RESUMEN
Purpose: To examine deep learning (DL)-based methods for accurate segmentation of geographic atrophy (GA) lesions using fundus autofluorescence (FAF) and near-infrared (NIR) images. Methods: This retrospective analysis utilized imaging data from study eyes of patients enrolled in Proxima A and B (NCT02479386; NCT02399072) natural history studies of GA. Two multimodal DL networks (UNet and YNet) were used to automatically segment GA lesions on FAF; segmentation accuracy was compared with annotations by experienced graders. The training data set comprised 940 image pairs (FAF and NIR) from 183 patients in Proxima B; the test data set comprised 497 image pairs from 154 patients in Proxima A. Dice coefficient scores, Bland-Altman plots, and Pearson correlation coefficient (r) were used to assess performance. Results: On the test set, Dice scores for the DL network to grader comparison ranged from 0.89 to 0.92 for screening visit; Dice score between graders was 0.94. GA lesion area correlations (r) for YNet versus grader, UNet versus grader, and between graders were 0.981, 0.959, and 0.995, respectively. Longitudinal GA lesion area enlargement correlations (r) for screening to 12 months (n = 53) were lower (0.741, 0.622, and 0.890, respectively) compared with the cross-sectional results at screening. Longitudinal correlations (r) from screening to 6 months (n = 77) were even lower (0.294, 0.248, and 0.686, respectively). Conclusions: Multimodal DL networks to segment GA lesions can produce accurate results comparable with expert graders. Translational Relevance: DL-based tools may support efficient and individualized assessment of patients with GA in clinical research and practice.
Asunto(s)
Aprendizaje Profundo , Atrofia Geográfica , Humanos , Estudios Transversales , Fondo de Ojo , Atrofia Geográfica/diagnóstico por imagen , Estudios Retrospectivos , Estudios Clínicos como AsuntoRESUMEN
PURPOSE: To develop a deep learning semantic segmentation network to automate the assessment of 8 periorbital measurements DESIGN: Development and validation of an artificial intelligence (AI) segmentation algorithm METHODS: A total of 418 photographs of periorbital areas were used to train a deep learning semantic segmentation model to segment iris, aperture, and brow areas. These data were used to develop a post-processing algorithm that measured margin reflex distance (MRD) 1 and 2, medial canthal height (MCH), lateral canthal height (LCH), medial brow height (MBH), lateral brow height (LBH), medial intercanthal distance (MID), and lateral intercanthal distance (LID). The algorithm validity was evaluated on a prospective hold-out test set against 3 graders. The main outcome measures were dice coefficient, mean absolute difference, intraclass correlation coefficient, and Bland-Altman analysis. A smartphone video was also segmented and evaluated as proof of concept. RESULTS: The AI algorithm performed in close agreement with all human graders, with a mean absolute difference of 0.5 mm for MRD1, MRD2, LCH, and MCH. The mean absolute difference between graders is approximately 1.5-2 mm for LBH and MBH and approximately 2-4 mm for MID and LID. The 95% confidence intervals for all graders overlapped in most cases, demonstrating that the algorithm performs similarly to human graders. The segmentation of a smartphone video demonstrated that MRD1 can be dynamically measured. CONCLUSIONS: We present, to our knowledge, the first open-sourced, artificial intelligence system capable of automating static and dynamic periorbital measurements. A fully automated tool stands to transform the delivery of clinical care and quantification of surgical outcomes.
Asunto(s)
Inteligencia Artificial , Párpados , Automatización , Párpados/diagnóstico por imagen , Cara , Humanos , Estudios ProspectivosRESUMEN
Purpose: Delayed rod-mediated dark adaptation (RMDA) is a functional biomarker for incipient age-related macular degeneration (AMD). We used anatomically restricted spectral domain optical coherence tomography (SD-OCT) imaging data to localize de novo imaging features associated with and to test hypotheses about delayed RMDA. Methods: Rod intercept time (RIT) was measured in participants with and without AMD at 5 degrees from the fovea, and macular SD-OCT images were obtained. A deep learning model was trained with anatomically restricted information using a single representative B-scan through the fovea of each eye. Mean-occlusion masking was utilized to isolate the relevant imaging features. Results: The model identified hyporeflective outer retinal bands on macular SD-OCT associated with delayed RMDA. The validation mean standard error (MSE) registered to the foveal B-scan localized the lowest error to 0.5 mm temporal to the fovea center, within an overall low-error region across the rod-free zone and adjoining parafovea. Mean absolute error (MAE) on the test set was 4.71 minutes (8.8% of the dynamic range). Conclusions: We report a novel framework for imaging biomarker discovery using deep learning and demonstrate its ability to identify and localize a previously undescribed biomarker in retinal imaging. The hyporeflective outer retinal bands in central macula on SD-OCT demonstrate a structural basis for dysfunctional rod vision that correlates to published histopathologic findings. Translational Relevance: This agnostic approach to anatomic biomarker discovery strengthens the rationale for RMDA as an outcome measure in early AMD clinical trials, and also expands the utility of deep learning beyond automated diagnosis to fundamental discovery.
