Long-term Treatment Patterns for Diabetic Macular Edema: Up to 6-Year Follow-Up in the IRIS® Registry.

OBJECTIVE: To characterize anti-vascular endothelial growth factor (VEGF) intravitreal therapy (IVT) patterns and long-term visual outcomes among patients with diabetic macular edema (DME) in routine clinical practice in the United States. DESIGN: Retrospective analysis of the American Academy of Ophthalmology's Intelligent Research in Sight (IRIS®) Registry.; Participants: Treatment-naïve patients with DME (no previous IVT in the past 12 months) initiating anti-VEGF IVT from 1/1/2015-3/31/2021. METHODS: Baseline characteristics, treatment patterns, and long-term visual acuity (VA) outcomes were reported for up to 6 years of follow-up. MAIN OUTCOME MEASURES: Outcomes included the annualized number of injections, change in VA, and anti-VEGF agents. RESULTS: A total of 190,345 eyes met inclusion criteria. After 1 year of anti-VEGF IVT initiation, eyes received a mean of 3.9 (±2.8) injections and gained +3.2 (±16.4) letters of vision. Of the 1,236 eyes with year 6 data, eyes received a mean of 2.9 (±2.1) injections in year 6 and gained +0.5 (±19.7) letters from baseline. The number of injections decreased, and injection intervals increased year over year up to 6 years regardless of baseline VA initiation. The average injection interval was 10-weeks in year 1, then widened to 13.2 in year 2, before plateauing in years 3-6 (12.6, 12.3, 12.2, and 12.3 weeks respectively). Improvements in VA from baseline were greatest in eyes that received 5 or more injections each year. At the end of follow-up, eyes with good baseline vision (> 20/25) lost vision, whereas those with worse baseline vision (< 20/25) gained vision. Although 51.7% of patients with DME discontinued IVT after a mean of 6 months, 32.8% re-initiated anti-VEGF IVT. Worse VA outcomes were associated with patients of Hispanic ethnicity (-1.08 [-1.34, -0.83] compared to non-Hispanic), Medicaid insurance (-1.15 [-1.48, -0.81] compared to Commercial), and older age (-0.06 [-0.07, -0.05] each additional year) CONCLUSIONS: Patients with DME in the routine clinical settings receive fewer injections than those in clinical trials and fewer than recommended per the label of FDA approved anti-VEGF IVT.

EYE-Llama, an in-domain large language model for ophthalmology.

Background: Training Large Language Models (LLMs) with in-domain data can significantly enhance their performance, leading to more accurate and reliable question-answering (QA) systems essential for supporting clinical decision-making and educating patients. Methods: This study introduces LLMs trained on in-domain, well-curated ophthalmic datasets. We also present an open-source substantial ophthalmic language dataset for model training. Our LLMs (EYE-Llama), first pre-trained on an ophthalmology-specific dataset, including paper abstracts, textbooks, EyeWiki, and Wikipedia articles. Subsequently, the models underwent fine-tuning using a diverse range of QA datasets. The LLMs at each stage were then compared to baseline Llama 2, ChatDoctor, and ChatGPT (GPT3.5) models, using four distinct test sets, and evaluated quantitatively (Accuracy, F1 score, and BERTScore) and qualitatively by two ophthalmologists. Results: Upon evaluating the models using the American Academy of Ophthalmology (AAO) test set and BERTScore as the metric, our models surpassed both Llama 2 and ChatDoctor in terms of F1 score and performed equally to ChatGPT, which was trained with 175 billion parameters (EYE-Llama: 0.57, Llama 2: 0.56, ChatDoctor: 0.56, and ChatGPT: 0.57). When evaluated on the MedMCQA test set, the fine-tuned models demonstrated a higher accuracy compared to the Llama 2 and ChatDoctor models (EYE-Llama: 0.39, Llama 2: 0.33, ChatDoctor: 0.29). However, ChatGPT outperformed EYE-Llama with an accuracy of 0.55. When tested with the PubmedQA set, the fine-tuned model showed improvement in accuracy over both the Llama 2, ChatGPT, and ChatDoctor models (EYE-Llama: 0.96, Llama 2: 0.90, ChatGPT: 0.93, ChatDoctor: 0.92). Conclusion: The study shows that pre-training and fine-tuning LLMs like EYE-Llama enhances their performance in specific medical domains. Our EYE-Llama models surpass baseline Llama 2 in all evaluations, highlighting the effectiveness of specialized LLMs in medical QA systems. (Funded by NEI R15EY035804 (MNA) and UNC Charlotte Faculty Research Grant (MNA).).

