To develop an objective glaucoma damage severity classification system based on OCT-derived retinal nerve fiber layer (RNFL) thickness measurements.
Algorithm development for RNFL damage severity ...classification based on multicenter OCT data.
A total of 6561 circumpapillary RNFL profiles from 2269 eyes of 1171 subjects to develop models, and 2505 RNFL profiles from 1099 eyes of 900 subjects to validate models.
We developed an unsupervised k-means model to identify clusters of eyes with similar RNFL thickness profiles. We annotated the clusters based on their respective global RNFL thickness. We computed the optimal global RNFL thickness thresholds that discriminated different severity levels based on Bayes’ minimum error principle. We validated the proposed pipeline based on an independent validation dataset with 2505 RNFL profiles from 1099 eyes of 900 subjects.
Accuracy, area under the receiver operating characteristic curve, and confusion matrix.
The k-means clustering discovered 4 clusters with 1382, 1613, 1727, and 1839 samples with mean (standard deviation) global RNFL thickness of 58.3 (8.9) μm, 78.9 (6.7) μm, 87.7 (8.2) μm, and 101.5 (7.9) μm. The Bayes’ minimum error classifier identified optimal global RNFL values of > 95 μm, 86 to 95 μm, 70 to 85 μm, and < 70 μm for discriminating normal eyes and eyes at the early, moderate, and advanced stages of RNFL thickness loss, respectively. About 4% of normal eyes and 98% of eyes with advanced RNFL loss had either global, or ≥ 1 quadrant, RNFL thickness outside of normal limits provided by the OCT instrument.
Unsupervised machine learning discovered that the optimal RNFL thresholds for separating normal eyes and eyes with early, moderate, and advanced RNFL loss were 95 μm, 85 μm, and 70 μm, respectively. This RNFL loss classification system is unbiased as there was no preassumption or human expert intervention in the development process. Additionally, it is objective, easy to use, and consistent, which may augment glaucoma research and day-to-day clinical practice.
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Myopia has been discussed as a risk factor for glaucoma. In this study, we characterized the relationship between ametropia and patterns of visual field (VF) loss in glaucoma. Reliable automated VFs ...(SITA Standard 24-2) of 120,019 eyes from 70,495 patients were selected from five academic institutions. The pattern deviation (PD) at each VF location was modeled by linear regression with ametropia (defined as spherical equivalent (SE) starting from extreme high myopia), mean deviation (MD), and their interaction (SE × MD) as regressors. Myopia was associated with decreased PD at the paracentral and temporal VF locations, whereas hyperopia was associated with decreased PD at the Bjerrum and nasal step locations. The severity of VF loss modulated the effect of ametropia: with decreasing MD and SE, paracentral/nasal step regions became more depressed and Bjerrum/temporal regions less depressed. Increasing degree of myopia was positively correlated with VF depression at four central points, and the correlation became stronger with increasing VF loss severity. With worsening VF loss, myopes have increased VF depressions at the paracentral and nasal step regions, while hyperopes have increased depressions at the Bjerrum and temporal locations. Clinicians should be aware of these effects of ametropia when interpreting VF loss.
The main factor causing irreversible blindness is glaucoma. Early detection greatly reduces the risk of further vision loss. To address this problem, we developed a domain adaptation-based deep ...learning model called Glaucoma Domain Adaptation (GDA) based on 66,742 fundus photographs collected from 3272 eyes of 1636 subjects. GDA learns domain-invariant and domain-specific representations to extract both general and specific features. We also developed a progressive weighting mechanism to accurately transfer the source domain knowledge while mitigating the transfer of negative knowledge from the source to the target domain. We employed low-rank coding for aligning the source and target distributions. We trained GDA based on three different scenarios including eyes annotated as glaucoma due to 1) optic disc abnormalities regardless of visual field abnormalities, 2) optic disc or visual field abnormalities except ones that are glaucoma due to both optic disc and visual field abnormalities at the same time, and 3) visual field abnormalities regardless of optic disc abnormalities. We then evaluate the generalizability of GDA based on two independent datasets. The AUCs of GDA in forecasting glaucoma based on the first, second, and third scenarios were 0.90, 0.88, and 0.80 and the Accuracies were 0.82, 0.78, and 0.72, respectively. The AUCs of GDA in diagnosing glaucoma based on the first, second, and third scenarios were 0.98, 0.96, and 0.93 and the Accuracies were 0.93, 0.91, and 0.88, respectively. The proposed GDA model achieved high performance and generalizability for forecasting and diagnosis of glaucoma disease from fundus photographs. GDA may augment glaucoma research and clinical practice in identifying patients with glaucoma and forecasting those who may develop glaucoma thus preventing future vision loss.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Abstract
Motivation
To develop and assess the accuracy of deep learning models that identify different retinal cell types, as well as different retinal ganglion cell (RGC) subtypes, based on patterns ...of single-cell RNA sequencing (scRNA-seq) in multiple datasets.
