Predicting glaucoma progression using deep learning framework guided by generative algorithm.

Glaucoma is a slowly progressing optic neuropathy that may eventually lead to blindness. To help patients receive customized treatment, predicting how quickly the disease will progress is important. Structural assessment using optical coherence tomography (OCT) can be used to visualize glaucomatous optic nerve and retinal damage, while functional visual field (VF) tests can be used to measure the extent of vision loss. However, VF testing is patient-dependent and highly inconsistent, making it difficult to track glaucoma progression. In this work, we developed a multimodal deep learning model comprising a convolutional neural network (CNN) and a long short-term memory (LSTM) network, for glaucoma progression prediction. We used OCT images, VF values, demographic and clinical data of 86 glaucoma patients with five visits over 12 months. The proposed method was used to predict VF changes 12 months after the first visit by combining past multimodal inputs with synthesized future images generated using generative adversarial network (GAN). The patients were classified into two classes based on their VF mean deviation (MD) decline: slow progressors (< 3 dB) and fast progressors (> 3 dB). We showed that our generative model-based novel approach can achieve the best AUC of 0.83 for predicting the progression 6 months earlier. Further, the use of synthetic future images enabled the model to accurately predict the vision loss even earlier (9 months earlier) with an AUC of 0.81, compared to using only structural (AUC = 0.68) or only functional measures (AUC = 0.72). This study provides valuable insights into the potential of using synthetic follow-up OCT images for early detection of glaucoma progression.

Three-Dimensional Structural Phenotype of the Optic Nerve Head as a Function of Glaucoma Severity.

Importance: The 3-dimensional (3-D) structural phenotype of glaucoma as a function of severity was thoroughly described and analyzed, enhancing understanding of its intricate pathology beyond current clinical knowledge. Objective: To describe the 3-D structural differences in both connective and neural tissues of the optic nerve head (ONH) between different glaucoma stages using traditional and artificial intelligence-driven approaches. Design, Setting, and Participants: This cross-sectional, clinic-based study recruited 541 Chinese individuals receiving standard clinical care at Singapore National Eye Centre, Singapore, and 112 White participants of a prospective observational study at Vilnius University Hospital Santaros Klinikos, Vilnius, Lithuania. The study was conducted from May 2022 to January 2023. All participants had their ONH imaged using spectral-domain optical coherence tomography and had their visual field assessed by standard automated perimetry. Main Outcomes and Measures: (1) Clinician-defined 3-D structural parameters of the ONH and (2) 3-D structural landmarks identified by geometric deep learning that differentiated ONHs among 4 groups: no glaucoma, mild glaucoma (mean deviation [MD], ≥-6.00 dB), moderate glaucoma (MD, -6.01 to -12.00 dB), and advanced glaucoma (MD, <-12.00 dB). Results: Study participants included 213 individuals without glaucoma (mean age, 63.4 years; 95% CI, 62.5-64.3 years; 126 females [59.2%]; 213 Chinese [100%] and 0 White individuals), 204 with mild glaucoma (mean age, 66.9 years; 95% CI, 66.0-67.8 years; 91 females [44.6%]; 178 Chinese [87.3%] and 26 White [12.7%] individuals), 118 with moderate glaucoma (mean age, 68.1 years; 95% CI, 66.8-69.4 years; 49 females [41.5%]; 97 Chinese [82.2%] and 21 White [17.8%] individuals), and 118 with advanced glaucoma (mean age, 68.5 years; 95% CI, 67.1-69.9 years; 43 females [36.4%]; 53 Chinese [44.9%] and 65 White [55.1%] individuals). The majority of ONH structural differences occurred in the early glaucoma stage, followed by a plateau effect in the later stages. Using a deep neural network, 3-D ONH structural differences were found to be present in both neural and connective tissues. Specifically, a mean of 57.4% (95% CI, 54.9%-59.9%, for no to mild glaucoma), 38.7% (95% CI, 36.9%-40.5%, for mild to moderate glaucoma), and 53.1 (95% CI, 50.8%-55.4%, for moderate to advanced glaucoma) of ONH landmarks that showed major structural differences were located in neural tissues with the remaining located in connective tissues. Conclusions and Relevance: This study uncovered complex 3-D structural differences of the ONH in both neural and connective tissues as a function of glaucoma severity. Future longitudinal studies should seek to establish a connection between specific 3-D ONH structural changes and fast visual field deterioration and aim to improve the early detection of patients with rapid visual field loss in routine clinical care.