Quantitative Characterization of Retinal Features in Translated OCTA.

Purpose: This study explores the feasibility of using generative machine learning (ML) to translate Optical Coherence Tomography (OCT) images into Optical Coherence Tomography Angiography (OCTA) images, potentially bypassing the need for specialized OCTA hardware. Methods: The method involved implementing a generative adversarial network framework that includes a 2D vascular segmentation model and a 2D OCTA image translation model. The study utilizes a public dataset of 500 patients, divided into subsets based on resolution and disease status, to validate the quality of TR-OCTA images. The validation employs several quality and quantitative metrics to compare the translated images with ground truth OCTAs (GT-OCTA). We then quantitatively characterize vascular features generated in TR-OCTAs with GT-OCTAs to assess the feasibility of using TR-OCTA for objective disease diagnosis. Result: TR-OCTAs showed high image quality in both 3 and 6 mm datasets (high-resolution, moderate structural similarity and contrast quality compared to GT-OCTAs). There were slight discrepancies in vascular metrics, especially in diseased patients. Blood vessel features like tortuosity and vessel perimeter index showed a better trend compared to density features which are affected by local vascular distortions. Conclusion: This study presents a promising solution to the limitations of OCTA adoption in clinical practice by using vascular features from TR-OCTA for disease detection. Translation relevance: This study has the potential to significantly enhance the diagnostic process for retinal diseases by making detailed vascular imaging more widely available and reducing dependency on costly OCTA equipment.

Retinal Detachment After Macular Hole Repair in Alport Syndrome.

Alport syndrome is characterized by type IV collagen network disruptions leading to renal, auditory, and ocular manifestations. This case report details a 24-year-old man with Alport syndrome who developed a rhegmatogenous retinal detachment following macular hole repair. The patient underwent a successful vitrectomy and internal limiting membrane peel for macular hole repair but returned with vision loss due to retinal detachment five weeks later, which necessitated a combined scleral buckle and vitrectomy. This is the first case describing rhegmatogenous retinal detachment post-macular hole repair in Alport syndrome. [Ophthalmic Surg Lasers Imaging Retina 2024;55:289-292.].

Impact of Anti-VEGF Treatment and Patient Characteristics on Vision Outcomes in Neovascular Age-related Macular Degeneration: Up to 6-Year Analysis of the AAO IRIS® Registry.

Purpose: To evaluate anti-VEGF treatment patterns and the influence of patient demographic and clinical characteristics on up to 6-year vision outcomes in neovascular age-related macular degeneration. Design: Retrospective, multicenter, noninterventional registry study with up to 6 years of follow-up. Participants: A cohort of 254 655 eyes (226 767 patients) with first anti-VEGF injection and at least 2 years of follow-up; 160 423 eyes had visual acuity (VA) data. Methods: Anonymized patient data were collected in the United States through the IRIS® Registry (Intelligent Research in Sight). Main Outcome Measures: Changes in VA from baseline; frequency of and gaps between intravitreal anti-VEGF injections; treatment discontinuations; switching anti-VEGF agents; and influence of baseline clinical and demographic characteristics on VA. Results: After a mean VA increase of 3.0 ETDRS letters at year 1, annual decreases led to a net loss from baseline of 4.6 letters after 6 years. Patients with longer follow-ups had better baseline and follow-up VA. From a mean of 7.2 in year 1 and 5.6 in year 2, mean injections plateaued between 4.2 to 4.6 in years 3 through 6. Treatment was discontinued in 38.8% of eyes and switched in 32.3%. When adjusting for differences at baseline, every additional injection resulted in a 0.68 letter improvement from baseline to year 1; thus, multiple injections in a year have the potential to be clinically meaningful. Older age, male gender, Medicaid insurance, and not being treated by a retina specialist were associated with a higher likelihood of vision loss at year 1. Of the patients, 58.5% lost ≥ 10 letters VA at least once during follow-up, with 14.5% of patients experiencing sustained poor vision after a median of 3.4 years. Conclusions: After modest mean VA improvement with intravitreal anti-VEGF injections at year 1, patients netted a loss of VA by year 6. Injection frequency decreased over time, and this was paired with a relatively high rate of discontinuation. Modeling suggested that more frequent injections were associated with better VA. Difficulty with continuous adherence to frequent intravitreal injections may have contributed to undertreatment resulting in less-than-optimal vision outcomes. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Feasibility of cross-vendor linkage of ophthalmic images with electronic health record data: an analysis from the IRIS Registry®.