Results
Deep domain adaptation models were developed and tested using three different datasets. The first dataset included 44 808 single retinal cells from mice (39 cell types) with 24 658 genes, the second dataset included 6225 single RGCs from mice (41 subtypes) with 13 616 genes and the third dataset included 35 699 single RGCs from mice (45 subtypes) with 18 222 genes. We used four loss functions in the learning process to align the source and target distributions, reduce misclassification errors and maximize robustness. Models were evaluated based on classification accuracy and confusion matrix. The accuracy of the model for correctly classifying 39 different retinal cell types in the first dataset was ∼92%. Accuracy in the second and third datasets reached ∼97% and 97% in correctly classifying 40 and 45 different RGCs subtypes, respectively. Across a range of seven different batches in the first dataset, the accuracy of the lead model ranged from 74% to nearly 100%. The lead model provided high accuracy in identifying retinal cell types and RGC subtypes based on scRNA-seq data. The performance was reasonable based on data from different batches as well. The validated model could be readily applied to scRNA-seq data to identify different retinal cell types and subtypes.
Availability and implementation
The code and datasets are available on https://github.com/DM2LL/Detecting-Retinal-Cell-Classes-and-Ganglion-Cell-Subtypes. We have also added the class labels of all samples to the datasets.
Supplementary information
Supplementary data are available at Bioinformatics online.
To test the hypothesis that visual field (VF) progression can be predicted from baseline and longitudinal optical coherence tomography (OCT) structural measurements.
Prospective cohort study.
A total ...of 104 eyes (104 patients) with ≥3 years of follow-up and ≥5 VF examinations were enrolled. We defined VF progression based on pointwise linear regression on 24-2 VF (≥3 locations with slope less than or equal to −1.0 dB/year and P < .01). We used elastic net logistic regression (ENR) and machine learning to predict VF progression with demographics, baseline circumpapillary retinal nerve fiber layer (RNFL), macular ganglion cell/inner plexiform layer (GCIPL) thickness, and RNFL and GCIPL change rates at central 24 superpixels and 3 eccentricities, 3.4°, 5.5°, and 6.8°, from fovea and hemimaculas. Areas-under-ROC curves (AUC) were used to compare models.
Average ± SD follow-up and VF examinations were 4.5 ± 0.9 years and 8.7 ± 1.6, respectively. VF progression was detected in 23 eyes (22%). ENR selected rates of change of superotemporal RNFL sector and GCIPL change rates in 5 central superpixels and at 3.4° and 5.6° eccentricities as the best predictor subset (AUC = 0.79 ± 0.12). Best machine learning predictors consisted of baseline superior hemimacular GCIPL thickness and GCIPL change rates at 3.4° eccentricity and 3 central superpixels (AUC = 0.81 ± 0.10). Models using GCIPL-only structural variables performed better than RNFL-only models.
VF progression can be predicted with clinically relevant accuracy from baseline and longitudinal structural data. Further refinement of proposed models would assist clinicians with timely prediction of functional glaucoma progression and clinical decision making.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Recent advances in artificial intelligence (AI), robotics, and chatbots have brought these technologies to the forefront of medicine, particularly ophthalmology. These technologies have been applied ...in diagnosis, prognosis, surgical operations, and patient-specific care in ophthalmology. It is thus both timely and pertinent to assess the existing landscape, recent advances, and trajectory of trends of AI, AI-enabled robots, and chatbots in ophthalmology.
Some recent developments have integrated AI enabled robotics with diagnosis, and surgical procedures in ophthalmology. More recently, large language models (LLMs) like ChatGPT have shown promise in augmenting research capabilities and diagnosing ophthalmic diseases. These developments may portend a new era of doctor-patient-machine collaboration.
Ophthalmology is undergoing a revolutionary change in research, clinical practice, and surgical interventions. Ophthalmic AI-enabled robotics and chatbot technologies based on LLMs are converging to create a new era of digital ophthalmology. Collectively, these developments portend a future in which conventional ophthalmic knowledge will be seamlessly integrated with AI to improve the patient experience and enhance therapeutic outcomes.
