Don-Il Ham
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Don-Il Ham has a total of 74 co-authors. Here are the top co-authors by number of shared publications:
Mingui Kong
Department of Ophthalmology, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 29, Saemunan-ro, Jongno-gu, Seoul, 03181, Korea
14 shared publicationsJe Moon Yoon
Department of Ophthalmology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea
10 shared publicationsHoon Noh
HanGil Eye Hospital, Incheon, Korea
8 shared publicationsSeung Wan Nam
Good Morning Light Eye Clinic, Ulsan, Korea
8 shared publicationsSungsoon Hwang
Department of Ophthalmology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea
6 shared publications- Je Moon Yoon
Department of Ophthalmology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
5 shared publications
Latest Publications
- Open Access
- 6 May 2026
- Published by Springer Nature in Eye
- Vol. 40 (10) , 1454-1460
Multimodal deep learning prediction of treatment response to anti-vascular endothelial growth factor in diabetic macular oedema
Objective To develop and validate a multimodal deep learning model that predicts treatment responses to intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections in patients with diabetic macular oedema (DMO) by combining optical coherence tomography images and clinical data. Methods This study included 107 DMO patients who received three consecutive anti-VEGF treatments. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, and specificity. The model’s predictions were compared with those of retinal specialists. Results Among 107 patients, 65 showed good response and 42 showed poor response to treatment. The multimodal model achieved an AUROC of 0.962 (95% CI, 0.945–0.979), accuracy of 0.953 (95% CI, 0.933–0.973), sensitivity of 0.969 (95% CI, 0.951–0.987), and specificity of 0.928 (95% CI, 0.903–0.953) in the internal validation. The model outperformed retinal specialists, who achieved accuracies ranging from 0.571 to 0.857. Conclusion The multimodal deep learning model demonstrated high accuracy in predicting anti-VEGF treatment responses in DMO patients. This approach could enable more personalised treatment strategies and optimal resource utilisation in ophthalmological care. Further validation with larger, multicentre datasets is warranted to confirm its clinical utility.
- 6 May 2026
- Published by Springer Nature in Eye
- Vol. 40 (10) , 1454-1460
Open AccessMultimodal deep learning prediction of treatment response to anti-vascular endothelial growth factor in diabetic macular oedema
Objective To develop and validate a multimodal deep learning model that predicts treatment responses to intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections in patients with diabetic macular oedema (DMO) by combining optical coherence tomography images and clinical data. Methods This study included 107 DMO patients who received three consecutive anti-VEGF treatments. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, and specificity. The model’s predictions were compared with those of retinal specialists. Results Among 107 patients, 65 showed good response and 42 showed poor response to treatment. The multimodal model achieved an AUROC of 0.962 (95% CI, 0.945–0.979), accuracy of 0.953 (95% CI, 0.933–0.973), sensitivity of 0.969 (95% CI, 0.951–0.987), and specificity of 0.928 (95% CI, 0.903–0.953) in the internal validation. The model outperformed retinal specialists, who achieved accuracies ranging from 0.571 to 0.857. Conclusion The multimodal deep learning model demonstrated high accuracy in predicting anti-VEGF treatment responses in DMO patients. This approach could enable more personalised treatment strategies and optimal resource utilisation in ophthalmological care. Further validation with larger, multicentre datasets is warranted to confirm its clinical utility.
