Artificial intelligence in screening, diagnosis, and classification of diabetic macular edema: A systematic review

A multimodal approach using fundus images and text meta-data in a machine learning classifier with embeddings to predict years with self-reported diabetes - An exploratory analysis

AIMS

Machine learning models can use image and text data to predict the number of years since diabetes diagnosis; such model can be applied to new patients to predict, approximately, how long the new patient may have lived with diabetes unknowingly. We aimed to develop a model to predict self-reported diabetes duration.

METHODS

We used the Brazilian Multilabel Ophthalmological Dataset. Unit of analysis was the fundus image and its meta-data, regardless of the patient. We included people 40 + years and fundus images without diabetic retinopathy. Fundus images and meta-data (sex, age, comorbidities and taking insulin) were passed to the MedCLIP model to extract the embedding representation. The embedding representation was passed to an Extra Tree Classifier to predict: 0-4, 5-9, 10-14 and 15 + years with self-reported diabetes.

RESULTS

There were 988 images from 563 people (mean age = 67 years; 64 % were women). Overall, the F1 score was 57 %. The group 15 + years of self-reported diabetes had the highest precision (64 %) and F1 score (63 %), while the highest recall (69 %) was observed in the group 0-4 years. The proportion of correctly classified observations was 55 % for the group 0-4 years, 51 % for 5-9 years, 58 % for 10-14 years, and 64 % for 15 + years with self-reported diabetes.

CONCLUSIONS

The machine learning model had acceptable accuracy and F1 score, and correctly classified more than half of the patients according to diabetes duration. Using large foundational models to extract image and text embeddings seems a feasible and efficient approach to predict years living with self-reported diabetes.

Recognition of diabetic retinopathy and macular edema using deep learning

Diabetic retinopathy (DR) and diabetic macular edema (DME) are both serious eye conditions associated with diabetes and if left untreated, and they can lead to permanent blindness. Traditional methods for screening these conditions rely on manual image analysis by experts, which can be time-consuming and costly due to the scarcity of such experts. To overcome the aforementioned challenges, we present the Modified CornerNet approach with DenseNet-100. This system aims to localize and classify lesions associated with DR and DME. To train our model, we first generate annotations for input samples. These annotations likely include information about the location and type of lesions within the retinal images. DenseNet-100 is a deep CNN used for feature extraction, and CornerNet is a one-stage object detection model. CornerNet is known for its ability to accurately localize small objects, which makes it suitable for detecting lesions in retinal images. We assessed our technique on two challenging datasets, EyePACS and IDRiD. These datasets contain a diverse range of retinal images, which is important to estimate the performance of our model. Further, the proposed model is also tested in the cross-corpus scenario on two challenging datasets named APTOS-2019 and Diaretdb1 to assess the generalizability of our system. According to the accomplished analysis, our method outperformed the latest approaches in terms of both qualitative and quantitative results. The ability to effectively localize small abnormalities and handle over-fitted challenges is highlighted as a key strength of the suggested framework which can assist the practitioners in the timely recognition of such eye ailments.

Machine learning and optical coherence tomography-derived radiomics analysis to predict persistent diabetic macular edema in patients undergoing anti-VEGF intravitreal therapy

BACKGROUND

Diabetic macular edema (DME) is a leading cause of vision loss in patients with diabetes. This study aimed to develop and evaluate an OCT-omics prediction model for assessing anti-vascular endothelial growth factor (VEGF) treatment response in patients with DME.

METHODS

A retrospective analysis of 113 eyes from 82 patients with DME was conducted. Comprehensive feature engineering was applied to clinical and optical coherence tomography (OCT) data. Logistic regression, support vector machine (SVM), and backpropagation neural network (BPNN) classifiers were trained using a training set of 79 eyes, and evaluated on a test set of 34 eyes. Clinical implications of the OCT-omics prediction model were assessed by decision curve analysis. Performance metrics (sensitivity, specificity, F1 score, and AUC) were calculated.

RESULTS

The logistic, SVM, and BPNN classifiers demonstrated robust discriminative abilities in both the training and test sets. In the training set, the logistic classifier achieved a sensitivity of 0.904, specificity of 0.741, F1 score of 0.887, and AUC of 0.910. The SVM classifier showed a sensitivity of 0.923, specificity of 0.667, F1 score of 0.881, and AUC of 0.897. The BPNN classifier exhibited a sensitivity of 0.962, specificity of 0.926, F1 score of 0.962, and AUC of 0.982. Similar discriminative capabilities were maintained in the test set. The OCT-omics scores were significantly higher in the non-persistent DME group than in the persistent DME group (p < 0.001). OCT-omics scores were also positively correlated with the rate of decline in central subfield thickness after treatment (Pearson's R = 0.44, p < 0.001).

