Artificial intelligence for radiological paediatric fracture assessment: a systematic review.
Artificial intelligence
Diagnostic accuracy
Fracture
Machine learning
Trauma
Journal
Insights into imaging
ISSN: 1869-4101
Titre abrégé: Insights Imaging
Pays: Germany
ID NLM: 101532453
Informations de publication
Date de publication:
03 Jun 2022
03 Jun 2022
Historique:
received:
04
03
2022
accepted:
12
05
2022
entrez:
3
6
2022
pubmed:
4
6
2022
medline:
4
6
2022
Statut:
epublish
Résumé
Majority of research and commercial efforts have focussed on use of artificial intelligence (AI) for fracture detection in adults, despite the greater long-term clinical and medicolegal implications of missed fractures in children. The objective of this study was to assess the available literature regarding diagnostic performance of AI tools for paediatric fracture assessment on imaging, and where available, how this compares with the performance of human readers. MEDLINE, Embase and Cochrane Library databases were queried for studies published between 1 January 2011 and 2021 using terms related to 'fracture', 'artificial intelligence', 'imaging' and 'children'. Risk of bias was assessed using a modified QUADAS-2 tool. Descriptive statistics for diagnostic accuracies were collated. Nine eligible articles from 362 publications were included, with most (8/9) evaluating fracture detection on radiographs, with the elbow being the most common body part. Nearly all articles used data derived from a single institution, and used deep learning methodology with only a few (2/9) performing external validation. Accuracy rates generated by AI ranged from 88.8 to 97.9%. In two of the three articles where AI performance was compared to human readers, sensitivity rates for AI were marginally higher, but this was not statistically significant. Wide heterogeneity in the literature with limited information on algorithm performance on external datasets makes it difficult to understand how such tools may generalise to a wider paediatric population. Further research using a multicentric dataset with real-world evaluation would help to better understand the impact of these tools.
Sections du résumé
BACKGROUND
BACKGROUND
Majority of research and commercial efforts have focussed on use of artificial intelligence (AI) for fracture detection in adults, despite the greater long-term clinical and medicolegal implications of missed fractures in children. The objective of this study was to assess the available literature regarding diagnostic performance of AI tools for paediatric fracture assessment on imaging, and where available, how this compares with the performance of human readers.
MATERIALS AND METHODS
METHODS
MEDLINE, Embase and Cochrane Library databases were queried for studies published between 1 January 2011 and 2021 using terms related to 'fracture', 'artificial intelligence', 'imaging' and 'children'. Risk of bias was assessed using a modified QUADAS-2 tool. Descriptive statistics for diagnostic accuracies were collated.
RESULTS
RESULTS
Nine eligible articles from 362 publications were included, with most (8/9) evaluating fracture detection on radiographs, with the elbow being the most common body part. Nearly all articles used data derived from a single institution, and used deep learning methodology with only a few (2/9) performing external validation. Accuracy rates generated by AI ranged from 88.8 to 97.9%. In two of the three articles where AI performance was compared to human readers, sensitivity rates for AI were marginally higher, but this was not statistically significant.
CONCLUSIONS
CONCLUSIONS
Wide heterogeneity in the literature with limited information on algorithm performance on external datasets makes it difficult to understand how such tools may generalise to a wider paediatric population. Further research using a multicentric dataset with real-world evaluation would help to better understand the impact of these tools.
Identifiants
pubmed: 35657439
doi: 10.1186/s13244-022-01234-3
pii: 10.1186/s13244-022-01234-3
pmc: PMC9166920
doi:
Types de publication
Journal Article
Langues
eng
Pagination
94Subventions
Organisme : Medical Research Council
ID : MR/R002118/1
Pays : United Kingdom
Organisme : National Institute for Health Research
ID : 301322
Organisme : National Institute for Health Research
ID : NIHR-CDF-2017-10-037
Informations de copyright
© 2022. The Author(s).
