Impact of defacing on automated brain atrophy estimation.
Atrophy
Brain
De-identification
Magnetic resonance imaging
Privacy
Journal
Insights into imaging
ISSN: 1869-4101
Titre abrégé: Insights Imaging
Pays: Germany
ID NLM: 101532453
Informations de publication
Date de publication:
26 Mar 2022
26 Mar 2022
Historique:
received:
20
09
2021
accepted:
19
02
2022
entrez:
29
3
2022
pubmed:
30
3
2022
medline:
30
3
2022
Statut:
epublish
Résumé
Defacing has become mandatory for anonymization of brain MRI scans; however, concerns regarding data integrity were raised. Thus, we systematically evaluated the effect of different defacing procedures on automated brain atrophy estimation. In total, 268 Alzheimer's disease patients were included from ADNI, which included unaccelerated (n = 154), within-session unaccelerated repeat (n = 67) and accelerated 3D T1 imaging (n = 114). Atrophy maps were computed using the open-source software veganbagel for every original, unmodified scan and after defacing using afni_refacer, fsl_deface, mri_deface, mri_reface, PyDeface or spm_deface, and the root-mean-square error (RMSE) between z-scores was calculated. RMSE values derived from unaccelerated and unaccelerated repeat imaging served as a benchmark. Outliers were defined as RMSE > 75th percentile and by using Grubbs's test. Benchmark RMSE was 0.28 ± 0.1 (range 0.12-0.58, 75th percentile 0.33). Outliers were found for unaccelerated and accelerated T1 imaging using the 75th percentile cutoff: afni_refacer (unaccelerated: 18, accelerated: 16), fsl_deface (unaccelerated: 4, accelerated: 18), mri_deface (unaccelerated: 0, accelerated: 15), mri_reface (unaccelerated: 0, accelerated: 2) and spm_deface (unaccelerated: 0, accelerated: 7). PyDeface performed best with no outliers (unaccelerated mean RMSE 0.08 ± 0.05, accelerated mean RMSE 0.07 ± 0.05). The following outliers were found according to Grubbs's test: afni_refacer (unaccelerated: 16, accelerated: 13), fsl_deface (unaccelerated: 10, accelerated: 21), mri_deface (unaccelerated: 7, accelerated: 20), mri_reface (unaccelerated: 7, accelerated: 6), PyDeface (unaccelerated: 5, accelerated: 8) and spm_deface (unaccelerated: 10, accelerated: 12). Most defacing approaches have an impact on atrophy estimation, especially in accelerated 3D T1 imaging. Only PyDeface showed good results with negligible impact on atrophy estimation.
Sections du résumé
BACKGROUND
BACKGROUND
Defacing has become mandatory for anonymization of brain MRI scans; however, concerns regarding data integrity were raised. Thus, we systematically evaluated the effect of different defacing procedures on automated brain atrophy estimation.
METHODS
METHODS
In total, 268 Alzheimer's disease patients were included from ADNI, which included unaccelerated (n = 154), within-session unaccelerated repeat (n = 67) and accelerated 3D T1 imaging (n = 114). Atrophy maps were computed using the open-source software veganbagel for every original, unmodified scan and after defacing using afni_refacer, fsl_deface, mri_deface, mri_reface, PyDeface or spm_deface, and the root-mean-square error (RMSE) between z-scores was calculated. RMSE values derived from unaccelerated and unaccelerated repeat imaging served as a benchmark. Outliers were defined as RMSE > 75th percentile and by using Grubbs's test.