- Open Access
- 12 August 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (8) , e0329907
Unveiling the drivers of geographic atrophy progression in eyes with reticular pseudodrusen
Purpose: To evaluate the progression rate of geographic atrophy (GA) and identify factors associated with GA expansion in eyes with reticular pseudodrusen (RPD) using fundus autofluorescence (FAF) photography. Methods: A total of 28 eyes from 28 patients diagnosed with GA and RPD, who completed a 3-year follow-up, was included. The eyes underwent thorough examination with color fundus photography, FAF and near infrared (NIR) imaging, and spectral domain optical coherence tomography (SD OCT). The areas of atrophic lesions were quantified using FAF images at years 1, 2, and 3 employing a semi-automated software (Region Finder). Results: The rate of increase in mean GA area in eyes with RPD was calculated to be 0.39 ± 0.21 mm per year over the 3-year period. The annual progression rate measured over three years showed a statistically significant increase each year. Univariate analysis revealed that subfoveal choroidal thickness (SFCT) and RPD distribution were independently associated with GA progression (P ≤ 0.037). In multivariate analysis, thin SFCT was the only significant factor linked to an increased rate of GA growth. The progression from extrafoveal GA to foveal GA was solely associated with the initial square root GA area. Conclusion: These results offer valuable insights for the development of interventional strategies aimed at mitigating GA enlargement in this specific subgroup of patients.
- 12 August 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (8) , e0329907
Open AccessUnveiling the drivers of geographic atrophy progression in eyes with reticular pseudodrusen
Purpose: To evaluate the progression rate of geographic atrophy (GA) and identify factors associated with GA expansion in eyes with reticular pseudodrusen (RPD) using fundus autofluorescence (FAF) photography. Methods: A total of 28 eyes from 28 patients diagnosed with GA and RPD, who completed a 3-year follow-up, was included. The eyes underwent thorough examination with color fundus photography, FAF and near infrared (NIR) imaging, and spectral domain optical coherence tomography (SD OCT). The areas of atrophic lesions were quantified using FAF images at years 1, 2, and 3 employing a semi-automated software (Region Finder). Results: The rate of increase in mean GA area in eyes with RPD was calculated to be 0.39 ± 0.21 mm per year over the 3-year period. The annual progression rate measured over three years showed a statistically significant increase each year. Univariate analysis revealed that subfoveal choroidal thickness (SFCT) and RPD distribution were independently associated with GA progression (P ≤ 0.037). In multivariate analysis, thin SFCT was the only significant factor linked to an increased rate of GA growth. The progression from extrafoveal GA to foveal GA was solely associated with the initial square root GA area. Conclusion: These results offer valuable insights for the development of interventional strategies aimed at mitigating GA enlargement in this specific subgroup of patients.
- Open Access
- 29 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 27553
Comprehensive analysis of pachyvessel morphology in central serous chorioretinopathy using multimodal imaging
The purpose of this study was to identify characteristic morphological features of large choroidal blood vessels (pachyvessels) in eyes with central serous chorioretinopathy (CSC) using enhanced depth imaging (EDI) optical coherence tomography (OCT). A total of 116 eyes from 58 patients with CSC and 116 eyes from 58 age- and gender-matched healthy subjects were included. EDI OCT raster scan images were analyzed for the presence of characteristic features of pachyvessels and accompanying retinal pigment epithelial (RPE) abnormalities, with additional imaging data obtained from fluorescein angiography (FA) and indocyanine green angiography (ICGA). Severely dilated pachyvessels abutting the RPE-Bruch's membrane-choriocapillaris complex, referred to as the "Hanging Drop (HD) sign," were identified in all acute CSC eyes and 86.4% of CSC eyes overall. The HD sign was observed either with ("active") or without ("inactive") vascular wall hyperreflectivity along the RPE-Bruch's membrane-choriocapillaris complex. The active HD sign was present exclusively in CSC eyes (56.1%) and was significantly more frequent in acute CSC compared to chronic and resolved CSC (88.9%, 47.4%, and 17.6%, respectively; P < 0.05). While the inactive HD sign was also more common in CSC eyes than in unaffected fellow eyes and healthy eyes (77.3%, 30.0%, and 9.5%, respectively), its prevalence did not differ significantly among CSC eyes with varying clinical conditions (P > 0.05). RPE abnormalities were consistently observed in OCT images showing the active HD sign. These findings suggest that characteristic morphological features of pachyvessels discerned using EDI OCT raster scans can aid in the diagnosis and evaluation of CSC.
