[Deep learning to support therapy decisions for intravitreal injections]. / Deep Learning zur Unterstützung der Therapieentscheidung bei intravitrealen Injektionen.
Ophthalmologe
; 115(9): 722-727, 2018 Sep.
Article
em De
| MEDLINE
| ID: mdl-29713804
ABSTRACT
Significant progress has been made in artificial intelligence and computer vision research in recent years. Machine learning methods excel in a wide variety of tasks where sufficient data are available. We describe the application of a deep convolutional neural network for the prediction of treatment indication with anti-vascular endothelial growth factor (VEGF) medications based on central retinal optical coherence tomography (OCT) scans. The neural network classifier was trained with OCT images acquired during routine treatment at the University of Regensburg over the years 2008-2016. In over 95% of the cases the treatment indication was accurately predicted based on a singular OCT B scan without human intervention. Despite promising classification the results of deep learning techniques, should always be controlled by the treating physician because false classification can never be excluded due to the probabilistic nature of the method.
Palavras-chave
Texto completo:
1
Coleções:
01-internacional
Base de dados:
MEDLINE
Assunto principal:
Aprendizado Profundo
Tipo de estudo:
Prognostic_studies
Limite:
Humans
Idioma:
De
Revista:
Ophthalmologe
Ano de publicação:
2018
Tipo de documento:
Article