Análisis de la capa de células ganglionares con deep learning en el diagnóstico de glaucoma / Ganglion cell layer analysis with deep learning in glaucoma diagnosis

Objetivo Determinar y comparar la precisión diagnóstica en glaucoma de dos modelos de aprendizaje profundo, usando imágenes en infrarrojo del nervio óptico, del fondo de ojo y de la capa de células ganglionares (CCG). Métodos Hemos seleccionado una muestra de pacientes normales y con glaucoma. Se recogieron tres imágenes en infrarrojo con un tomógrafo de coherencia óptica de tipo spectral-domain (SD-OCT). La primera corresponde a la imagen de barrido confocal del fondo de ojo, la segunda es un recorte de la primera centrada en el nervio óptico, y la tercera fue la imagen del corte SD-OCT de la CCG. Nuestros modelos de aprendizaje profundo se desarrollaron en la plataforma MATLAB con las redes neuronales preentrenadas ResNet50 y VGG19. Resultados Se recogieron 498 ojos de 298 pacientes. De los 498 ojos, 312 son glaucomatosos y 186 son normales. En la prueba, la precisión de los modelos fue de 96% (ResNet50) y 96% (VGG19) para las imágenes de la CCG, de 90% (ResNet50) y 90% (VGG19) para las imágenes del nervio óptico y de 82% (ResNet50) y 84% (VGG19) para las de fondo de ojo. El área ROC en la prueba fue de 0,96 (ResNet50) y 0,97 (VGG19) para las imágenes de la CCG, de 0,87 (ResNet50) y 0,88 (VGG19) para las imágenes del nervio óptico, y de 0,79 (ResNet50) y 0,81 (VGG19) para las imágenes de fondo de ojo. Conclusiones Los dos modelos de aprendizaje profundo analizados, aplicados sobre las imágenes de la CCG, ofrecen una alta precisión diagnóstica, sensibilidad y especificidad en el diagnóstico de glaucoma (AU)

Objective To determine and compare the diagnostic precision in glaucoma of two deep learning models using infrared images of the optic nerve, eye fundus, and the ganglion cell layer (GCL). Methods We have selected a sample of normal and glaucoma patients. Three infrared images were registered with a spectral-domain optical coherence tomography (SD-OCT). The first corresponds to the confocal scan image of the fundus, the second is a cut-out of the first centered on the optic nerve, and the third was the SD-OCT image of the GCL. Our deep learning models are developed on the MatLab platform with the ResNet50 and VGG19 pre-trained neural networks. Results 498 eyes of 298 patients were collected. Of the 498 eyes, 312 are glaucoma and 186 are normal. In the test, the precision of the models was 96% (ResNet50) and 96% (VGG19) for the GCL images, 90% (ResNet50) and 90% (VGG19) for the optic nerve images and 82% (ResNet50) and 84% (VGG19) for the fundus images. The ROC area in the test was 0.96 (ResNet50) and 0.97 (VGG19) for the GCL images, 0.87 (ResNet50) and 0.88 (VGG19) for the optic nerve images, and 0.79 (ResNet50) and 0.81 (VGG19) for the fundus images. Conclusions Both deep learning models, applied to the GCL images, achieve high diagnostic precision, sensitivity and specificity in the diagnosis of glaucoma (AU)

SD-OCT peripapillary nerve fibre layer and ganglion cell complex parameters in glaucoma: principal component analysis.