- 29 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 27553
Open AccessComprehensive analysis of pachyvessel morphology in central serous chorioretinopathy using multimodal imaging
The purpose of this study was to identify characteristic morphological features of large choroidal blood vessels (pachyvessels) in eyes with central serous chorioretinopathy (CSC) using enhanced depth imaging (EDI) optical coherence tomography (OCT). A total of 116 eyes from 58 patients with CSC and 116 eyes from 58 age- and gender-matched healthy subjects were included. EDI OCT raster scan images were analyzed for the presence of characteristic features of pachyvessels and accompanying retinal pigment epithelial (RPE) abnormalities, with additional imaging data obtained from fluorescein angiography (FA) and indocyanine green angiography (ICGA). Severely dilated pachyvessels abutting the RPE-Bruch's membrane-choriocapillaris complex, referred to as the "Hanging Drop (HD) sign," were identified in all acute CSC eyes and 86.4% of CSC eyes overall. The HD sign was observed either with ("active") or without ("inactive") vascular wall hyperreflectivity along the RPE-Bruch's membrane-choriocapillaris complex. The active HD sign was present exclusively in CSC eyes (56.1%) and was significantly more frequent in acute CSC compared to chronic and resolved CSC (88.9%, 47.4%, and 17.6%, respectively; P < 0.05). While the inactive HD sign was also more common in CSC eyes than in unaffected fellow eyes and healthy eyes (77.3%, 30.0%, and 9.5%, respectively), its prevalence did not differ significantly among CSC eyes with varying clinical conditions (P > 0.05). RPE abnormalities were consistently observed in OCT images showing the active HD sign. These findings suggest that characteristic morphological features of pachyvessels discerned using EDI OCT raster scans can aid in the diagnosis and evaluation of CSC.
- Open Access
- 1 July 2025
- Published by Springer Nature in BMC Ophthalmology
- Vol. 25 (1) , 1-11
Deep learning-based automated classification of choroidal layers in en face swept-source optical coherence tomography images
This study aims to develop a deep learning-based algorithm dedicated to the automated classification of choroidal layers in en face swept-source optical coherence tomography (SS-OCT) images of the eye. A cohort of 117 healthy subjects (117 eyes), underwent SS-OCT volume scans covering a 12 × 9 mm range. En face SS-OCT images of the choroid were acquired at 2.6 μm intervals from Bruch’s membrane to the chorioscleral border. The images were classified into 5 layers using landmarks, such as the start of the choriocapillaris, the start of Sattler’s layer, and the start and end of Haller’s layer. The dataset consisted of 16,025 en face SS-OCT images, and model performance was evaluated using 5-fold cross validation. We developed the deep learning system based on ResNet, incorporating boundary-enhancing undersampling and subclass ensemble techniques. The developed algorithm successfully classified choroidal layers with a balanced accuracy of 84.30% and 92.61% within error ranges of 0 μm and 5.2 μm, respectively. Automated stratification of choroidal layers from en face SS-OCT images can be achieved accurately through deep learning.
- 1 July 2025
- Published by Springer Nature in BMC Ophthalmology
- Vol. 25 (1) , 1-11
Open AccessDeep learning-based automated classification of choroidal layers in en face swept-source optical coherence tomography images
This study aims to develop a deep learning-based algorithm dedicated to the automated classification of choroidal layers in en face swept-source optical coherence tomography (SS-OCT) images of the eye. A cohort of 117 healthy subjects (117 eyes), underwent SS-OCT volume scans covering a 12 × 9 mm range. En face SS-OCT images of the choroid were acquired at 2.6 μm intervals from Bruch’s membrane to the chorioscleral border. The images were classified into 5 layers using landmarks, such as the start of the choriocapillaris, the start of Sattler’s layer, and the start and end of Haller’s layer. The dataset consisted of 16,025 en face SS-OCT images, and model performance was evaluated using 5-fold cross validation. We developed the deep learning system based on ResNet, incorporating boundary-enhancing undersampling and subclass ensemble techniques. The developed algorithm successfully classified choroidal layers with a balanced accuracy of 84.30% and 92.61% within error ranges of 0 μm and 5.2 μm, respectively. Automated stratification of choroidal layers from en face SS-OCT images can be achieved accurately through deep learning.
