Using Artificial Neural Networks to Predict Intra-Abdominal Abscess Risk Post-Appendectomy.
artificial intelligence
intraabdominal abscess
pediatric
Journal
Annals of surgery open : perspectives of surgical history, education, and clinical approaches
ISSN: 2691-3593
Titre abrégé: Ann Surg Open
Pays: United States
ID NLM: 101769928
Informations de publication
Date de publication:
Jun 2022
Jun 2022
Historique:
received:
27
09
2021
accepted:
18
04
2022
medline:
23
5
2022
pubmed:
23
5
2022
entrez:
21
8
2023
Statut:
epublish
Résumé
To determine if artificial neural networks (ANN) could predict the risk of intra-abdominal abscess (IAA) development post-appendectomy. IAA formation occurs in 13.6% to 14.6% of appendicitis cases with "complicated" appendicitis as the most common cause of IAA. There remains inconsistency in describing the severity of appendicitis with variation in treatment with respect to perforated appendicitis. Two "reproducible" ANN with different architectures were developed on demographic, clinical, and surgical information from a retrospective surgical dataset of 1574 patients less than 19 years old classified as either negative (n = 1,328) or positive (n = 246) for IAA post-appendectomy for appendicitis. Of 34 independent variables initially, 12 variables with the highest influence on the outcome selected for the final dataset for ANN model training and testing. A total of 1574 patients were used for training and test sets (80%/20% split). Model 1 achieved accuracy of 89.84%, sensitivity of 70%, and specificity of 93.61% on the test set. Model 2 achieved accuracy of 84.13%, sensitivity of 81.63%, and specificity of 84.6%. ANN applied to selected variables can accurately predict patients who will have IAA post-appendectomy. Our reproducible and explainable ANNs potentially represent a state-of-the-art method for optimizing post-appendectomy care.
Sections du résumé
Objective
UNASSIGNED
To determine if artificial neural networks (ANN) could predict the risk of intra-abdominal abscess (IAA) development post-appendectomy.
Background
UNASSIGNED
IAA formation occurs in 13.6% to 14.6% of appendicitis cases with "complicated" appendicitis as the most common cause of IAA. There remains inconsistency in describing the severity of appendicitis with variation in treatment with respect to perforated appendicitis.
Methods
UNASSIGNED
Two "reproducible" ANN with different architectures were developed on demographic, clinical, and surgical information from a retrospective surgical dataset of 1574 patients less than 19 years old classified as either negative (n = 1,328) or positive (n = 246) for IAA post-appendectomy for appendicitis. Of 34 independent variables initially, 12 variables with the highest influence on the outcome selected for the final dataset for ANN model training and testing.
Results
UNASSIGNED
A total of 1574 patients were used for training and test sets (80%/20% split). Model 1 achieved accuracy of 89.84%, sensitivity of 70%, and specificity of 93.61% on the test set. Model 2 achieved accuracy of 84.13%, sensitivity of 81.63%, and specificity of 84.6%.
Conclusions
UNASSIGNED
ANN applied to selected variables can accurately predict patients who will have IAA post-appendectomy. Our reproducible and explainable ANNs potentially represent a state-of-the-art method for optimizing post-appendectomy care.
Identifiants
pubmed: 37601615
doi: 10.1097/AS9.0000000000000168
pmc: PMC10431380
doi:
Types de publication
Journal Article
Langues
eng
Pagination
e168Informations de copyright
Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.
Références
Patient Saf Surg. 2019 Dec 07;13:41
pubmed: 31827618
Mol Cancer. 2005 Aug 06;4:29
pubmed: 16083507
Med Arh. 2009;63(5):249-51
pubmed: 20380121
NPJ Digit Med. 2020 Mar 26;3:47
pubmed: 32258429
Drug Discov Today. 2017 Nov;22(11):1680-1685
pubmed: 28881183
Arch Surg. 2001 Apr;136(4):438-41
pubmed: 11296116
J Gastrointest Surg. 2012 Oct;16(10):1929-39
pubmed: 22890606
JAMA Surg. 2018 Nov 1;153(11):1021-1027
pubmed: 30046808
Nat Commun. 2019 Mar 11;10(1):1096
pubmed: 30858366
JAMA. 2016 Dec 13;316(22):2353-2354
pubmed: 27898975
Surg Infect (Larchmt). 2017 Jan;18(1):1-76
pubmed: 28085573
IEEE J Biomed Health Inform. 2015 Nov;19(6):1893-905
pubmed: 25095272
J Pediatr Surg. 2013 Jan;48(1):74-80
pubmed: 23331796
BMC Bioinformatics. 2013 Mar 22;14:106
pubmed: 23522326
PLoS One. 2019 Feb 19;14(2):e0212356
pubmed: 30779785
J Laparoendosc Adv Surg Tech A. 2009 Apr;19 Suppl 1:S15-8
pubmed: 19371148
Pediatrics. 2005 Apr;115(4):920-5
pubmed: 15805365
Annu Int Conf IEEE Eng Med Biol Soc. 2017 Jul;2017:2566-2569
pubmed: 29060423
Asian Pac J Cancer Prev. 2015;16(12):5095-9
pubmed: 26163648
Cochrane Database Syst Rev. 2018 Nov 28;11:CD001546
pubmed: 30484855
BMC Med Inform Decis Mak. 2020 Nov 30;20(1):310
pubmed: 33256715
Int J Methods Psychiatr Res. 2011 Mar;20(1):40-9
pubmed: 21499542
Front Big Data. 2021 Jul 01;4:688969
pubmed: 34278297
Aust N Z J Surg. 1999 May;69(5):373-4
pubmed: 10353555
Cochrane Database Syst Rev. 2005 Jul 20;(3):CD001439
pubmed: 16034862
Acad Radiol. 1998 Jul;5(7):473-9
pubmed: 9653463
Asian Pac J Cancer Prev. 2014;15(13):5349-53
pubmed: 25041000
J Dig Dis. 2019 Sep;20(9):486-494
pubmed: 31328389
J Pediatr Surg. 2008 Dec;43(12):2242-5
pubmed: 19040944
Surg Laparosc Endosc Percutan Tech. 2010 Dec;20(6):362-70
pubmed: 21150411
J Pediatr. 2014 Jun;164(6):1286-91.e2
pubmed: 24565425
J Pediatr Surg. 2007 May;42(5):857-61
pubmed: 17502199
Am J Surg. 2002 Jun;183(6):608-13
pubmed: 12095586
World J Surg. 2017 May;41(5):1254-1258
pubmed: 28074278
Hepat Mon. 2011 Jul;11(7):544-7
pubmed: 22087192