Springer Nature Switzerland AG.

Références

J Nucl Med. 2001 Aug;42(8):1288-94
pubmed: 11483693
Nuklearmedizin. 1990 Dec;29(6):274-7
pubmed: 2075089
J Nucl Med Technol. 2017 Sep;45(3):185-192
pubmed: 28408698
Nucl Med Commun. 2008 Apr;29(4):323-30
pubmed: 18317295
Chest. 2013 May;143(5 Suppl):e166S-e190S
pubmed: 23649437
Cardiovasc Diagn Ther. 2017 Jun;7(3):317-324
pubmed: 28567357
J Nucl Med Technol. 2000 Dec;28(4):233-44
pubmed: 11142324
Respirology. 2010 Oct;15(7):1079-83
pubmed: 20636308
Lung Cancer. 2018 Apr;118:155-160
pubmed: 29571995
J Thorac Cardiovasc Surg. 2014 Nov;148(5):2345-52
pubmed: 24882061
J Nucl Med. 2013 Sep;54(9):1588-96
pubmed: 23907760
Eur Respir J. 2015 Jan;45(1):262-5
pubmed: 25359337
Eur J Nucl Med Mol Imaging. 2019 Nov;46(12):2429-2451
pubmed: 31410539
Ann Nucl Med. 2008 Jun;22(5):437-45
pubmed: 18600424
Respiration. 2017;93(2):138-150
pubmed: 27992862
J Nucl Med. 2002 Oct;43(10):1343-58
pubmed: 12368373
Chest. 2021 May;159(5):1833-1842
pubmed: 33345947

Auteurs

Emilie Verrecchia-Ramos (E)

Department of Medical Physics, Mercy Hospital, CHR Metz-Thionville, 1, Allée du Château, 57530, Ars-Laquenexy, France. e.verrecchiaramos@chr-metz-thionville.fr.

Olivier Morel (O)

Department of Nuclear Medicine, Mercy Hospital, CHR Metz-Thionville, 1, Allée du Château, 57530, Ars-Laquenexy, France.

Merwan Ginet (M)

Department of Nuclear Medicine, Mercy Hospital, CHR Metz-Thionville, 1, Allée du Château, 57530, Ars-Laquenexy, France.

Paul Retif (P)

Department of Medical Physics, Mercy Hospital, CHR Metz-Thionville, 1, Allée du Château, 57530, Ars-Laquenexy, France.
CNRS, CRAN, Université de Lorraine, 54000, Nancy, France.

Sinan Ben Mahmoud (S)

Department of Nuclear Medicine, Mercy Hospital, CHR Metz-Thionville, 1, Allée du Château, 57530, Ars-Laquenexy, France.

Classifications MeSH