- 3 May 2025
- Published by Springer Nature in Graefe's Archive for Clinical and Experimental Ophthalmology
- Vol. 263 (8) , 2209-2218
Peripheral retinal and choroidal thickness of eyes with reticular pseudodrusen
Purpose This study aimed to investigate the peripheral retinal and choroidal thicknesses in eyes with reticular pseudodrusen (RPD). Methods Imaging data of ultrawide-field swept-source optical coherence tomography taken from 40 patients with age-related macular degeneration (AMD) were analyzed. Thirty-nine eyes of 20 patients had RPD (RPD group) and 39 eyes of 20 patients had no RPD (control group). Eyes with RPD were divided into localized/intermediate and diffuse types according to the fundus distribution. The retinal and choroidal thicknesses were measured at the fovea and six peripheral points in each eye, and the results were statistically analyzed. Results The mean ages were 78.0 ± 6.1 years in the RPD group and 74.4 ± 7.2 years in the control group. The RPD group showed a thinner retina only at the nasal point (p = 0.002) than the control group. The choroid was significantly thinner in the RPD group than in the control group at most peripheral points, except for the far temporal and far inferior points. Diffuse and localized/intermediate types of RPD eyes showed no significant difference in choroidal thickness, except at the nasal point, which was thinner in the diffuse type (p = 0.049). Conclusion In eyes with RPD, most peripheral retinal thicknesses were not different from those of eyes with AMD without RPD; however, most peripheral choroidal thicknesses were thinner regardless of the fundus distribution status of the RPD. Nasal choroidal thickness may be associated with RPD severity.
- 3 May 2025
- Published by Springer Nature in Graefe's Archive for Clinical and Experimental Ophthalmology
- Vol. 263 (8) , 2209-2218
Peripheral retinal and choroidal thickness of eyes with reticular pseudodrusen
Purpose This study aimed to investigate the peripheral retinal and choroidal thicknesses in eyes with reticular pseudodrusen (RPD). Methods Imaging data of ultrawide-field swept-source optical coherence tomography taken from 40 patients with age-related macular degeneration (AMD) were analyzed. Thirty-nine eyes of 20 patients had RPD (RPD group) and 39 eyes of 20 patients had no RPD (control group). Eyes with RPD were divided into localized/intermediate and diffuse types according to the fundus distribution. The retinal and choroidal thicknesses were measured at the fovea and six peripheral points in each eye, and the results were statistically analyzed. Results The mean ages were 78.0 ± 6.1 years in the RPD group and 74.4 ± 7.2 years in the control group. The RPD group showed a thinner retina only at the nasal point (p = 0.002) than the control group. The choroid was significantly thinner in the RPD group than in the control group at most peripheral points, except for the far temporal and far inferior points. Diffuse and localized/intermediate types of RPD eyes showed no significant difference in choroidal thickness, except at the nasal point, which was thinner in the diffuse type (p = 0.049). Conclusion In eyes with RPD, most peripheral retinal thicknesses were not different from those of eyes with AMD without RPD; however, most peripheral choroidal thicknesses were thinner regardless of the fundus distribution status of the RPD. Nasal choroidal thickness may be associated with RPD severity.
- Open Access
- 30 January 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (1) , e0318369
Genetic influence on choroidal vascularity index
Purpose: To investigate the heritability of genetic influence on macular choroidal vascularity index (CVI). Methods: Total choroidal area (TCA), luminal area (LA), and CVI was measured using horizontal scan of spectral-domain optical coherence tomography with enhanced depth imaging in the 373 healthy twin participants. Characteristics of the participants were investigated, including age, sex, axial length, hypertension, diabetes, drinking habits, and smoking status. Univariate and subsequent multivariate regression analyses were performed to evaluate the associations of these factors with TCA, LA, and CVI. Results: Patients who were older and had a higher intraocular pressure, higher diastolic blood pressure, and lower pulse pressure showed associations with lower CVI (p<0.001, p = 0.014, p = 0.005, and p = 0.015, respectively). The covariate-adjusted heritability (±standard error) of the CVI was 0.716 ± 0.091, and the heritabilities of the TCA and LA were 0.691 ± 0.089 and 0.634 ± 0.100, respectively. Conclusion: The TCA, LA, and CVI are highly heritable.