Geometric Deep Learning to Identify the Critical 3D Structural Features of the Optic Nerve Head for Glaucoma Diagnosis.

PURPOSE: To compare the performance of 2 relatively recent geometric deep learning techniques in diagnosing glaucoma from a single optical coherence tomographic (OCT) scan of the optic nerve head (ONH); and to identify the 3-dimensional (3D) structural features of the ONH that are critical for the diagnosis of glaucoma. DESIGN: Comparison and evaluation of deep learning diagnostic algorithms. METHODS: In this study, we included a total of 2247 nonglaucoma and 2259 glaucoma scans from 1725 participants. All participants had their ONHs imaged in 3D with Spectralis OCT. All OCT scans were automatically segmented using deep learning to identify major neural and connective tissues. Each ONH was then represented as a 3D point cloud. We used PointNet and dynamic graph convolutional neural network (DGCNN) to diagnose glaucoma from such 3D ONH point clouds and to identify the critical 3D structural features of the ONH for glaucoma diagnosis. RESULTS: Both the DGCNN (area under the curve [AUC]: 0.97±0.01) and PointNet (AUC: 0.95±0.02) were able to accurately detect glaucoma from 3D ONH point clouds. The critical points (ie, critical structural features of the ONH) formed an hourglass pattern, with most of them located within the neuroretinal rim in the inferior and superior quadrant of the ONH. CONCLUSIONS: The diagnostic accuracy of both geometric deep learning approaches was excellent. Moreover, we were able to identify the critical 3D structural features of the ONH for glaucoma diagnosis that tremendously improved the transparency and interpretability of our method. Consequently, our approach may have strong potential to be used in clinical applications for the diagnosis and prognosis of a wide range of ophthalmic disorders.

The three-dimensional structural configuration of the central retinal vessel trunk and branches as a glaucoma biomarker.

PURPOSE: To assess whether the 3-dimensional (3D) structural configuration of the central retinal vessel trunk and its branches (CRVT&B) could be used as a diagnostic marker for glaucoma. DESIGN: Retrospective, deep-learning approach diagnosis study. METHODS: We trained a deep learning network to automatically segment the CRVT&B from the B-scans of the optical coherence tomography (OCT) volume of the optic nerve head. Subsequently, 2 different approaches were used for glaucoma diagnosis using the structural configuration of the CRVT&B as extracted from the OCT volumes. In the first approach, we aimed to provide a diagnosis using only 3D convolutional neural networks and the 3D structure of the CRVT&B. For the second approach, we projected the 3D structure of the CRVT&B orthographically onto sagittal, frontal, and transverse planes to obtain 3 two-dimensional (2D) images, and then a 2D convolutional neural network was used for diagnosis. The segmentation accuracy was evaluated using the Dice coefficient, whereas the diagnostic accuracy was assessed using the area under the receiver operating characteristic curves (AUCs). The diagnostic performance of the CRVT&B was also compared with that of retinal nerve fiber layer (RNFL) thickness (calculated in the same cohorts). RESULTS: Our segmentation network was able to efficiently segment retinal blood vessels from OCT scans. On a test set, we achieved a Dice coefficient of 0.81 ± 0.07. The 3D and 2D diagnostic networks were able to differentiate glaucoma from nonglaucoma subjects with accuracies of 82.7% and 83.3%, respectively. The corresponding AUCs for the CRVT&B were 0.89 and 0.90, higher than those obtained with RNFL thickness alone (AUCs ranging from 0.74 to 0.80). CONCLUSIONS: Our work demonstrated that the diagnostic power of the CRVT&B is superior to that of a gold-standard glaucoma parameter, that is, RNFL thickness. Our work also suggested that the major retinal blood vessels form a "skeleton"-the configuration of which may be representative of major optic nerve head structural changes as typically observed with the development and progression of glaucoma.

Acquired nasolacrimal duct obstruction: clinical and histological findings of 275 cases.