Purpose: To link compliant, universal Digital Imaging and Communications in Medicine (DICOM) ophthalmic imaging data at the individual patient level with the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight). Design: A retrospective study using de-identified EHR registry data. Subjects Participants Controls: IRIS Registry records. Materials and Methods: DICOM files of several imaging modalities were acquired from two large retina ophthalmology practices. Metadata tags were extracted and harmonized to facilitate linkage to the IRIS Registry using a proprietary, heuristic patient-matching algorithm, adhering to HITRUST guidelines. Linked patients and images were assessed by image type and clinical diagnosis. Reasons for failed linkage were assessed by examining patients' records. Main Outcome Measures: Success rate of linking clinicoimaging and EHR data at the patient level. Results: A total of 2 287 839 DICOM files from 54 896 unique patients were available. Of these, 1 937 864 images from 46 196 unique patients were successfully linked to existing patients in the registry. After removing records with abnormal patient names and invalid birthdates, the success linkage rate was 93.3% for images. 88.2% of all patients at the participating practices were linked to at least one image. Conclusions and Relevance: Using identifiers from DICOM metadata, we created an automated pipeline to connect longitudinal real-world clinical data comprehensively and accurately to various imaging modalities from multiple manufacturers at the patient and visit levels. The process has produced an enriched and multimodal IRIS Registry, bridging the gap between basic research and clinical care by enabling future applications in artificial intelligence algorithmic development requiring large linked clinicoimaging datasets.

Acute Subretinal Fluid Accumulation Induced by Futibatinib Therapy for Malignant Metastatic Breast Cancer.

Futibatinib is an irreversible inhibitor of fibroblast growth factor receptors and is currently the subject of phase II clinical trials for the treatment of metastatic carcinomas. We report a case of a 59-year-old woman with metastatic malignant breast cancer who developed acute symptomatic subretinal fluid (SRF) accumulation after two weeks of futibatinib therapy. The SRF resolved within two weeks after futibatinib cessation. The medication was subsequently restarted at a lower dose, and SRF recurred within two weeks. To our knowledge, this is the first case depicting rapidly reversible SRF accumulation with the use of futibatinib in a real-world clinical setting. [Ophthalmic Surg Lasers Imaging Retina 2024;55:109-111.].

Artificial Intelligence Improves Patient Follow-Up in a Diabetic Retinopathy Screening Program.

Purpose: We examine the rate of and reasons for follow-up in an Artificial Intelligence (AI)-based workflow for diabetic retinopathy (DR) screening relative to two human-based workflows. Patients and Methods: A DR screening program initiated September 2019 between one institution and its affiliated primary care and endocrinology clinics screened 2243 adult patients with type 1 or 2 diabetes without a diagnosis of DR in the previous year in the San Francisco Bay Area. For patients who screened positive for more-than-mild-DR (MTMDR), rates of follow-up were calculated under a store-and-forward human-based DR workflow ("Human Workflow"), an AI-based workflow involving IDx-DR ("AI Workflow"), and a two-step hybrid workflow ("AI-Human Hybrid Workflow"). The AI Workflow provided results within 48 hours, whereas the other workflows took up to 7 days. Patients were surveyed by phone about follow-up decisions. Results: Under the AI Workflow, 279 patients screened positive for MTMDR. Of these, 69.2% followed up with an ophthalmologist within 90 days. Altogether 70.5% (N=48) of patients who followed up chose their location based on primary care referral. Among the subset of patients that were seen in person at the university eye institute under the Human Workflow and AI-Human Hybrid Workflow, 12.0% (N=14/117) and 11.7% (N=12/103) of patients with a referrable screening result followed up compared to 35.5% of patients under the AI Workflow (N=99/279; χ2df=2 = 36.70, p < 0.00000001). Conclusion: Ophthalmology follow-up after a positive DR screening result is approximately three-fold higher under the AI Workflow than either the Human Workflow or AI-Human Hybrid Workflow. Improved follow-up behavior may be due to the decreased time to screening result.

Vitreoretinal Specialist Use of Ancillary Testing: An IRIS® Registry Analysis.