Purpose: To identify ocular hypertension (OHT) subtypes with different trends of visual field (VF) progression based on unsupervised machine learning and to discover factors associated with fast VF ...progression. Design: Cross-sectional and longitudinal study. Participants: A total of 3133 eyes of 1568 ocular hypertension treatment study (OHTS) participants with at least five follow-up VF tests were included in the study. Methods: We used a latent class mixed model (LCMM) to identify OHT subtypes using standard automated perimetry (SAP) mean deviation (MD) trajectories. We characterized the subtypes based on demographic, clinical, ocular, and VF factors at the baseline. We then identified factors driving fast VF progression using generalized estimating equation (GEE) and justified findings qualitatively and quantitatively. Main Outcome Measure: Rates of SAP mean deviation (MD) change. Results: The LCMM model discovered four clusters (subtypes) of eyes with different trajectories of MD worsening. The number of eyes in clusters were 794 (25%), 1675 (54%), 531 (17%) and 133 (4%). We labeled the clusters as Improvers (cluster 1), Stables (cluster 2), Slow progressors (cluster 3), and Fast progressors (cluster 4) based on their mean of MD decline rate, which were 0.08, −0.06, −0.21, and -0.45 dB/year, respectively. Eyes with fast VF progression had higher baseline age, intraocular pressure (IOP), pattern standard deviation (PSD) and refractive error (RE), but lower central corneal thickness (CCT). Fast progression was associated with being male, heart disease history, diabetes history, African American race, and stroke history. Conclusion: Unsupervised clustering can objectively identify OHT subtypes including those with fast VF worsening without human expert intervention. Fast VF progression was associated with higher history of stroke, heart disease and diabetes. Fast progressors were more from African American race , males and had higher incidence of glaucoma conversion. Subtyping can provide guidance for adjusting treatment plans to slow vision loss and improve quality of life of patients with a faster progression course.
To investigate the capability of ChatGPT for forecasting the conversion from ocular hypertension (OHT) to glaucoma based on the Ocular Hypertension Treatment Study (OHTS).
Retrospective case-control ...study.
A total of 3008 eyes of 1504 subjects from the OHTS were included in the study.
We selected demographic, clinical, ocular, optic nerve head, and visual field (VF) parameters 1 year before glaucoma development from the OHTS participants. Subsequently, we developed queries by converting tabular parameters into textual format based on both eyes of all participants. We used the ChatGPT application program interface (API) to automatically perform ChatGPT prompting for all subjects. We then investigated whether ChatGPT can accurately forecast conversion from OHT to glaucoma based on various objective metrics.
Accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and weighted F1 score.
ChatGPT4.0 demonstrated an accuracy of 75%, AUC of 0.67, sensitivity of 56%, specificity of 78%, and weighted F1 score of 0.77 in predicting conversion to glaucoma 1 year before onset. ChatGPT3.5 provided an accuracy of 61%, AUC of 0.62, sensitivity of 64%, specificity of 59%, and weighted F1 score of 0.63 in predicting conversion to glaucoma 1 year before onset.
The performance of ChatGPT4.0 in forecasting development of glaucoma 1 year before onset was reasonable. The overall performance of ChatGPT4.0 was consistently higher than ChatGPT3.5. Large language models (LLMs) hold great promise for augmenting glaucoma research capabilities and enhancing clinical care. Future efforts in creating ophthalmology-specific LLMs that leverage multimodal data in combination with active learning may lead to more useful integration with clinical practice and deserve further investigations.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
To identify patterns of visual field (VF) loss based on unsupervised machine learning and to identify patterns that are associated with rapid progression.
Cross-sectional and longitudinal study.
A ...total of 2231 abnormal VFs from 205 eyes of 176 Ocular Hypertension Treatment Study (OHTS) participants followed over approximately 16 years.
Visual fields were assessed by an unsupervised deep archetypal analysis algorithm and an OHTS-certified VF reader to identify prevalent patterns of VF loss. Machine-identified patterns of glaucoma damage were compared against those patterns previously identified (expert-identified) in the OHTS in 2003. Based on the longitudinal VFs of each eye, VF loss patterns that were strongly associated with rapid glaucoma progression were identified.
Machine-expert correspondence and type of patterns of VF loss associated with rapid progression.
The average VF mean deviation (MD) at conversion to glaucoma was -2.7 decibels (dB) (standard deviation SD = 2.4 dB), whereas the average MD of the eyes at the last visit was -5.2 dB (SD = 5.5 dB). Fifty out of 205 eyes had MD rate of -1 dB/year or worse and were considered rapid progressors. Eighteen machine-identified patterns of VF loss were compared with expert-identified patterns, in which 13 patterns of VF loss were similar. The most prevalent expert-identified patterns included partial arcuate, paracentral, and nasal step defects, and the most prevalent machine-identified patterns included temporal wedge, partial arcuate, nasal step, and paracentral VF defects. One of the machine-identified patterns of VF loss predicted future rapid VF progression after adjustment for age, sex, and initial MD.
An automated machine learning system can identify patterns of VF loss and could provide objective and reproducible nomenclature for characterizing early signs of visual defects and rapid progression in patients with glaucoma.