Evaluation of dry eye subtypes and characteristics using conventional assessments and dynamic tear interferometry.

BACKGROUND/AIMS: To evaluate subtypes and characteristics of dry eye (DE) using conventional tests and dynamic tear interferometry, and to investigate determinants of disease severity in each DE subtype. METHODS: 309 patients diagnosed with DE and 69 healthy controls were prospectively enrolled. All eyes were evaluated using Ocular Surface Disease Index (OSDI), Schirmer's test I (ST1) and Meibomian gland dysfunction (MGD) grade were analysed. The tear interferometric pattern and lipid layer thickness were determined using DR-1α and LipiView II, respectively. RESULTS: Dynamic interferometric analysis revealed 56.6% of patients with DE exhibited Jupiter patterns, indicative of aqueous-deficiency, while 43.4% exhibited crystal patterns, indicative of lipid deficiency. These findings were in accordance with classification based on ST1 scores and MGD grade. Conventional assessment indicated 286 patients exhibited evidence of evaporative DE (EDE) due to MGD, while only 11 exhibited signs of pure aqueous-deficient DE (pure ADDE, only ST1 ≤5 mm). Interestingly, of 286 patients with EDE, 144 were categorised into the mixed-ADDE/EDE group, in which ST1 was identified as a strong negative determinant of OSDI. In contrast, 72.2% of patients with mixed-ADDE/EDE exhibited Jupiter patterns (Jupiter mixed), while 27.8% exhibited crystal patterns (crystal mixed). OSDI values were significantly higher in the crystal-mixed group than in the Jupiter mixed, in which OSDI scores were independently associated with ST1 values only. CONCLUSIONS: Our findings indicate that majority of EDE patients also exhibit aqueous deficiency, which can aggravate symptoms even in patients with lipid-deficient mixed-ADDE/EDE. Conventional assessments should be combined with interferometric tear analysis to determine the most appropriate treatment for each DE patient.

Diagnostic value of ganglion cell-inner plexiform layer for early detection of ethambutol-induced optic neuropathy.

AIM: To evaluate the diagnostic value of macular ganglion cell-inner plexiform layer (mGCIPL) thickness versus peripapillary retinal nerve fibre layer (pRNFL) thickness for the early detection of ethambutol-induced optic neuropathy (EON). METHODS: Twenty-eight eyes of 15 patients in the EON group and 100 eyes of 53 healthy subjects in the control group were included. All patients with EON demonstrated the onset of visual symptoms within 3 weeks. Diagnostic power for pRNFL and mGCIPL thicknesses measured by Cirrus spectral-domain optical coherence tomography was assessed by area under the receiver operating characteristic (AUROC) curves and sensitivity. RESULTS: All of the mGCIPL thickness measurements were thinner in the EON group than in the control group in early EON (p<0.001). All of pRNFL thicknesses except inferior RNFL showed AUROC curves above 0.5, and all of the mGCIPL thicknesses showed AUROC curves above 0.5. The AUROC of the average mGCIPL (0.812) thickness was significantly greater than that of the average pRNFL (0.507) thickness (p<0.001). Of all the mGCIPL-related parameters considered, the minimum thickness showed the greatest AUROC value (0.863). The average mGCIPL thickness showed a weak correlation with visual field pattern standard deviations (r2=0.158, p<0.001). CONCLUSIONS: In challenging cases of EON, the mGCIPL thickness has better diagnostic performance in detecting early-onset EON as compared with using pRNFL thickness. Among the early detection ability of mGCIPL thickness, minimum GCIPL thickness has high diagnostic ability.

The possibility of the combination of OCT and fundus images for improving the diagnostic accuracy of deep learning for age-related macular degeneration: a preliminary experiment.