Purpose: To investigate patterns of ancillary imaging testing among vitreoretinal specialists for patients with vitreoretinal disease in the United States (US). Methods: Optical coherence tomography (OCT), color fundus photography (CFP), and fluorescein angiography (FA), ordered by vitreoretinal specialists and documented within the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight) between 01 January 2018 and 31 December 2020, were retrospectively assessed. Trends in imaging modality choice were analyzed by payer type, geographic region, and practice type. Sub-analyses were conducted according to categorization of vitreoretinal specialists into those treating a high versus low volume of patients with neovascular age-related macular degeneration (nAMD). Results: OCT was the most common imaging modality used, followed by CFP and FA. Following normalization, the highest volume of OCT procedures performed were identified among Medicare Advantage and Medicare Fee-for-Service beneficiaries, within the South of the US, and at medium and large practices. Minimal differences were observed for CFP and FA volume across payer types and regions. Across practice types, the largest volume of CFP and FA procedures were identified in small and private equity owned practices, respectively. Vitreoretinal specialists with a high nAMD volume more frequently performed OCT than those with a low nAMD volume. Conclusion: Vitreoretinal specialists demonstrated a strong preference for OCT, with real-world associations according to payer type, geographic location, and practice type. Volume of nAMD patients seen impacted the likelihood of specialists ordering OCTs.

Federated learning for diagnosis of age-related macular degeneration.

This paper presents a federated learning (FL) approach to train deep learning models for classifying age-related macular degeneration (AMD) using optical coherence tomography image data. We employ the use of residual network and vision transformer encoders for the normal vs. AMD binary classification, integrating four unique domain adaptation techniques to address domain shift issues caused by heterogeneous data distribution in different institutions. Experimental results indicate that FL strategies can achieve competitive performance similar to centralized models even though each local model has access to a portion of the training data. Notably, the Adaptive Personalization FL strategy stood out in our FL evaluations, consistently delivering high performance across all tests due to its additional local model. Furthermore, the study provides valuable insights into the efficacy of simpler architectures in image classification tasks, particularly in scenarios where data privacy and decentralization are critical using both encoders. It suggests future exploration into deeper models and other FL strategies for a more nuanced understanding of these models' performance. Data and code are available at https://github.com/QIAIUNCC/FL_UNCC_QIAI.

Ophthalmology at the Forefront of Big Data Integration in Medicine: Insights from the IRIS Registry Database.

Ophthalmology stands at the vanguard of incorporating big data into medicine, as exemplified by the integration of The Intelligent Research in Sight (IRIS) Registry. This synergy cultivates patient-centered care, demonstrates real world efficacy and safety data for new therapies, and facilitates comprehensive population health insights. By evaluating the creation and utilization of the world's largest specialty clinical data registry, we underscore the transformative capacity of data-driven medical paradigms, current shortcomings, and future directions. We aim to provide a scaffold for other specialties to adopt big data integration into medicine.

AI-Human Hybrid Workflow Enhances Teleophthalmology for the Detection of Diabetic Retinopathy.

Objective: Detection of diabetic retinopathy (DR) outside of specialized eye care settings is an important means of access to vision-preserving health maintenance. Remote interpretation of fundus photographs acquired in a primary care or other nonophthalmic setting in a store-and-forward manner is a predominant paradigm of teleophthalmology screening programs. Artificial intelligence (AI)-based image interpretation offers an alternative means of DR detection. IDx-DR (Digital Diagnostics Inc) is a Food and Drug Administration-authorized autonomous testing device for DR. We evaluated the diagnostic performance of IDx-DR compared with human-based teleophthalmology over 2 and a half years. Additionally, we evaluated an AI-human hybrid workflow that combines AI-system evaluation with human expert-based assessment for referable cases. Design: Prospective cohort study and retrospective analysis. Participants: Diabetic patients ≥ 18 years old without a prior DR diagnosis or DR examination in the past year presenting for routine DR screening in a primary care clinic. Methods: Macula-centered and optic nerve-centered fundus photographs were evaluated by an AI algorithm followed by consensus-based overreading by retina specialists at the Stanford Ophthalmic Reading Center. Detection of more-than-mild diabetic retinopathy (MTMDR) was compared with in-person examination by a retina specialist. Main Outcome Measures: Sensitivity, specificity, accuracy, positive predictive value, and gradability achieved by the AI algorithm and retina specialists. Results: The AI algorithm had higher sensitivity (95.5% sensitivity; 95% confidence interval [CI], 86.7%-100%) but lower specificity (60.3% specificity; 95% CI, 47.7%-72.9%) for detection of MTMDR compared with remote image interpretation by retina specialists (69.5% sensitivity; 95% CI, 50.7%-88.3%; 96.9% specificity; 95% CI, 93.5%-100%). Gradability of encounters was also lower for the AI algorithm (62.5%) compared with retina specialists (93.1%). A 2-step AI-human hybrid workflow in which the AI algorithm initially rendered an assessment followed by overread by a retina specialist of MTMDR-positive encounters resulted in a sensitivity of 95.5% (95% CI, 86.7%-100%) and a specificity of 98.2% (95% CI, 94.6%-100%). Similarly, a 2-step overread by retina specialists of AI-ungradable encounters improved gradability from 63.5% to 95.6% of encounters. Conclusions: Implementation of an AI-human hybrid teleophthalmology workflow may both decrease reliance on human specialist effort and improve diagnostic accuracy. Financial Disclosures: Proprietary or commercial disclosure may be found after the references.