Références
J Child Orthop. 2010 Oct;4(5):471-6
pubmed: 21966313
Pediatr Emerg Care. 1999 Aug;15(4):245-8
pubmed: 10460076
J Pediatr Orthop. 2019 Jul;39(6):e482-e486
pubmed: 30730444
Ann Intern Med. 2011 Oct 18;155(8):529-36
pubmed: 22007046
Pediatr Radiol. 2021 May;51(6):1061-1064
pubmed: 33904953
Skeletal Radiol. 2019 Feb;48(2):239-244
pubmed: 29955910
AJR Am J Roentgenol. 2012 Oct;199(4):916-20
pubmed: 22997387
Radiology. 2016 Jan;278(1):64-73
pubmed: 26172532
J Pediatr Surg. 2021 Jun;56(6):1180-1184
pubmed: 33771371
Children (Basel). 2021 May 21;8(6):
pubmed: 34063945
Comput Biol Med. 2016 Nov 1;78:120-125
pubmed: 27684324
Orthopedics. 2019 Mar 1;42(2):e260-e267
pubmed: 30763449
J Bone Joint Surg Br. 2012 Jul;94(7):961-8
pubmed: 22733954
Arch Dis Child. 2000 Jun;82(6):452-5
pubmed: 10833175
J Pediatr Orthop. 2016 Mar;36(2):213-7
pubmed: 25705809
Arch Dis Child. 2015 Nov;100(11):1016-7
pubmed: 26194358
NPJ Digit Med. 2022 Jan 27;5(1):11
pubmed: 35087178
Acta Orthop. 2021 Oct;92(5):615-620
pubmed: 34082661
Clin Radiol. 2009 Jul;64(7):690-8
pubmed: 19520213
Neuro Oncol. 2021 Feb 25;23(2):214-225
pubmed: 33075135
Radiol Artif Intell. 2019 Jan 30;1(1):e180015
pubmed: 33937781
Pediatr Radiol. 2022 May;52(5):924-931
pubmed: 35043263
J Bone Miner Res. 2004 Dec;19(12):1976-81
pubmed: 15537440
Arch Dis Child. 2017 Feb;102(2):170-173
pubmed: 27789460
Eur Radiol. 2019 Dec;29(12):6780-6789
pubmed: 31119416
Insights Imaging. 2021 Feb 10;12(1):13
pubmed: 33564955
Pediatr Radiol. 2020 Jun;50(7):907-912
pubmed: 32166463
BMJ. 2021 Mar 29;372:n71
pubmed: 33782057
Orthop Traumatol Surg Res. 2020 Nov;106(7):1245-1249
pubmed: 33060015
Spine (Phila Pa 1976). 2007 Oct 1;32(21):2339-47
pubmed: 17906576
AI Soc. 2021 Oct 18;:1-25
pubmed: 34690449
J Am Coll Radiol. 2020 May;17(5S):S125-S137
pubmed: 32370957
Acta Orthop. 2019 Aug;90(4):394-400
pubmed: 30942136
Pediatr Radiol. 2017 Jun;47(7):808-816
pubmed: 28536766
BMC Musculoskelet Disord. 2022 Jan 28;23(1):96
pubmed: 35090422
Osteoporos Int. 2002 Dec;13(12):990-5
pubmed: 12459942
Pediatr Radiol. 2021 Jul 1;:
pubmed: 34196729
Radiol Artif Intell. 2020 Nov 11;2(6):e200004
pubmed: 33937846
Clin Radiol. 2016 Dec;71(12):1263-1267
pubmed: 27499464
Pediatr Radiol. 2022 May;52(6):1095-1103
pubmed: 35147714
Acta Orthop. 2018 Aug;89(4):468-473
pubmed: 29577791
Radiology. 2021 Jul;300(1):120-129
pubmed: 33944629
Clin Radiol. 2019 Jul;74(7):496-502
pubmed: 31126587
Korean J Radiol. 2022 Mar;23(3):343-354
pubmed: 35029078
Radiol Artif Intell. 2020 Mar 25;2(2):e200029
pubmed: 33937821
J Spinal Disord Tech. 2004 Dec;17(6):477-82
pubmed: 15570118
Clin Orthop Relat Res. 2019 Nov;477(11):2482-2491
pubmed: 31283727
Nat Med. 2020 Jun;26(6):807-808
pubmed: 32514173
BMJ Health Care Inform. 2021 Aug;28(1):
pubmed: 34426417
Nat Med. 2021 Oct;27(10):1663-1665
pubmed: 34635854
Diagn Interv Imaging. 2022 Mar;103(3):151-159
pubmed: 34810137
Acta Orthop. 2020 Oct;91(5):598-604
pubmed: 32589095
AJR Am J Roentgenol. 2018 Dec;211(6):1361-1368
pubmed: 30300006
CJEM. 2018 May;20(3):420-424
pubmed: 28625198
Acta Orthop. 2017 Dec;88(6):581-586
pubmed: 28681679
Invest Radiol. 2020 Feb;55(2):101-110
pubmed: 31725064
Acad Emerg Med. 2020 Feb;27(2):128-138
pubmed: 31702075
Clin Radiol. 2017 Oct;72(10):904.e11-904.e20
pubmed: 28506798
Clin Radiol. 2020 Sep;75(9):713.e17-713.e28
pubmed: 32591230
Pediatr Radiol. 2021 May;51(5):695-696
pubmed: 33666734
Pediatr Radiol. 2021 May;51(5):773-781
pubmed: 33442781