CONCLUSION

The developed OCT-omics model accurately assesses anti-VEGF treatment response in DME patients. The model's robust performance and clinical implications highlight its utility as a non-invasive tool for personalized treatment prediction and retinal pathology assessment.

OCTess: AN OPTICAL CHARACTER RECOGNITION ALGORITHM FOR AUTOMATED DATA EXTRACTION OF SPECTRAL DOMAIN OPTICAL COHERENCE TOMOGRAPHY REPORTS

PURPOSE

Manual extraction of spectral domain optical coherence tomography (SD-OCT) reports is time and resource intensive. This study aimed to develop an optical character recognition (OCR) algorithm for automated data extraction from Cirrus SD-OCT macular cube reports.

METHODS

SD-OCT monocular macular cube reports (n = 675) were randomly selected from a single-center database of patients from 2020 to 2023. Image processing and bounding box operations were performed, and Tesseract (an OCR library) was used to develop the algorithm, OCTess. The algorithm was validated using a separate test data set.

RESULTS

The long short-term memory deep learning version of Tesseract achieved the best performance. After reverifying all discrepancies between human and algorithmic data extractions, OCTess achieved accuracies of 100.00% and 99.98% in the training (n = 125) and testing (n = 550) datasets, while the human error rate was 1.11% (98.89% accuracy) and 0.49% (99.51% accuracy) in each, respectively. OCTess extracted data in 3.1 seconds, compared with 94.3 seconds per report for human evaluators.

CONCLUSION

We developed an OCR and machine learning algorithm that extracted SD-OCT data with near-perfect accuracy, outperforming humans in both accuracy and efficiency. This algorithm can be used for efficient construction of large-scale SD-OCT data sets for researchers and clinicians.

Recommendations for diabetic macular edema management by retina specialists and large language model-based artificial intelligence platforms

PURPOSE

To study the role of artificial intelligence (AI) in developing diabetic macular edema (DME) management recommendations by creating and comparing responses to clinicians in hypothetical AI-generated case scenarios. The study also examined whether its joint recommendations followed national DME management guidelines.

METHODS

The AI hypothetically generated 50 ocular case scenarios from 25 patients using keywords like age, gender, type, duration and control of diabetes, visual acuity, lens status, retinopathy stage, coexisting ocular and systemic co-morbidities, and DME-related retinal imaging findings. For DME and ocular co-morbidity management, we calculated inter-rater agreements (kappa analysis) separately for clinician responses, AI-platforms, and the "majority clinician response" (the maximum number of identical clinician responses) and "majority AI-platform" (the maximum number of identical AI responses). Treatment recommendations for various situations were compared to the Indian national guidelines.

RESULTS

For DME management, clinicians (ĸ=0.6), AI platforms (ĸ=0.58), and the 'majority clinician response' and 'majority AI response' (ĸ=0.69) had moderate to substantial inter-rate agreement. The study showed fair to substantial agreement for ocular co-morbidity management between clinicians (ĸ=0.8), AI platforms (ĸ=0.36), and the 'majority clinician response' and 'majority AI response' (ĸ=0.49). Many of the current study's recommendations and national clinical guidelines agreed and disagreed. When treating center-involving DME with very good visual acuity, lattice degeneration, renal disease, anaemia, and a recent history of cardiovascular disease, there were clear disagreements.

CONCLUSION

For the first time, this study recommends DME management using large language model-based generative AI. The study's findings could guide in revising the global DME management guidelines.

Novel Artificial Intelligence-Based Assessment of Imaging Biomarkers in Full-Thickness Macular Holes: Preliminary Data from a Pivotal Trial

Artificial intelligence (AI)- and deep learning (DL)-based systems have shown significant progress in the field of macular disorders, demonstrating high performance in detecting retinal fluid and assessing anatomical changes during disease progression. This study aimed to validate an AI algorithm for identifying and quantifying prognostic factors in visual recovery after macular hole (MH) surgery by analyzing major optical coherence tomography (OCT) biomarkers. This study included 20 patients who underwent vitrectomy for a full-thickness macular hole (FTMH). The mean diameter of the FTMH was measured at 285.36 ± 97.4 μm. The preoperative best-corrected visual acuity (BCVA) was 0.76 ± 0.06 logMAR, improving to 0.38 ± 0.16 postoperatively, with a statistically significant difference (p = 0.001). AI software was utilized to assess biomarkers, such as intraretinal fluid (IRF) and subretinal fluid (SRF) volume, external limiting membrane (ELM) and ellipsoid zone (EZ) integrity, and retinal hyperreflective foci (HRF). The AI analysis showed a significant decrease in IRF volume, from 0.08 ± 0.12 mm3 preoperatively to 0.01 ± 0.01 mm3 postoperatively. ELM interruption improved from 79% ± 18% to 34% ± 37% after surgery (p = 0.006), whereas EZ interruption improved from 80% ± 22% to 40% ± 36% (p = 0.007) postoperatively. Additionally, the study revealed a negative correlation between preoperative IRF volume and postoperative BCVA recovery, suggesting that greater preoperative fluid volumes may hinder visual improvement. The integrity of the ELM and EZ was found to be essential for postoperative visual acuity improvement, with their disruption negatively impacting visual recovery. The study highlights the potential of AI in quantifying OCT biomarkers for managing MHs and improving patient care.