Recently, researchers have built new deep learning (DL) models using a single image modality to diagnose age-related macular degeneration (AMD). Retinal fundus and optical coherence tomography (OCT) images in clinical settings are the most important modalities investigating AMD. Whether concomitant use of fundus and OCT data in DL technique is beneficial has not been so clearly identified. This experimental analysis used OCT and fundus image data of postmortems from the Project Macula. The DL based on OCT, fundus, and combination of OCT and fundus were invented to diagnose AMD. These models consisted of pre-trained VGG-19 and transfer learning using random forest. Following the data augmentation and training process, the DL using OCT alone showed diagnostic efficiency with area under the curve (AUC) of 0.906 (95% confidence interval, 0.891-0.921) and 82.6% (81.0-84.3%) accuracy rate. The DL using fundus alone exhibited AUC of 0.914 (0.900-0.928) and 83.5% (81.8-85.0%) accuracy rate. Combined usage of the fundus with OCT increased the diagnostic power with AUC of 0.969 (0.956-0.979) and 90.5% (89.2-91.8%) accuracy rate. The Delong test showed that the DL using both OCT and fundus data outperformed the DL using OCT alone (P value < 0.001) and fundus image alone (P value < 0.001). This multimodal random forest model showed even better performance than a restricted Boltzmann machine (P value = 0.002) and deep belief network algorithms (P value = 0.042). According to Duncan's multiple range test, the multimodal methods significantly improved the performance obtained by the single-modal methods. In this preliminary study, a multimodal DL algorithm based on the combination of OCT and fundus image raised the diagnostic accuracy compared to this data alone. Future diagnostic DL needs to adopt the multimodal process to combine various types of imaging for a more precise AMD diagnosis. Graphical abstract The basic architectural structure of the tested multimodal deep learning model based on pre-trained deep convolutional neural network and random forest using the combination of OCT and fundus image.

Multi-categorical deep learning neural network to classify retinal images: A pilot study employing small database.

Deep learning emerges as a powerful tool for analyzing medical images. Retinal disease detection by using computer-aided diagnosis from fundus image has emerged as a new method. We applied deep learning convolutional neural network by using MatConvNet for an automated detection of multiple retinal diseases with fundus photographs involved in STructured Analysis of the REtina (STARE) database. Dataset was built by expanding data on 10 categories, including normal retina and nine retinal diseases. The optimal outcomes were acquired by using a random forest transfer learning based on VGG-19 architecture. The classification results depended greatly on the number of categories. As the number of categories increased, the performance of deep learning models was diminished. When all 10 categories were included, we obtained results with an accuracy of 30.5%, relative classifier information (RCI) of 0.052, and Cohen's kappa of 0.224. Considering three integrated normal, background diabetic retinopathy, and dry age-related macular degeneration, the multi-categorical classifier showed accuracy of 72.8%, 0.283 RCI, and 0.577 kappa. In addition, several ensemble classifiers enhanced the multi-categorical classification performance. The transfer learning incorporated with ensemble classifier of clustering and voting approach presented the best performance with accuracy of 36.7%, 0.053 RCI, and 0.225 kappa in the 10 retinal diseases classification problem. First, due to the small size of datasets, the deep learning techniques in this study were ineffective to be applied in clinics where numerous patients suffering from various types of retinal disorders visit for diagnosis and treatment. Second, we found that the transfer learning incorporated with ensemble classifiers can improve the classification performance in order to detect multi-categorical retinal diseases. Further studies should confirm the effectiveness of algorithms with large datasets obtained from hospitals.

CLINICAL OUTCOMES OF INTRAVITREAL BEVACIZUMAB VERSUS PHOTODYNAMIC THERAPY WITH OR WITHOUT BEVACIZUMAB FOR MYOPIC CHOROIDAL NEOVASCULARIZATION: A 7-Year Follow-up Study.

PURPOSE: To compare the long-term visual outcomes of intravitreal bevacizumab (IVB) with those of photodynamic therapy (PDT) for myopic choroidal neovascularization over a 7-year period. METHODS: Eyes treated with IVB (17 eyes) or PDT (20 eyes) that were followed up for at least 7 years were included in this retrospective study. Myopic maculopathy was classified according to the international photographic classification before treatment. The best-corrected visual acuity (BCVA) and the chorioretinal atrophy (CRA) size were measured before and after treatment over a 7-year period. RESULTS: The mean change in BCVA at 7 years was greater in the IVB group than in the PDT group (P = 0.044). While BCVA improved from baseline throughout the 7-year period in the IVB group (P = 0.029), there was no improvement in the PDT group (P = 0.266). In subgroup analysis for 18 eyes with preoperative tessellated fundi (Category 1), there was no difference in BCVA improvement (P = 0.166) and CRA size between the 2 groups at 7 years. However, for 17 eyes with preoperative diffuse CRA (Category 2), BCVA remained unchanged in the IVB group and decreased in the PDT group (P = 0.030) at 7 years. CONCLUSION: IVB resulted in superior long-term functional and anatomical outcomes compared with PDT. In particular, PDT resulted in a greater BCVA decrease and CRA increase compared with IVB in eyes with preoperative diffuse CRA. However, the clinical outcomes were not different in eyes with preoperative tessellated fundi.