- 30 January 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (1) , e0318369
Open AccessGenetic influence on choroidal vascularity index
Purpose: To investigate the heritability of genetic influence on macular choroidal vascularity index (CVI). Methods: Total choroidal area (TCA), luminal area (LA), and CVI was measured using horizontal scan of spectral-domain optical coherence tomography with enhanced depth imaging in the 373 healthy twin participants. Characteristics of the participants were investigated, including age, sex, axial length, hypertension, diabetes, drinking habits, and smoking status. Univariate and subsequent multivariate regression analyses were performed to evaluate the associations of these factors with TCA, LA, and CVI. Results: Patients who were older and had a higher intraocular pressure, higher diastolic blood pressure, and lower pulse pressure showed associations with lower CVI (p<0.001, p = 0.014, p = 0.005, and p = 0.015, respectively). The covariate-adjusted heritability (±standard error) of the CVI was 0.716 ± 0.091, and the heritabilities of the TCA and LA were 0.691 ± 0.089 and 0.634 ± 0.100, respectively. Conclusion: The TCA, LA, and CVI are highly heritable.
- 29 June 2024
- Published by Springer Nature in Eye
- Vol. 38 (15) , 2960-2967
Choroidal venous overload in Vogt‒Koyanagi‒Harada disease
Background/Objectives This study aimed to investigate the change of choroidal venous overload in Vogt‒Koyanagi‒Harada (VKH) disease. Clinical records of 52 patients with VKH disease (52 eyes) and 24 control subjects (24 eyes) who underwent multimodal imaging, including fluorescein angiography (FA) and indocyanine green angiography (ICGA), were retrospectively reviewed. Subjects/Methods Imaging data were assessed for signs associated with choroidal venous overload, e.g., choroidal perfusion delay, choroidal vascular hyperpermeability, dilated choroidal veins, and intervortex venous anastomosis (IVA). Dual FA and ICGA scoring for active posterior segment inflammation was performed. Clinical and imaging features associated with choroidal venous overload were compared between early- and late-stage VKH disease. Results Choroidal perfusion delay, choroidal vascular hyperpermeability, dilated choroidal veins, and IVA were more prevalent in eyes with VKH disease (69.2%, 67.3%, 61.5%, and 65.4%, respectively) than in control eyes (25.0%, 20.8%, 25.0%, and 37.5%, respectively) (p < 0.05). All eyes with IVA in the early-stage of VKH disease had got other 3 signs. All choroidal venous overload signs were more prevalent in patients with early-stage (20 eyes) than in those with late-stage VKH disease (32 eyes) (p < 0.05). The number of choroidal venous overload signs were inversely related to disease duration (p < 0.001) and proportionally related to the total ICGA score (p < 0.001). IVA was significantly associated with the total ICGA score in logistic regression (p = 0.014). Conclusions Choroidal venous overload occurs early in VKH disease. Angiographic signs of choroidal venous overload may be useful markers to assess the status of VKH disease.
