RESUMO
PURPOSE: To compare the diagnostic performance of human gradings vs predictions provided by a machine-to-machine (M2M) deep learning (DL) algorithm trained to quantify retinal nerve fiber layer (RNFL) damage on fundus photographs. DESIGN: Evaluation of a machine learning algorithm. METHODS: An M2M DL algorithm trained with RNFL thickness parameters from spectral-domain optical coherence tomography was applied to a subset of 490 fundus photos of 490 eyes of 370 subjects graded by 2 glaucoma specialists for the probability of glaucomatous optical neuropathy (GON), and estimates of cup-to-disc (C/D) ratios. Spearman correlations with standard automated perimetry (SAP) global indices were compared between the human gradings vs the M2M DL-predicted RNFL thickness values. The area under the receiver operating characteristic curves (AUC) and partial AUC for the region of clinically meaningful specificity (85%-100%) were used to compare the ability of each output to discriminate eyes with repeatable glaucomatous SAP defects vs eyes with normal fields. RESULTS: The M2M DL-predicted RNFL thickness had a significantly stronger absolute correlation with SAP mean deviation (rho=0.54) than the probability of GON given by human graders (rho=0.48; P < .001). The partial AUC for the M2M DL algorithm was significantly higher than that for the probability of GON by human graders (partial AUC = 0.529 vs 0.411, respectively; P = .016). CONCLUSION: An M2M DL algorithm performed as well as, if not better than, human graders at detecting eyes with repeatable glaucomatous visual field loss. This DL algorithm could potentially replace human graders in population screening efforts for glaucoma.
Assuntos
Aprendizado Profundo , Glaucoma de Ângulo Aberto/diagnóstico , Fibras Nervosas/patologia , Doenças do Nervo Óptico/diagnóstico , Exame Físico , Células Ganglionares da Retina/patologia , Idoso , Algoritmos , Área Sob a Curva , Estudos Transversais , Feminino , Fundo de Olho , Glaucoma de Ângulo Aberto/diagnóstico por imagem , Gonioscopia , Humanos , Pressão Intraocular/fisiologia , Masculino , Pessoa de Meia-Idade , Doenças do Nervo Óptico/diagnóstico por imagem , Fotografação , Curva ROC , Estudos Retrospectivos , Tomografia de Coerência Óptica , Transtornos da Visão/diagnóstico , Testes de Campo Visual/métodos , Campos Visuais/fisiologiaRESUMO
PURPOSE: Older people present significant declines in face recognition with age. Spatial vision (high-contrast acuity) and age are the best predictors of face recognition. Visual disabilities are more common in the older population due to aging eye diseases. The purpose of the study was to compare the face recognition memory deficit between primary open angle glaucoma (POAG) and age-related macular degeneration (ARMD) patients living in a developing country. METHODS: This was a cross-sectional, observational study. The sample comprised 64 patients with POAG, 48 with ARMD, and 52 controls. All groups were matched for age, gender, comorbidity, and ethnic distribution. Evidence of cognitive impairment was ruled out and subjects with even mild cognitive impairment were not included in the study. After a complete eye examination including measurement of the best-corrected visual acuity, fundus evaluation, and automated visual field, all subjects underwent the Cambridge face memory test (CFMT). CFMT score in percentage (%) was the main outcome measure and data were compared with ANOVA. RESULTS: The mean age was 66.6 ± 9.2, 69.8 ± 9.3, and 63.4 ± 7.3 years, for POAG, ARMD, and controls, respectively (P = 0.152). Gender, ethnicity, and comorbidity were evenly distributed among the groups. The CFMT score was 53.3 ± 15.2%, 49.8 ± 14.2%, and 62.1 ± 15.9% for POAG, ARMD, and controls, respectively (P < 0.001). CONCLUSION: ARMD and POAG patients have higher face recognition memory deficit as compared to normal controls. This might be due to a visual disability.
Assuntos
Percepção de Forma/fisiologia , Glaucoma de Ângulo Aberto/fisiopatologia , Degeneração Macular/fisiopatologia , Transtornos da Memória/etiologia , Acuidade Visual , Campos Visuais/fisiologia , Idoso , Brasil/epidemiologia , Estudos Transversais , Países em Desenvolvimento , Feminino , Glaucoma de Ângulo Aberto/complicações , Gonioscopia , Humanos , Pressão Intraocular/fisiologia , Degeneração Macular/complicações , Masculino , Transtornos da Memória/epidemiologia , Transtornos da Memória/fisiopatologia , Pessoa de Meia-IdadeRESUMO
PURPOSE: Falls are very prevalent in the older population. Visually impaired elderly patients are prone to falls as the result of visual loss and ageing. The purpose of the study was to compare the fear of falling (FoF) between primary open angle glaucoma (POAG) and age-related macular degeneration (ARMD) patients who live in a developing country. METHODS: This was a cross-sectional observational study. After a complete eye examination including measurement of best-corrected visual acuity, ophthalmoscopy, and automated visual field, all subjects completed the Fall Efficacy Scale International Brazil (FES-I-Brazil) questionnaire. RESULTS: The sample comprised 64 patients with POAG, 48 with ARMD, and 52 controls. All groups were matched for age, sex, comorbidity, and ethnic distribution. The FES-I score was 24.6 ± 8.7, 25.3 ± 6.3, and 24.2 ± 7.7 for glaucoma, ARMD, and controls, respectively (P = 0.894). A post hoc analysis comparing all subjects with advanced visual field defect (mean deviation [MD] < -12 dB) revealed a higher FES-I score in ARMD patients as compared to POAG ones (46.2 ± 16.8 and 24.0 ± 7.7 for ARMD and POAG, respectively, P < 0.001). CONCLUSION: In this cohort of elderly subjects with eye diseases, the FoF was similar among groups; however, ARMD patients with more compromised visual field had higher FoF as compared to POAG patients and controls. TRANSLATIONAL RELEVANCE: A high rate of fear of falling exists in ARMD patients with compromised visual field. This finding may be useful in developing multidimensional strategies to decrease fear of falling and improve quality of life in older persons living in a developing country.