A Deep Learning Approach for Accurate Discrimination Between Optic Disc Drusen and Papilledema on Fundus Photographs.
Journal
Journal of neuro-ophthalmology : the official journal of the North American Neuro-Ophthalmology Society
ISSN: 1536-5166
Titre abrégé: J Neuroophthalmol
Pays: United States
ID NLM: 9431308
Informations de publication
Date de publication:
02 Aug 2024
02 Aug 2024
Historique:
medline:
2
8
2024
pubmed:
2
8
2024
entrez:
2
8
2024
Statut:
aheadofprint
Résumé
Optic disc drusen (ODD) represent an important differential diagnosis of papilledema caused by intracranial hypertension, but their distinction may be difficult in clinical practice. The aim of this study was to train, validate, and test a dedicated deep learning system (DLS) for binary classification of ODD vs papilledema (including various subgroups within each category), on conventional mydriatic digital ocular fundus photographs collected in a large international multiethnic population. This retrospective study included 4,508 color fundus images in 2,180 patients from 30 neuro-ophthalmology centers (19 countries) participating in the Brain and Optic Nerve Study with Artificial Intelligence (BONSAI) Group. For training and internal validation, we used 857 ODD images and 3,230 papilledema images, in 1,959 patients. External testing was performed on an independent data set (221 patients), including 207 images with ODD (96 visible and 111 buried), provided by 3 centers of the Optic Disc Drusen Studies Consortium, and 214 images of papilledema (92 mild-to-moderate and 122 severe) from a previously validated study. The DLS could accurately distinguish between all ODD and papilledema (all severities included): area under the receiver operating characteristic curve (AUC) 0.97 (95% confidence interval [CI], 0.96-0.98), accuracy 90.5% (95% CI, 88.0%-92.9%), sensitivity 86.0% (95% CI, 82.1%-90.1%), and specificity 94.9% (95% CI, 92.3%-97.6%). The performance of the DLS remained high for discrimination of buried ODD from mild-to-moderate papilledema: AUC 0.93 (95% CI, 0.90-0.96), accuracy 84.2% (95% CI, 80.2%-88.6%), sensitivity 78.4% (95% CI, 72.2%-84.7%), and specificity 91.3% (95% CI, 87.0%-96.4%). A dedicated DLS can accurately distinguish between ODD and papilledema caused by intracranial hypertension, even when considering buried ODD vs mild-to-moderate papilledema.
Sections du résumé
BACKGROUND
BACKGROUND
Optic disc drusen (ODD) represent an important differential diagnosis of papilledema caused by intracranial hypertension, but their distinction may be difficult in clinical practice. The aim of this study was to train, validate, and test a dedicated deep learning system (DLS) for binary classification of ODD vs papilledema (including various subgroups within each category), on conventional mydriatic digital ocular fundus photographs collected in a large international multiethnic population.
METHODS
METHODS
This retrospective study included 4,508 color fundus images in 2,180 patients from 30 neuro-ophthalmology centers (19 countries) participating in the Brain and Optic Nerve Study with Artificial Intelligence (BONSAI) Group. For training and internal validation, we used 857 ODD images and 3,230 papilledema images, in 1,959 patients. External testing was performed on an independent data set (221 patients), including 207 images with ODD (96 visible and 111 buried), provided by 3 centers of the Optic Disc Drusen Studies Consortium, and 214 images of papilledema (92 mild-to-moderate and 122 severe) from a previously validated study.
RESULTS
RESULTS
The DLS could accurately distinguish between all ODD and papilledema (all severities included): area under the receiver operating characteristic curve (AUC) 0.97 (95% confidence interval [CI], 0.96-0.98), accuracy 90.5% (95% CI, 88.0%-92.9%), sensitivity 86.0% (95% CI, 82.1%-90.1%), and specificity 94.9% (95% CI, 92.3%-97.6%). The performance of the DLS remained high for discrimination of buried ODD from mild-to-moderate papilledema: AUC 0.93 (95% CI, 0.90-0.96), accuracy 84.2% (95% CI, 80.2%-88.6%), sensitivity 78.4% (95% CI, 72.2%-84.7%), and specificity 91.3% (95% CI, 87.0%-96.4%).
CONCLUSIONS
CONCLUSIONS
A dedicated DLS can accurately distinguish between ODD and papilledema caused by intracranial hypertension, even when considering buried ODD vs mild-to-moderate papilledema.