Progression of Geographic Atrophy: Retrospective Analysis of Patients from the IRIS® Registry (Intelligent Research in Sight).

Purpose: To evaluate disease progression and associated vision changes in patients with geographic atrophy (GA) secondary to age-related macular degeneration (AMD) in 1 eye and GA or neovascular AMD (nAMD) in the fellow eye using a large dataset from routine clinical practice. Design: Retrospective analysis of clinical data over 24 months. Subjects: A total of 256 635 patients with GA from the American Academy of Ophthalmology (Academy) IRIS® Registry (Intelligent Research in Sight) Registry (January 2016 to December 2017). Methods: Patients with ≥ 24 months of follow-up were grouped by fellow-eye status: Cohort 1, GA:GA; Cohort 2, GA:nAMD, each with (subfoveal) and without subfoveal (nonsubfoveal) involvement. Eyes with history of retinal disease other than AMD were excluded. Sensitivity analysis included patients who were managed by retina specialists and had a record of imaging within 30 days of diagnosis. Main Outcome Measures: Change in visual acuity (VA), occurrence of new-onset nAMD, and GA progression from nonsubfoveal to subfoveal. Results: In total, 69 441 patients were included: 44 120 (64%) GA:GA and 25 321 (36%) GA:nAMD. Otherwise eligible patients (57 788) were excluded due to follow-up < 24 months. In both GA:GA and GA:nAMD cohorts, nonsubfoveal study eyes had better mean (standard deviation) VA at index (67 [19.3] and 66 [20.3] letters) than subfoveal eyes (59 [23.9] and 47 [26.9] letters), and 24-month mean VA changes were similar for nonsubfoveal (-7.6 and -6.2) and subfoveal (-7.9 and -6.5) subgroups. Progression to subfoveal GA occurred in 16.7% of nonsubfoveal study eyes in the GA:GA cohort and 12.5% in the GA:nAMD cohort. More new-onset study-eye nAMD was observed in the GA:nAMD (21.6%) versus GA:GA (8.2%) cohorts. Sensitivity analysis supported the robustness of the observations in the study. Conclusions: This retrospective analysis describes the natural progression of GA lesions and the decline in VA associated with the disease. Financial Disclosures: Proprietary or commercial disclosure may be found after the references.

Contrastive learning-based pretraining improves representation and transferability of diabetic retinopathy classification models.

Diabetic retinopathy (DR) is a major cause of vision impairment in diabetic patients worldwide. Due to its prevalence, early clinical diagnosis is essential to improve treatment management of DR patients. Despite recent demonstration of successful machine learning (ML) models for automated DR detection, there is a significant clinical need for robust models that can be trained with smaller cohorts of dataset and still perform with high diagnostic accuracy in independent clinical datasets (i.e., high model generalizability). Towards this need, we have developed a self-supervised contrastive learning (CL) based pipeline for classification of referable vs non-referable DR. Self-supervised CL based pretraining allows enhanced data representation, therefore, the development of robust and generalized deep learning (DL) models, even with small, labeled datasets. We have integrated a neural style transfer (NST) augmentation in the CL pipeline to produce models with better representations and initializations for the detection of DR in color fundus images. We compare our CL pretrained model performance with two state of the art baseline models pretrained with Imagenet weights. We further investigate the model performance with reduced labeled training data (down to 10 percent) to test the robustness of the model when trained with small, labeled datasets. The model is trained and validated on the EyePACS dataset and tested independently on clinical datasets from the University of Illinois, Chicago (UIC). Compared to baseline models, our CL pretrained FundusNet model had higher area under the receiver operating characteristics (ROC) curve (AUC) (CI) values (0.91 (0.898 to 0.930) vs 0.80 (0.783 to 0.820) and 0.83 (0.801 to 0.853) on UIC data). At 10 percent labeled training data, the FundusNet AUC was 0.81 (0.78 to 0.84) vs 0.58 (0.56 to 0.64) and 0.63 (0.60 to 0.66) in baseline models, when tested on the UIC dataset. CL based pretraining with NST significantly improves DL classification performance, helps the model generalize well (transferable from EyePACS to UIC data), and allows training with small, annotated datasets, therefore reducing ground truth annotation burden of the clinicians.

