Potential of high dimensional radiomic features to assess blood components in intraaortic vessels in non-contrast CT scans.
Anemia
Artificial intelligence
Blood
CT
Radiomics
Journal
BMC medical imaging
ISSN: 1471-2342
Titre abrégé: BMC Med Imaging
Pays: England
ID NLM: 100968553
Informations de publication
Date de publication:
12 08 2021
12 08 2021
Historique:
received:
03
04
2021
accepted:
30
07
2021
entrez:
13
8
2021
pubmed:
14
8
2021
medline:
22
1
2022
Statut:
epublish
Résumé
To assess the potential of radiomic features to quantify components of blood in intraaortic vessels to non-invasively predict moderate-to-severe anemia in non-contrast enhanced CT scans. One hundred patients (median age, 69 years; range, 19-94 years) who received CT scans of the thoracolumbar spine and blood-testing for hemoglobin and hematocrit levels ± 24 h between 08/2018 and 11/2019 were retrospectively included. Intraaortic blood was segmented using a spherical volume of interest of 1 cm diameter with consecutive radiomic analysis applying PyRadiomics software. Feature selection was performed applying analysis of correlation and collinearity. The final feature set was obtained to differentiate moderate-to-severe anemia. Random forest machine learning was applied and predictive performance was assessed. A decision-tree was obtained to propose a cut-off value of CT Hounsfield units (HU). High correlation with hemoglobin and hematocrit levels was shown for first-order radiomic features (p < 0.001 to p = 0.032). The top 3 features showed high correlation to hemoglobin values (p) and minimal collinearity (r) to the top ranked feature Median (p < 0.001), Energy (p = 0.002, r = 0.387), Minimum (p = 0.032, r = 0.437). Median (p < 0.001) and Minimum (p = 0.003) differed in moderate-to-severe anemia compared to non-anemic state. Median yielded superiority to the combination of Median and Minimum (p(AUC) = 0.015, p(precision) = 0.017, p(accuracy) = 0.612) in the predictive performance employing random forest analysis. A Median HU value ≤ 36.5 indicated moderate-to-severe anemia (accuracy = 0.90, precision = 0.80). First-order radiomic features correlate with hemoglobin levels and may be feasible for the prediction of moderate-to-severe anemia. High dimensional radiomic features did not aid augmenting the data in our exemplary use case of intraluminal blood component assessment. Trial registration Retrospectively registered.
Sections du résumé
BACKGROUND
To assess the potential of radiomic features to quantify components of blood in intraaortic vessels to non-invasively predict moderate-to-severe anemia in non-contrast enhanced CT scans.
METHODS
One hundred patients (median age, 69 years; range, 19-94 years) who received CT scans of the thoracolumbar spine and blood-testing for hemoglobin and hematocrit levels ± 24 h between 08/2018 and 11/2019 were retrospectively included. Intraaortic blood was segmented using a spherical volume of interest of 1 cm diameter with consecutive radiomic analysis applying PyRadiomics software. Feature selection was performed applying analysis of correlation and collinearity. The final feature set was obtained to differentiate moderate-to-severe anemia. Random forest machine learning was applied and predictive performance was assessed. A decision-tree was obtained to propose a cut-off value of CT Hounsfield units (HU).
RESULTS
High correlation with hemoglobin and hematocrit levels was shown for first-order radiomic features (p < 0.001 to p = 0.032). The top 3 features showed high correlation to hemoglobin values (p) and minimal collinearity (r) to the top ranked feature Median (p < 0.001), Energy (p = 0.002, r = 0.387), Minimum (p = 0.032, r = 0.437). Median (p < 0.001) and Minimum (p = 0.003) differed in moderate-to-severe anemia compared to non-anemic state. Median yielded superiority to the combination of Median and Minimum (p(AUC) = 0.015, p(precision) = 0.017, p(accuracy) = 0.612) in the predictive performance employing random forest analysis. A Median HU value ≤ 36.5 indicated moderate-to-severe anemia (accuracy = 0.90, precision = 0.80).
CONCLUSIONS
First-order radiomic features correlate with hemoglobin levels and may be feasible for the prediction of moderate-to-severe anemia. High dimensional radiomic features did not aid augmenting the data in our exemplary use case of intraluminal blood component assessment. Trial registration Retrospectively registered.
Identifiants
pubmed: 34384385
doi: 10.1186/s12880-021-00654-9
pii: 10.1186/s12880-021-00654-9
pmc: PMC8359593
doi:
Substances chimiques
Hemoglobins
0
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
123Informations de copyright
© 2021. The Author(s).
Références
Can Assoc Radiol J. 2003 Feb;54(1):26-30
pubmed: 12625080
Eur J Radiol. 2012 Dec;81(12):4196-202
pubmed: 22889590
J Clin Pathol. 1996 Apr;49(4):271-4
pubmed: 8655699
Theranostics. 2019 Feb 12;9(5):1303-1322
pubmed: 30867832
J Trauma. 2007 Aug;63(2):312-5
pubmed: 17693829
Cancer Res. 2017 Nov 1;77(21):e104-e107
pubmed: 29092951
Nat Commun. 2014 Jun 03;5:4006
pubmed: 24892406
Sci Rep. 2019 Jul 1;9(1):9441
pubmed: 31263116
Hematol Oncol Clin North Am. 2016 Apr;30(2):247-308
pubmed: 27040955
Blood Transfus. 2020 Dec 22;:
pubmed: 33370231
CA Cancer J Clin. 2019 Mar;69(2):127-157
pubmed: 30720861
Sci Rep. 2015 Aug 17;5:13087
pubmed: 26278466
Public Health Nutr. 2009 Apr;12(4):444-54
pubmed: 18498676
Sci Rep. 2015 Jun 05;5:11044
pubmed: 26251068
PLoS One. 2016 Nov 15;11(11):e0166635
pubmed: 27846276
Educ Health (Abingdon). 2003 Jul;16(2):230
pubmed: 14741909
Can Assoc Radiol J. 2013 Feb;64(1):46-50
pubmed: 22397828
Radiology. 2016 Feb;278(2):563-77
pubmed: 26579733
Transfus Med Rev. 2017 Jan;31(1):62-71
pubmed: 27317382
AJR Am J Roentgenol. 2005 Nov;185(5):1240-4
pubmed: 16247142
Magn Reson Imaging. 2012 Nov;30(9):1323-41
pubmed: 22770690
PLoS One. 2016 Dec 29;11(12):e0166550
pubmed: 28033372
Med Phys. 2017 Mar;44(3):1050-1062
pubmed: 28112418
J Clin Monit Comput. 2018 Dec;32(6):1025-1031
pubmed: 29335914
Magn Reson Imaging. 2012 Nov;30(9):1234-48
pubmed: 22898692
Adv Ther. 2021 Jan;38(1):201-225
pubmed: 33247314
Phys Med Biol. 2016 Jul 7;61(13):R150-66
pubmed: 27269645
Eur J Cancer. 2012 Mar;48(4):441-6
pubmed: 22257792
Sci Rep. 2018 Aug 29;8(1):13047
pubmed: 30158540
J Infus Nurs. 2012 Mar-Apr;35(2):84-91
pubmed: 22382792
Radiology. 1976 Dec;121(3 Pt. 1):635-40
pubmed: 981659
Lung Cancer. 2020 Aug;146:197-208
pubmed: 32563015
Lung Cancer. 2018 Jan;115:34-41
pubmed: 29290259
Ulster Med J. 2001 Nov;70(2):116-8
pubmed: 11795761
Nat Rev Clin Oncol. 2017 Dec;14(12):749-762
pubmed: 28975929