BACKGROUND: Acquired nasolacrimal duct obstruction is a blockage of the lacrimal outflow system usually caused by local nonspecific inflammation of the lacrimal sac and the nasolacrimal duct. However, cases exist where the primary nasolacrimal system obstruction is caused by malignancies. Our aim was to investigate lacrimal sac pathologies in patients with acquired nasolacrimal duct obstruction and compare their clinical manifestations. METHODS: This retrospective clinical study included 275 patients with acquired nasolacrimal duct obstruction who underwent external dacryocystorhinostomy and lacrimal sac biopsy. Cases were classified into tumor or nonspecific pathology groups and subdivided according to the level of inflammation. Histological and clinical data were analyzed. RESULTS: Three tumors (1.1%) (an adenoid cystic carcinoma, an eccrine spiradenoma and small B cell lymphoma) were diagnosed. Chronic nongranulomatous inflammation was the most common histological finding, corresponding to 194 cases (70.5%). The other 81 (29.5%) were subacute, acute forms of nongranulomatous inflammation, tumors and fibrosis cases. Epiphora with continuous purulent discharge was the most common clinical sign reported by 144 (52.4%) patients, and two (0.7%) patients had a palpable mass near the medial canthal tendon, which was identified as an eccrine spiradenoma and small B cell lymphoma. There was no significant difference in the clinical symptoms, duration or case history between the nonspecific pathology and tumor groups (p = 0.292). CONCLUSIONS: Chronic nongranulomatous inflammation of the lacrimal sac was the most common finding among acquired nasolacrimal duct obstruction cases. There were no associations between the histological findings and clinical presentation. The authors recommend a lacrimal sac biopsy only in cases when a tumor is clinically suspected.

Association Between Structure-function Characteristics and Visual Field Outcomes in Glaucoma Subjects With Intraocular Pressure Reduction After Trabeculectomy.

PRECIS: Improvements in post-trabeculectomy visual field (VF) outcomes were found to be significantly associated with preoperative nerve fiber layer thickness parameters extracted from the sectorized structure-function relationship, baseline VF, and severity of glaucoma. OBJECTIVE: To determine whether the preoperative structure-function relationship helps to predict visual outcomes at 1-year post-trabeculectomy. PATIENTS AND METHODS: In total, 91 eyes from 87 participants who successfully underwent trabeculectomy were included in our study. All eyes received optical coherence tomography imaging and VF assessment using 30-2 standard automated perimetry preoperatively at baseline and postoperatively 1 year after trabeculectomy. The linear mixed-model analysis was used to assess the association of structure and function at baseline, and multivariate analysis to investigate factors associated with postoperative VF outcomes. RESULTS: Results from multivariate and univariate analysis for VF 1 year after trabeculectomy showed that a positive preoperative retinal nerve fiber layer thickness deviation from the structure-function model was found to be significantly associated with improved postoperative VF outcomes [ß=0.06 dB/µm; 95% confidence interval (CI), 0.03-0.09]. Other significant factors included baseline VF MD (ß=-0.18; 95% CI, -0.23 to -0.13) and the presence of severe glaucoma (ß=-1.69; 95% CI, -2.80 to -0.57). Intraocular pressure was positively associated with improved VF outcomes only in univariate analysis (ß=0.06; 95% CI, 0.01-0.11). CONCLUSIONS AND RELEVANCE: Characteristics derived from the baseline structure-function relationship were found to be strongly associated with postoperative VF outcomes. These findings suggest that the structure-function relationship could potentially have a role in predicting VF progression after trabeculectomy.

One year structural and functional glaucoma progression after trabeculectomy.

We evaluated the changes in visual field mean deviation (VF MD) and retinal nerve fibre layer (RNFL) thickness in glaucoma patients undergoing trabeculectomy. One hundred patients were examined with VF and spectral-domain optical coherence tomography (OCT) before trabeculectomy and 4 follow-up visits over one year. Linear mixed models were used to investigate factors associated with VF and RNFL. VF improved during the first 3 months of follow-up (2.55 ± 1.06 dB/year) and worsened at later visits (-1.14 ± 0.29 dB/year). RNFL thickness reduced by -4.21 ± 0.25 µm/year from 1st month of follow-up. Eyes with an absence of initial VF improvement (ß = 0.64; 0.30-0.98), RNFL thinning (ß = 0.15; 0.08-0.23), increasing intraocular pressure (IOP; ß = -0.11; -0.18 to -0.03) and severe glaucoma (ß = -10.82; -13.61 to -8.02) were associated with VF deterioration. Eyes with VF deterioration (ß = 0.19; 0.08-0.29), increasing IOP (ß = -0.09; -0.17 to -0.01), and moderate (ß = -6.33; -12.17 to -0.49) or severe glaucoma (ß = -19.58; -24.63 to -14.52) were associated with RNFL thinning. Changes in RNFL structure and function occur over a 1-year follow-up period after trabeculectomy. Early VF improvement is more likely to occur in patients with mild/moderate glaucoma, whereas those with severe glaucoma show greater decline over one year. Our findings indicate that progression is observable using OCT, even in late-stage glaucoma.