Identifying the mechanism of missingness for unspecified diabetic retinopathy disease severity in the electronic health record: an IRIS® Registry analysis.

Observational studies of diabetic retinopathy (DR) using electronic health record data often determine disease severity using International Classification of Disease (ICD) codes. We investigated the mechanism of missingness for DR severity based on ICD coding using the American Academy of Ophthalmology IRIS® Registry. We included all patient encounters in the registry with a DR ICD-9 or ICD-10 code between January 1, 2014 and June 30, 2021. Demographic, clinical, and practice-level characteristics were compared between encounters with specified and unspecified disease severity. Practices were divided into quartiles based on the proportion of clinical encounters with unspecified DR severity. Encounters with unspecified disease severity were associated with significantly older patient age, better visual acuity, and lower utilization of ophthalmic procedures. Higher volume practices and retina specialist practices had lower proportions of clinical encounters with unspecified disease severity. Results strongly suggest that DR disease severity related to ICD coding is missing not at random.

Emulating VIEW 1 and VIEW 2 Clinical Trial Outcome Data Using the American Academy of Ophthalmology IRIS Registry.

BACKGROUND AND OBJECTIVE: A retrospective, noninterventional cohort study of the American Academy of Ophthalmology IRIS Registry, an electronic health record (EHR)-based comprehensive eye disease and condition registry, intended to assess whether the IRIS® Registry (Intelligent Research in Sight) could emulate the VIEW randomized clinical trials (VIEW RCTs) eligibility criteria, treatment protocol regimen, and primary endpoint. PATIENTS AND METHODS: Deidentified patients having an anti-VEGF injection of aflibercept or ranibizumab between January 1, 2013, and December 31, 2018, from the IRIS Registry. Patients were treated in accordance with one of three treatment regimens from the VIEW RCT: monthly intravitreal aflibercept injection (IAI 2Q4), intravitreal aflibercept every 2 months after 3 initial monthly doses (IAI 2Q8), or monthly ranibizumab (RQ4) injection. The main outcome measures are the number and proportion of patients meeting VIEW RCT eligibility and treatment group criteria, demographic, and clinical differences between IRIS Registry treatment groups, mean change in best documented visual acuity at one year, and evaluation of the primary endpoint of the VIEW RCT: difference in the proportion of patients maintaining vision. RESULTS: Among the 90,900 patients who met VIEW RCT eligibility criteria, 4,457 (4.85%) met treatment group criteria. The percentage of patients maintaining vision at one year was over 90%. No statistically significant difference was observed when comparing the proportion of patients maintaining vision among the RQ4 treatment group to the IAI 2Q4 or IAI 2Q8 treatment group. CONCLUSIONS: A small percentage of real-world patients met VIEW RCT study eligibility criteria and treatment protocol regimen. Among patients meeting all available criteria, the primary endpoint interpretation yielded by an observational EHR-based dataset suggested comparable results to the VIEW RCT. [Ophthalmic Surg Lasers Imaging Retina 2023;54:6-14.].

Dry Age-Related Macular Degeneration: Distribution of Visual Acuity and Progression Risk in a Large Registry.