BACKGROUND/AIMS: To identify objective glaucoma-related structural features based on peripapillary (p) and macular (m) spectral domain optical coherence tomography (SD-OCT) parameters and assess their discriminative ability between healthy and glaucoma patients. METHODS: Two hundred and sixty eyes (91 controls and 169 glaucoma) were included in this prospective study. After a complete examination, all participants underwent the posterior pole and the peripapillary retinal nerve fibre layer (pRNFL) protocols of the Spectralis SD-OCT. Principal component analysis (PCA), a data reduction method, was applied to identify and characterise the main information provided by the ganglion cell complex (GCC). The discriminative ability between healthy and glaucomatous eyes of the first principal components (PCs) was compared with that of conventional SD-OCT parameters (pRNFL, macular RNFL (mRNFL), macular ganglion cell layer (mGCL)and macular inner plexiform layer (mIPL)) using 10-fold cross-validated areas under the curve (AUC). RESULTS: The first PC explained 58% of the total information contained in the GCC and the pRNFL parameters and was the result of a general combination of almost all variables studied (diffuse distribution). Other PCs were driven mainly by pRNFL and mRNFL measurements. PCs and pRNFL had similar AUC (0.95 vs 0.96, p=0.88), and outperformed the other structural measurements: mRNFL (0.91, p=0.002), mGCL (0.92, p=0.02) and mIPL (0.92, p=0.0001). CONCLUSIONS: PCA identified a diffuse representation of the papillary and macular SD-OCT parameters as the most important PC to summarise structural data in healthy and glaucomatous eyes. PCs and pRNFL parameters showed the greatest discriminative ability between healthy and glaucoma cases.

Ganglion cell layer analysis with deep learning in glaucoma diagnosis. / Análisis de la capa de células ganglionares con deep learning en el diagnóstico de glaucoma.

OBJECTIVE: To determine and compare the diagnostic precision in glaucoma of two deep learning models using infrared images of the optic nerve, eye fundus, and the ganglion cell layer (GCL). METHODS: We have selected a sample of normal and glaucoma patients. Three infrared images were registered with a spectral-domain optical coherence tomography (SD-OCT). The first corresponds to the confocal scan image of the fundus, the second is a cut-out of the first centered on the optic nerve, and the third was the SD-OCT image of the GCL. Our deep learning models are developed on the MatLab platform with the ResNet50 and VGG19 pre-trained neural networks. RESULTS: 498 eyes of 298 patients were collected. Of the 498 eyes, 312 are glaucoma and 186 are normal. In the test, the precision of the models was 96% (ResNet50) and 96% (VGG19) for the GCL images, 90% (ResNet50) and 90% (VGG19) for the optic nerve images and 82% (ResNet50) and 84% (VGG19) for the fundus images. The ROC area in the test was 0.96 (ResNet50) and 0.97 (VGG19) for the GCL images, 0.87 (ResNet50) and 0.88 (VGG19) for the optic nerve images, and 0.79 (ResNet50) and 0.81 (VGG19) for the fundus images. CONCLUSIONS: Both deep learning models, applied to the GCL images, achieve high diagnostic precision, sensitivity and specificity in the diagnosis of glaucoma.

Perimetric progression in open angle glaucoma and the Visual Field Index (VFI).

PURPOSE: To evaluate the changes in the Visual Field Index (VFI) in eyes with perimetric glaucomatous progression, and to compare these against stable glaucoma patients. PATIENTS AND METHODS: Consecutive patients with open angle glaucoma with a minimum of 6 reliable visual fields and 2 years of follow-up were identified. Perimetric progression was assessed by 4 masked glaucoma experts from different units, and classified into 3 categories: "definite progression," "suspected progression," or "no progression." This was compared with the Glaucoma Progression Analysis (GPA) II and VFI linear regression analysis, where progression was defined as a negative slope with significance of <5%. RESULTS: Three hundred ninety-seven visual fields from 51 eyes of 39 patients were assessed. The mean number of visual fields was 7.8 (SD 1.1) per eye, and the mean follow-up duration was 63.7 (SD 13.4) months. The mean VFI linear regression slope showed an overall statistically significant difference (P<0.001, analysis of variance) for each category of progression. Using expert consensus opinion as the reference standard, both VFI analysis and GPA II had high specificity (0.93 and 0.90, respectively), but relatively low sensitivity (0.45 and 0.41, respectively). CONCLUSIONS: The mean VFI regression slope in our cohort of eyes without perimetric progression showed a statistically significant difference compared with those with suspected and definite progression. VFI analysis and GPA II both had similarly high specificity but low sensitivity when compared with expert consensus opinion.