Long-Term Shape, Curvature, and Depth Changes of the Lamina Cribrosa after Trabeculectomy.

PURPOSE: To evaluate changes in lamina cribrosa (LC) shape, curvature, and depth after trabeculectomy. DESIGN: Prospective, observational case series. PARTICIPANTS: A total of 112 patients (118 eyes) with open- or closed-angle glaucoma undergoing trabeculectomy. METHODS: The optic nerve head was imaged using enhanced depth imaging spectral-domain OCT before trabeculectomy and at 6 follow-up visits throughout the first postoperative year. The anterior LC surface and Bruch's membrane opening were marked in the serial horizontal B scans for the analysis of LC parameters using Morphology 1.0 software. Postoperative morphologic LC changes were assessed. MAIN OUTCOME MEASURES: The postoperative LC global shape index (GSI), nasal-temporal (N-T) and superior-inferior (S-I) curvatures, and mean and sectoral LC depth (LCD). RESULTS: The mean LC GSI increased only during the early postoperative period (P = 0.02), resulting in a change toward the saddle-rut shape. There was a flattening of the LC curvature in N-T (P < 0.001) and S-I (P = 0.003) meridians 12 months after trabeculectomy. A shallowing of the mean and sectoral LCD from baseline was significant throughout the entire follow-up period (P < 0.001) and progressed up to postoperative month 6. Twenty-eight patients showed a deepening of the LC from baseline in at least 1 visit. Eyes with shallower LCD compared with baseline responded to intraocular pressure (IOP) reduction with greater movement anteriorly than eyes with deeper LCD (P = 0.002). Greater IOP reduction (P = 0.007), less retinal nerve fiber layer thinning over the year (P = 0.003), and more superiorly-inferiorly curved baseline LC (P = 0.001) were associated with an increase in GSI. Younger age and IOP reduction were related to LC shallowing (P < 0.001, P = 0.002) and N-T flattening (P < 0.001). CONCLUSIONS: In most eyes, trabeculectomy resulted in long-term flattening and shallowing of the LC. However, in some eyes, LC deepened from baseline. Change in LC global shape appeared to be temporal. Reduction in IOP plays an important role in the early phase of LC change; however, in the later phase, LC remodeling may play a crucial role in view of stable IOP.

Changes in macular thickness after trabeculectomy with or without adjunctive 5-fluorouracil.

CONFLICT OF INTEREST: None of the authors have any conflict of interest to declare, financial or otherwise. No financial or other support was received for the study. BACKGROUND: The aim of the study was to assess changes in macular thickness after trabeculectomy in respect to the use of 5-fluorouracil (5-FU) as well as to analyse possible associations between the postoperative changes in macular thickness and intraocular pressure (IOP). MATERIALS AND METHODS: The prospective observational study included 106 eyes (100 patients) with glaucoma who underwent trabeculectomy with or without 5-FU. Subsequently 5-FU needling was performed if failure of the filtrating bleb occurred. Macular thickness and the IOP were evaluated before, one week, and six months after the surgery. The mean and sectoral macular thickness was assessed using spectral domain optical coherence tomography. RESULTS: The mean (±SD) IOP reduced from 27.71 (±6.88) mmHg at baseline to 18.3 (±8.1) mmHg one week (p < 0.001) and 15.1 (±7.6) mmHg six months (p < 0.001) after trabeculectomy. One week postoperatively, the mean macular thickness increased from 285.19 (±15.98) µm to 288.9 (±16.31) µm (p < 0.001); macular thickening was significant in all subfields (p < 0.001) and correlated positively with IOP reduction (rho = 0.312, p = 0.001 for central subfield). After six months, macula remained thicker only at the central and inner nasal subfields (p < 0.05). The changes in macular thickness were not affected by the use of 5-fluorouracil. CONCLUSIONS: Trabeculectomy may induce a slight macular thickening which is more pronounced in the early postoperative period. The IOP reduction plays an important role in this process and is associated with thicker postoperative macula. However, the use of adjunctive 5-FU has no influence on macular thickness after glaucoma surgery despite its potential hypotonic, inflammatory and cytotoxic effects.