INTRODUCTION: Understanding the progression to geographic atrophy (GA) in late dry age-related macular degeneration (dAMD) can support development opportunities for dAMD treatments. We characterized dAMD by distribution of visual acuity (VA) categories and evaluated VA progression risk by disease stage. METHODS: This retrospective observational study used data from the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight) to identify patients diagnosed with dAMD in ≥ 1 eye from January 2016 through December 2019 (index date) with ≥ 1 visit and ≥ 1 VA measurement recorded post-index date. Patients were followed until the date of last visit, last contribution for diagnosing provider, or diagnosis of neovascular AMD post-index. Models were utilized to describe the distribution of VA categories and progression to worsening VA. RESULTS: Data from 593,277 patients were analyzed. At baseline, 64.4% had mild disease, 29.4% intermediate, and 2.9%/3.3% had GA with/without subfoveal involvement. Most patients with mild (88.4%) and intermediate (79.7%) disease and GA without subfoveal involvement (57.1%) had baseline VA ≥ 20/63 in the study eye; 72.0% of patients with GA with subfoveal involvement had VA < 20/63. Modeled results showed lower VA with more progressive stage at baseline. Annual probability of stable dAMD based on baseline stage ranged from 82.1% (GA without) to 92.3% (GA with subfoveal involvement). Annual progression probability to GA without/with subfoveal involvement was 0.4% for mild and 5.5% for intermediate disease and from dry to neovascular AMD, 0.5% for mild and 8.0% for intermediate disease. CONCLUSIONS: Results from this analysis of a large database of electronic health records complement those from randomized trials and show that patients with more advanced dAMD have lower VA at baseline and that VA progression is generally faster with each progressive stage. Together these findings highlight the disease burden and trajectory of dAMD as well as opportunities for addressing unmet needs.

Dry age-related macular degeneration (dAMD) is a disease that progressively worsens over time. As the disease progresses, patients start to lose their vision, leading to a substantial burden on their quality of life and finances due to the need for increased healthcare services. As of 2022, there are no medications available to reverse or stop worsening of dAMD. This study used real-world data from a large registry of electronic health records to increase the understanding of how patients progress through the stages of dAMD. By reviewing patient records, we were able to identify approximately 600,000 patients with confirmed dAMD. These patients were then followed over time, and we were able to confirm that patients with a lower ability to see at the beginning of our review period had more advanced dAMD. We also found that as patients' disease worsened, their vision also decreased. These findings highlight the need for new medication options to reverse or delay the worsening of dAMD and improve the quality of life for patients.

Depth-resolved extraction of optical attenuation for glaucoma assessment in clinical settings: a pilot study.

We assessed the ability of the optical attenuation coefficient (AC) to detect early-stage glaucoma with two AC estimation algorithms: retinal layer intensity ratio (LIR) and depth-resolved confocal (DRC). We also introduced new depth-dependent AC parameters for retinal nerve fiber layer assessment. Optical coherence tomography B-scans were collected from 44 eyes of age-similar participants with eye health ranging from healthy to severe glaucoma, including glaucoma suspect patients. Mean AC values estimated from the DRC method are comparable to ratio-extracted values (p > 0.5 for all study groups), and the depth-dependent ACDRC parameters enhance the utility of the AC for detection of early-stage glaucoma.

Predicting Systemic Health Features from Retinal Fundus Images Using Transfer-Learning-Based Artificial Intelligence Models.

While color fundus photos are used in routine clinical practice to diagnose ophthalmic conditions, evidence suggests that ocular imaging contains valuable information regarding the systemic health features of patients. These features can be identified through computer vision techniques including deep learning (DL) artificial intelligence (AI) models. We aim to construct a DL model that can predict systemic features from fundus images and to determine the optimal method of model construction for this task. Data were collected from a cohort of patients undergoing diabetic retinopathy screening between March 2020 and March 2021. Two models were created for each of 12 systemic health features based on the DenseNet201 architecture: one utilizing transfer learning with images from ImageNet and another from 35,126 fundus images. Here, 1277 fundus images were used to train the AI models. Area under the receiver operating characteristics curve (AUROC) scores were used to compare the model performance. Models utilizing the ImageNet transfer learning data were superior to those using retinal images for transfer learning (mean AUROC 0.78 vs. 0.65, p-value < 0.001). Models using ImageNet pretraining were able to predict systemic features including ethnicity (AUROC 0.93), age > 70 (AUROC 0.90), gender (AUROC 0.85), ACE inhibitor (AUROC 0.82), and ARB medication use (AUROC 0.78). We conclude that fundus images contain valuable information about the systemic characteristics of a patient. To optimize DL model performance, we recommend that even domain specific models consider using transfer learning from more generalized image sets to improve accuracy.