Asunto(s)
Aprendizaje Profundo , Mácula Lútea , Degeneración Macular , Adaptación a la Oscuridad , Humanos , Mácula Lútea/diagnóstico por imagen , Degeneración Macular/diagnóstico por imagen , Agudeza VisualRESUMEN
PURPOSE: To determine the prevalence and significance of subretinal drusenoid deposits (reticular pseudodrusen) among patients with age-related macular degeneration (AMD). DESIGN: A prospective study with a nested case-control study of consecutive patients with AMD seen in a referral retinal practice. PARTICIPANTS: There were 153 patients with AMD, 131 of whom had > or =1 eye with late AMD, which was defined as either central geographic atrophy or choroidal neovascularization. The control group consisted of 101 patients who did not have AMD as their primary diagnosis, central serous chorioretinopathy, high myopia, retinal detachment, or laser treatment in the macular area. METHODS: The presence of subretinal drusenoid deposits was determined by 2 methods, using the blue channel of color fundus photograph and the spectral domain optical coherence tomography (SD-OCT) sections. Soft drusen were determined from color fundus photographs and confirmed by SD-OCT. MAIN OUTCOME MEASURES: Prevalence of ocular risk factors and subretinal drusenoid deposits in eyes with AMD and their association with late AMD. RESULTS: There were 153 patients who had any form of AMD, with a mean age of 80.3 years. Subretinal drusenoid deposits were diagnosed in the case group in 13 (8.7%) of right and 18 (12.0%) of left eyes using the blue channel of the color photograph and in 58 (38.4%) of right and 54 (35.8%) of left eyes using SD-OCT. Soft drusen and subretinal drusenoid deposits detected by SD-OCT were found to be independently correlated with late AMD (soft drusen odds ratio = 16.66 [P<0.001]; subretinal drusenoid deposits as detected by OCT odds ratio = 2.64 [P = 0.034]). In the control group, subretinal drusenoid deposits were diagnosed in 6 (6.5%) of right and 6 (6.3%) of left eyes using SD-OCT. CONCLUSIONS: Both soft drusen and subretinal drusenoid deposits occur in patients with AMD and both are significantly associated with late AMD. These findings suggest that detection and classification of drusen and consequently assignment of risk should be based on a methodology that includes SD-OCT.
Asunto(s)
Degeneración Macular/epidemiología , Drusas Retinianas/epidemiología , Anciano , Anciano de 80 o más Años , Estudios de Casos y Controles , Neovascularización Coroidal/diagnóstico , Neovascularización Coroidal/epidemiología , Técnicas de Diagnóstico Oftalmológico , Femenino , Atrofia Geográfica/diagnóstico , Atrofia Geográfica/epidemiología , Humanos , Degeneración Macular/diagnóstico , Masculino , New York/epidemiología , Fotograbar , Prevalencia , Estudios Prospectivos , Drusas Retinianas/diagnóstico , Factores de Riesgo , Tomografía de Coherencia ÓpticaRESUMEN
BACKGROUND AND OBJECTIVES: To compare the fluorescence measurements acquired from a fundus camera with those from a scanning laser ophthalmoscope (SLO) camera. MATERIALS AND METHODS: The fundus camera and the SLO camera were used to capture images of 29 cuvettes each containing serially diluted sodium fluorescein dye in normal saline. The intensity levels of the resulting images were plotted as a function of concentration to compare the two cameras. Ten samples of serially diluted indocyanine green (ICG) dye in bovine serum were also measured. RESULTS: Both cameras revealed that fluorescence intensity varied as a function of the logarithmic concentration of the dye, independent of the actual dye used, with expected decrease in fluorescence at very high concentrations of dye due to quenching of fluorescence. There were very small variations on repeated trials with the fundus camera, whereas the SLO camera exhibited marked variability, particularly at higher concentrations of dye. Measurements acquired with the SLO camera varied as a function of time, which did not occur with the fundus camera. The image averaging software on the SLO camera caused shifts in the grayscale values measured that depended on the initial amount of fluorescence measured in the raw samples. CONCLUSIONS: Although these differences may cause modest qualitative differences in imaging the ocular fundus, the variation in data obtained from the SLO camera would seem problematic if quantification of the amounts of fluorescence is required.