Changes in choroidal thickness after intraocular pressure reduction following trabeculectomy.

PURPOSE: To evaluate the changes of peripapillary and subfoveal choroidal thickness (CT) after trabeculectomy. METHODS: Prospective longitudinal study included 37 eyes with open-angle glaucoma. The subfoveal and peripapillary CT was measured using enhanced depth imaging spectral domain optical coherence tomography before trabeculectomy and 1 week, 3 and 6 months postoperatively. The associations between changes in the CT, intraocular pressure (IOP) and axial length were analysed. RESULTS: The medium subfoveal CT (IQR) increased from 182 (97) µm at baseline to 267 (107) µm 1 week, 213 (97) µm 3 months and 207 (91) µm 6 months postoperatively (p < 0.001). The peripapillary CT increased in all four quadrants at all follow-ups (p < 0.05). The subfoveal and peripapillary choroidal thickening correlated with the magnitude of IOP reduction (p < 0.05) and axial length shortening (p < 0.01) during whole follow-up period. There was a peripapillary CT increase of 2.9 µm per mmHg of IOP reduction (p < 0.001, CI 1.5-4.4) and 4.8 µm per mm of baseline axial length (p = 0.049, CI 0.03-9.6) 1 week postoperatively after adjustment for baseline IOP. Six months postoperatively, the decrease in axial length was the only factor associated with peripapillary choroidal thickening (p = 0.031; regression coefficient: 73.29 µm/mm, CI 7.1-139.5). CONCLUSION: Intraocular pressure (IOP) reduction after trabeculectomy caused the increase in subfoveal and peripapillary CT for at least 6 months postoperatively correlating with greater IOP reduction and axial length shortening. In the long term, the decrease in axial eye length, but not IOP, was the only factor to be associated with peripapillary choroidal thickening.

Quantitative histopathological assessment of ocular surface squamous neoplasia using digital image analysis.

The aim of this retrospective pilot study was to evaluate the Aperio nuclear V9 algorithm as an image analysis tool to observe the histopathological changes of ocular surface squamous neoplasia (OSSN). A histopathological assessment, including the Ki-67 proliferative index (PI) of immunohistochemically-stained tumor conjunctiva (TC) and healthy conjunctiva (HC) tissues, was performed in six cases of OSSN. The Aperio V9 algorithm was applied to digital images of the tissue specimens to count the Ki-67 PI and to measure the nuclear area indices. This digital algorithm was validated using stereological and visual analysis methods. The visual scoring of Ki-67 PI ranged from 22 to 60% (mean, 38.5%), and from 5 to 20% (mean 9.5%) in TC and HC tissue, respectively. The computer-aided analysis, using the Aperio nuclear V9 algorithm, revealed that the Ki-67 PI ranged from 21.5 to 43.5% (mean, 33.6%), and from 1.9 to 21.0% (mean, 11.8%) in the TC and HC tissue, respectively. The stereological method demonstrated that the Ki-67 PI ranged from 30.1 to 51.5% (mean, 41.0%), and from 3.2 to 30.1% (mean, 15.1%) in the TC and HC tissues, respectively. The strongest association in the collinearity of regression analysis was observed between the Aperio nuclear V9 algorithm/stereological models in the TC tissue (r2=0.7; P=0.04) and the HC tissue (r2=0.7; P=0.03), and the visual/stereological models in the TC tissue (r2=0.7; P=0.04) and the visual/Aperio nuclear V9 algorithm models in the HC tissue (r2=0.7; P=0.04). A weak and statistically insignificant association was identified between the visual/Aperio nuclear V9 algorithm analysis in the TC tissue (r2=0.4; P=0.2) and the visual/stereological models in the HC tissue (r2=0.5; P=0.13). No significant difference was observed between the nuclear area of the TC (mean, 36.5 µm2) and HC (mean, 35.7 µm2; P=0.88) tissues. It was concluded that the Aperio nuclear V9 algorithm is a useful tool for the reliable analysis of histopathological changes of OSSN. The results of this computer-aided algorithm correlate strongly with the stereological method when assessing the Ki-67 PI.