- 29 June 2024
- Published by Springer Nature in Eye
- Vol. 38 (15) , 2960-2967
Choroidal venous overload in Vogt‒Koyanagi‒Harada disease
Background/Objectives This study aimed to investigate the change of choroidal venous overload in Vogt‒Koyanagi‒Harada (VKH) disease. Clinical records of 52 patients with VKH disease (52 eyes) and 24 control subjects (24 eyes) who underwent multimodal imaging, including fluorescein angiography (FA) and indocyanine green angiography (ICGA), were retrospectively reviewed. Subjects/Methods Imaging data were assessed for signs associated with choroidal venous overload, e.g., choroidal perfusion delay, choroidal vascular hyperpermeability, dilated choroidal veins, and intervortex venous anastomosis (IVA). Dual FA and ICGA scoring for active posterior segment inflammation was performed. Clinical and imaging features associated with choroidal venous overload were compared between early- and late-stage VKH disease. Results Choroidal perfusion delay, choroidal vascular hyperpermeability, dilated choroidal veins, and IVA were more prevalent in eyes with VKH disease (69.2%, 67.3%, 61.5%, and 65.4%, respectively) than in control eyes (25.0%, 20.8%, 25.0%, and 37.5%, respectively) (p < 0.05). All eyes with IVA in the early-stage of VKH disease had got other 3 signs. All choroidal venous overload signs were more prevalent in patients with early-stage (20 eyes) than in those with late-stage VKH disease (32 eyes) (p < 0.05). The number of choroidal venous overload signs were inversely related to disease duration (p < 0.001) and proportionally related to the total ICGA score (p < 0.001). IVA was significantly associated with the total ICGA score in logistic regression (p = 0.014). Conclusions Choroidal venous overload occurs early in VKH disease. Angiographic signs of choroidal venous overload may be useful markers to assess the status of VKH disease.
- Open Access
- 13 March 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-9
Advanced age-related macular degeneration and risk factors in eyes with pachydrusen
The risk of progression to advanced age-related macular degeneration (AMD) varies depending on the type of drusen. This retrospective longitudinal study included 248 eyes of 156 patients with pachydrusen without advanced AMD at baseline. Macular neovascularization (MNV) and geographic atrophy (GA) were evaluated. Risk factors for progression to advanced AMD were determined using multivariate Cox regression analysis. The mean age at baseline was 65.4 ± 9.1 years, and the mean follow-up duration was 6.40 ± 3.58 years. The mean total number of pachydrusen and macular pachydrusen were 4.10 ± 2.85 and 2.27 ± 1.81 per eye, respectively. Pachydrusen was accompanied by other types of drusen in 4.8% (12 eyes) of eyes at baseline. During follow-up, MNVs occurred in 2.8% (seven eyes), including polypoidal choroidal vasculopathy (PCV six eyes); however, no GA occurred. Regarding risk factors for progression to neovascular AMD, age (p = 0.023) and macular pigmentary changes (p = 0.014) were significantly associated with MNV development. The cumulative incidence of MNV was significantly higher in the group with macular pigmentary changes (17.39% vs. 0.57% at 10 years; p = 0.0005). The number of macular pachydrusen and the presence of MNV in the fellow eye did not show a statistically significant relationship with MNV development. Age and macular pigmentary changes are risk factors for MNV development in the eyes with pachydrusen. Eyes with pachydrusen appear to have a risk profile for advanced AMD that is different from that of AMD eyes with drusen or drusenoid deposits other than pachydrusen.
- 13 March 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-9
Open AccessAdvanced age-related macular degeneration and risk factors in eyes with pachydrusen
The risk of progression to advanced age-related macular degeneration (AMD) varies depending on the type of drusen. This retrospective longitudinal study included 248 eyes of 156 patients with pachydrusen without advanced AMD at baseline. Macular neovascularization (MNV) and geographic atrophy (GA) were evaluated. Risk factors for progression to advanced AMD were determined using multivariate Cox regression analysis. The mean age at baseline was 65.4 ± 9.1 years, and the mean follow-up duration was 6.40 ± 3.58 years. The mean total number of pachydrusen and macular pachydrusen were 4.10 ± 2.85 and 2.27 ± 1.81 per eye, respectively. Pachydrusen was accompanied by other types of drusen in 4.8% (12 eyes) of eyes at baseline. During follow-up, MNVs occurred in 2.8% (seven eyes), including polypoidal choroidal vasculopathy (PCV six eyes); however, no GA occurred. Regarding risk factors for progression to neovascular AMD, age (p = 0.023) and macular pigmentary changes (p = 0.014) were significantly associated with MNV development. The cumulative incidence of MNV was significantly higher in the group with macular pigmentary changes (17.39% vs. 0.57% at 10 years; p = 0.0005). The number of macular pachydrusen and the presence of MNV in the fellow eye did not show a statistically significant relationship with MNV development. Age and macular pigmentary changes are risk factors for MNV development in the eyes with pachydrusen. Eyes with pachydrusen appear to have a risk profile for advanced AMD that is different from that of AMD eyes with drusen or drusenoid deposits other than pachydrusen.