RESULTS
RESULTS
Benchmark RMSE was 0.28 ± 0.1 (range 0.12-0.58, 75th percentile 0.33). Outliers were found for unaccelerated and accelerated T1 imaging using the 75th percentile cutoff: afni_refacer (unaccelerated: 18, accelerated: 16), fsl_deface (unaccelerated: 4, accelerated: 18), mri_deface (unaccelerated: 0, accelerated: 15), mri_reface (unaccelerated: 0, accelerated: 2) and spm_deface (unaccelerated: 0, accelerated: 7). PyDeface performed best with no outliers (unaccelerated mean RMSE 0.08 ± 0.05, accelerated mean RMSE 0.07 ± 0.05). The following outliers were found according to Grubbs's test: afni_refacer (unaccelerated: 16, accelerated: 13), fsl_deface (unaccelerated: 10, accelerated: 21), mri_deface (unaccelerated: 7, accelerated: 20), mri_reface (unaccelerated: 7, accelerated: 6), PyDeface (unaccelerated: 5, accelerated: 8) and spm_deface (unaccelerated: 10, accelerated: 12).
CONCLUSION
CONCLUSIONS
Most defacing approaches have an impact on atrophy estimation, especially in accelerated 3D T1 imaging. Only PyDeface showed good results with negligible impact on atrophy estimation.
Identifiants
pubmed: 35348936
doi: 10.1186/s13244-022-01195-7
pii: 10.1186/s13244-022-01195-7
pmc: PMC8964867
doi:
Types de publication
Journal Article
Langues
eng
Pagination
54Subventions
Organisme : NIA NIH HHS
ID : U01 AG024904
Pays : United States
Organisme : Deutsche Forschungsgemeinschaft
ID : DA 2167/1-1
Informations de copyright
© 2022. The Author(s).
Références
Magn Reson Med. 2002 Jun;47(6):1202-10
pubmed: 12111967
Neuroimage Clin. 2018;20:851-860
pubmed: 30278372
J Magn Reson Imaging. 2012 Jul;36(1):55-72
pubmed: 22696125
N Engl J Med. 2019 Oct 24;381(17):1684-1686
pubmed: 31644852
J Digit Imaging. 2016 Jun;29(3):365-79
pubmed: 26628083
Eur Radiol. 2020 Feb;30(2):1062-1074
pubmed: 31691120
Arch Neurol. 2006 Jan;63(1):81-6
pubmed: 16401739
Mov Disord. 2009 Jan 15;24(1):99-103
pubmed: 18823048
Neuroimaging Clin N Am. 2005 Nov;15(4):869-77, xi-xii
pubmed: 16443497
Eur Radiol. 2020 May;30(5):2821-2829
pubmed: 32002640
Behav Neurol. 2000;12(4):191-200
pubmed: 11568431
J Neuroradiol. 2021 Mar 13;:
pubmed: 33727023
Prog Mol Biol Transl Sci. 2019;165:321-369
pubmed: 31481169
J Alzheimers Dis. 2021;83(1):281-290
pubmed: 34308908
Eur Radiol. 2021 Feb;31(2):1043-1048
pubmed: 32852588
Hum Brain Mapp. 2007 Sep;28(9):892-903
pubmed: 17295313
Neuroimage. 2018 Feb 1;166:400-424
pubmed: 29079522
Neuroimage. 2015 Jun;113:61-9
pubmed: 25797830
Magn Reson Med. 1999 Nov;42(5):952-62
pubmed: 10542355
Eur Neurol. 1997;37(2):95-9
pubmed: 9058064
Neuroimage. 2013 Oct 15;80:62-79
pubmed: 23684880
Comput Biomed Res. 1996 Jun;29(3):162-73
pubmed: 8812068
Nat Rev Neurol. 2020 Mar;16(3):171-182
pubmed: 32094485
Neuroimage. 2015 Feb 15;107:46-53
pubmed: 25481794
Neuroimage. 2021 May 1;231:117845
pubmed: 33582276
Front Psychiatry. 2021 Feb 24;12:617997
pubmed: 33716819
Alzheimers Res Ther. 2018 May 24;10(1):46
pubmed: 29793546
Front Neurosci. 2012 Oct 16;6:152
pubmed: 23087608
Neuroinformatics. 2013 Jan;11(1):65-75
pubmed: 22968671
IEEE Trans Inf Technol Biomed. 2009 Jan;13(1):5-9
pubmed: 19129018
Neuroinformatics. 2017 Apr;15(2):215-226
pubmed: 28316055