Identifiants
pubmed: 39090774
doi: 10.1097/WNO.0000000000002223
pii: 00041327-990000000-00692
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Subventions
Organisme : Singapore National Medical Research Centre
ID : CIRG18Nov-0013
Organisme : Duke-NUS Medical School
ID : 05/FY2019/P2/06-A60
Informations de copyright
Copyright © 2024 by North American Neuro-Ophthalmology Society.
Déclaration de conflit d'intérêts
The authors report no conflicts of interest.
Références
Biousse V, Bruce BB, Newman NJ. Ophthalmoscopy in the 21st century: the 2017 H. Houston Merritt lecture. Neurology. 2018;90:167–175.
Fisayo A, Bruce BB, Newman NJ, Biousse V. Overdiagnosis of idiopathic intracranial hypertension. Neurology. 2016;86:341–350.
Skougaard M, Heegaard S, Malmqvist L, Hamann S. Prevalence and histopathological signatures of optic disc drusen based on microscopy of 1713 enucleated eyes. Acta Ophthalmol. 2020;98:195–200.
Friedman AH, Henkind P, Gartner S. Drusen of the optic disc. A histopathological study. Trans Ophthalmol Soc U K (1962). 1975;95:4–9.
Fraser JA, Rueløkke LL, Malmqvist L, Hamann S. Prevalence of optic disc drusen in young patients with nonarteritic anterior ischemic optic neuropathy: a 10-year retrospective study. J Neuro Ophthalmol. 2021;41:200–205.
Neudorfer M, Ben‐Haim MS, Leibovitch I, Kesler A. The efficacy of optic nerve ultrasonography for differentiating papilloedema from pseudopapilloedema in eyes with swollen optic discs. Acta Ophthalmol. 2013;91:376–380.
Costello F, Rothenbuehler S, Sibony P, Hamann S, Optic Disc Drusen Studies Consortium. Diagnosing optic disc drusen in the modern imaging era: a practical approach. Neuroophthalmol. 2021;45:1–16.
Costello F, Malmqvist L, Hamann S. The role of optical coherence tomography in differentiating optic disc drusen from optic disc edema. Asia Pac J Ophthalmol (Phila). 2018;7:271–279.
Girard MJA, Panda S, Tun TA, et al. Discriminating between papilledema and optic disc drusen using 3D structural analysis of the optic nerve head. Neurology. 2023;100:e192–e202.
Lee KM, Woo SJ, Hwang J. Differentiation between optic disc drusen and optic disc oedema using fundus photography. Acta Ophthalmol 2017;95:e329–e335.
Leong YY, Vasseneix C, Finkelstein MT, Milea D, Najjar RP. Artificial intelligence meets neuro-ophthalmology. Asia Pacific J Ophthalmol. 2022;11:111–125.
Milea D, Najjar RP, Zhubo J, et al., BONSAI Group. Artificial intelligence to detect papilledema from ocular fundus photographs. N Engl J Med. 2020;382:1687–1695.
Biousse V, Newman NJ, Najjar RP, et al., BONSAI Brain and Optic Nerve Study with Artificial Intelligence Study Group. Optic disc classification by deep learning versus expert neuro‐ophthalmologists. Ann Neurol. 2020;88:785–795.
Vasseneix C, Najjar RP, Xu X, et al., BONSAI Group. Accuracy of a deep learning system for classification of papilledema severity on ocular fundus photographs. Neurology. 2021;97:e369–e377.
Milea L, Najjar RP. Classif-Eye: a semi-automated image classification application. GitHub Repository. 2020. Available at: github.com/milealeonard/Classif-Eye/. Accessed June 8, 2024
Friedman DI, Liu GT, Digre KB. Revised diagnostic criteria for the pseudotumor cerebri syndrome in adults and children. Neurology. 2013;81:1159–1165.
Pineles SL, Arnold AC. Fluorescein angiographic identification of optic disc drusen with and without optic disc edema. J Neuroophthalmol. 2012;32:17–22.
Sibony PA, Kupersmith MJ, Kardon RH. Optical coherence tomography neuro-toolbox for the diagnosis and management of papilledema, optic disc edema, and pseudopapilledema. J Neuroophthalmol. 2021;41:77–92.
Vasseneix C, Nusinovici S, Xu X, et al., BONSAI Brain and Optic Nerve Study With Artificial Intelligence Group. Deep learning system outperforms clinicians in identifying optic disc abnormalities. J Neuroophthalmol. 2023;43:159–167.
Biousse V, Najjar RP, Tang Z, et al., BONSAI Study Group. Application of a deep learning system to detect papilledema on nonmydriatic ocular fundus photographs in an emergency department. Am J Ophthalmol. 2024;261:199–207.