- Open Access
- 22 January 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-16
Enhancing foveal avascular zone analysis for Alzheimer’s diagnosis with AI segmentation and machine learning using multiple radiomic features
We propose a hybrid technique that employs artificial intelligence (AI)-based segmentation and machine learning classification using multiple features extracted from the foveal avascular zone (FAZ)—a retinal biomarker for Alzheimer’s disease—to improve the disease diagnostic performance. Imaging data of optical coherence tomography angiography from 37 patients with Alzheimer’s disease and 48 healthy controls were investigated. The presence or absence of brain amyloids was confirmed using amyloid positron emission tomography. In the superficial capillary plexus of the angiography scans, the FAZ was automatically segmented using an AI method to extract multiple biomarkers (area, solidity, compactness, roundness, and eccentricity), which were paired with clinical data (age and sex) as common correction variables. We used a light-gradient boosting machine (a light-gradient boosting machine is a machine learning algorithm based on trees utilizing gradient boosting) to diagnose Alzheimer’s disease by integrating the corresponding multiple radiomic biomarkers. Fivefold cross-validation was applied for analysis, and the diagnostic performance for Alzheimer’s disease was determined by the area under the curve. The proposed hybrid technique achieved an area under the curve of $$72.2\pm 4.2$$ %, outperforming the existing single-feature (area) criteria by over 13%. Furthermore, in the holdout test set, the proposed technique exhibited a 14% improvement compared to single features, achieving an area under the curve of 72.0± 4.8%. Based on these facts, we have demonstrated the effectiveness of our technology in achieving significant performance improvements in FAZ-based Alzheimer’s diagnosis research through the use of multiple radiomic biomarkers (area, solidity, compactness, roundness, and eccentricity).- 22 January 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-16
Open AccessEnhancing foveal avascular zone analysis for Alzheimer’s diagnosis with AI segmentation and machine learning using multiple radiomic features
We propose a hybrid technique that employs artificial intelligence (AI)-based segmentation and machine learning classification using multiple features extracted from the foveal avascular zone (FAZ)—a retinal biomarker for Alzheimer’s disease—to improve the disease diagnostic performance. Imaging data of optical coherence tomography angiography from 37 patients with Alzheimer’s disease and 48 healthy controls were investigated. The presence or absence of brain amyloids was confirmed using amyloid positron emission tomography. In the superficial capillary plexus of the angiography scans, the FAZ was automatically segmented using an AI method to extract multiple biomarkers (area, solidity, compactness, roundness, and eccentricity), which were paired with clinical data (age and sex) as common correction variables. We used a light-gradient boosting machine (a light-gradient boosting machine is a machine learning algorithm based on trees utilizing gradient boosting) to diagnose Alzheimer’s disease by integrating the corresponding multiple radiomic biomarkers. Fivefold cross-validation was applied for analysis, and the diagnostic performance for Alzheimer’s disease was determined by the area under the curve. The proposed hybrid technique achieved an area under the curve of $$72.2\pm 4.2$$ %, outperforming the existing single-feature (area) criteria by over 13%. Furthermore, in the holdout test set, the proposed technique exhibited a 14% improvement compared to single features, achieving an area under the curve of 72.0± 4.8%. Based on these facts, we have demonstrated the effectiveness of our technology in achieving significant performance improvements in FAZ-based Alzheimer’s diagnosis research through the use of multiple radiomic biomarkers (area, solidity, compactness, roundness, and eccentricity). - Open Access
- 29 September 2023
- Published by Springer Nature in Eye
- Vol. 38 (4) , 691-697
Macular lesions associated with age-related macular degeneration in pachydrusen eyes
Background To investigate the prevalence of macular lesions associated with age-related macular degeneration (AMD) in eyes with pachydrusen. Methods Clinical records and multimodal imaging data of patients over 50 years old with drusen or drusenoid deposits were retrospectively assessed, and eyes with pachydrusen were included in this study. The presence of AMD features, including drusen or drusenoid deposits, macular pigmentary abnormalities, geographic atrophy (GA), and macular neovascularization (MNV), were evaluated. Results Out of 967 eyes of 494 patients with drusen or drusenoid deposits, 330 eyes of 183 patients had pachydrusen (34.1%). The mean age was 66.1 ± 9.3 years, and the subfoveal choroidal thickness (SFCT) was 292.7 ± 100.1 μm. The mean number of pachydrusen per eye was 2.22 ± 1.73. The majority of eyes with pachydrusen had no other drusen or drusenoid deposits (95.2%). Only 16 eyes (4.8%) had other deposits, including soft drusen (10 eyes, 3.0%), cuticular drusen (3 eyes, 0.9%), and reticular pseudodrusen (RPD; 3 eyes, 0.9%). Macular pigmentary abnormalities accompanied pachydrusen in 68 eyes (27.4%). None of the eyes had GA, and 82 eyes (24.8%) had MNV. The majority of MNV was polypoidal choroidal vasculopathy (PCV; 65 eyes, 19.7%), followed by type 1 (10 eyes, 3.0%), type 2 (5 eyes, 1.5%), and type 3 MNV (2 eyes, 0.6%). Conclusions Eyes with pachydrusen in Korean population have several characteristic AMD lesions in low frequencies. These findings indicate that pachydrusen might have diagnostic and prognostic values that are different from those of other drusen or drusenoid deposits.
- 29 September 2023
- Published by Springer Nature in Eye
- Vol. 38 (4) , 691-697
Open AccessMacular lesions associated with age-related macular degeneration in pachydrusen eyes
Background To investigate the prevalence of macular lesions associated with age-related macular degeneration (AMD) in eyes with pachydrusen. Methods Clinical records and multimodal imaging data of patients over 50 years old with drusen or drusenoid deposits were retrospectively assessed, and eyes with pachydrusen were included in this study. The presence of AMD features, including drusen or drusenoid deposits, macular pigmentary abnormalities, geographic atrophy (GA), and macular neovascularization (MNV), were evaluated. Results Out of 967 eyes of 494 patients with drusen or drusenoid deposits, 330 eyes of 183 patients had pachydrusen (34.1%). The mean age was 66.1 ± 9.3 years, and the subfoveal choroidal thickness (SFCT) was 292.7 ± 100.1 μm. The mean number of pachydrusen per eye was 2.22 ± 1.73. The majority of eyes with pachydrusen had no other drusen or drusenoid deposits (95.2%). Only 16 eyes (4.8%) had other deposits, including soft drusen (10 eyes, 3.0%), cuticular drusen (3 eyes, 0.9%), and reticular pseudodrusen (RPD; 3 eyes, 0.9%). Macular pigmentary abnormalities accompanied pachydrusen in 68 eyes (27.4%). None of the eyes had GA, and 82 eyes (24.8%) had MNV. The majority of MNV was polypoidal choroidal vasculopathy (PCV; 65 eyes, 19.7%), followed by type 1 (10 eyes, 3.0%), type 2 (5 eyes, 1.5%), and type 3 MNV (2 eyes, 0.6%). Conclusions Eyes with pachydrusen in Korean population have several characteristic AMD lesions in low frequencies. These findings indicate that pachydrusen might have diagnostic and prognostic values that are different from those of other drusen or drusenoid deposits.