Optical coherence tomography in the management of diabetic macular oedema

Diabetic macular oedema (DMO) is the major cause of visual impairment in people with diabetes. Optical coherence tomography (OCT) is now the most widely used modality to assess presence and severity of DMO. DMO is currently broadly classified based on the involvement to the central 1 mm of the macula into non-centre or centre involved DMO (CI-DMO) and DMO can occur with or without visual acuity (VA) loss. This classification forms the basis of management strategies of DMO. Despite years of research on quantitative and qualitative DMO related features assessed by OCT, these do not fully inform physicians of the prognosis and severity of DMO relative to visual function. Having said that, recent research on novel OCT biomarkers development and re-defined classification of DMO show better correlation with visual function and treatment response. This review summarises the current evidence of the association of OCT biomarkers in DMO management and its potential clinical importance in predicting VA and anatomical treatment response. The review also discusses some future directions in this field, such as the use of artificial intelligence to quantify and monitor OCT biomarkers and retinal fluid and identify phenotypes of DMO, and the need for standardisation and classification of OCT biomarkers to use in future clinical trials and clinical practice settings as prognostic markers and secondary treatment outcome measures in the management of DMO.

Machine learning regression algorithms to predict short-term efficacy after anti-VEGF treatment in diabetic macular edema based on real-world data

The objective of this retrospective study was to predict short-term efficacy of anti-vascular endothelial growth factor (VEGF) treatment in diabetic macular edema (DME) using machine learning regression models. Real-world data from 279 DME patients who received anti-VEGF treatment at Ineye Hospital of Chengdu University of TCM between April 2017 and November 2022 were analyzed. Eight machine learning regression models were established to predict four clinical efficacy indicators. The accuracy of the models was evaluated using mean absolute error (MAE), mean square error (MSE) and coefficient of determination score (R2). Multilayer perceptron had the highest R2 and lowest MAE among all models. Regression tree and lasso regression had similar R2, with lasso having lower MAE and MSE. Ridge regression, linear regression, support vector machines and polynomial regression had lower R2 and higher MAE. Support vector machine had the lowest MSE, while polynomial regression had the highest MSE. Stochastic gradient descent had the lowest R2 and high MAE and MSE. The results indicate that machine learning regression algorithms are valuable and effective in predicting short-term efficacy in DME patients through anti-VEGF treatment, and the lasso regression is the most effective ML algorithm for developing predictive regression models.

Patients' Perception of Robot-Driven Technology in the Management of Retinal Diseases

INTRODUCTION

There is increasing application of robots and other artificial intelligence-driven technologies in the management of retinal disease. These technologies have the potential to meet increasing demands for retinal diseases. However, there is currently a lack of understanding of patients' attitudes towards use of robots in ophthalmology. This study investigates patients' attitudes towards robot-led management of retinal disease.

METHODS

Paper questionnaires were distributed to 177 patients attending intravitreal treatment (IVT) at the Princess Alexandra Eye Pavilion between 1 October 2022 and 31 January 2023. The questionnaire collected information on age, sex, diagnosis and postcode. In the questionnaire, patients responded to questions about their attitudes towards robot-led diagnosis, treatment decisions and IVT injections. Responses were collected using a 5-category Likert scale which was analysed using ordinal logistic regression with adjustments for age, sex and deprivation status.

RESULTS

Those from affluent socioeconomic backgrounds were significantly (p < 0.001) more accepting of robots diagnosing and deciding on treatment, although the total number of patients who were accepting was only 26 (14.7%). Furthermore, there was an increased proportion of patients who would accept robots if the robot made fewer mistakes than doctors, if the robot reduced waiting or appointment time and if the robot was able to communicate well and have empathy; the same association with socioeconomic background remains (p < 0.001). Lastly, 116 patients (65.5%) would not be happy if IVT injections were performed by a robot; this was more likely the case if the patient was female (p = 0.04) or from a more deprived socioeconomic background (p < 0.001).

CONCLUSION

Attitudes towards robot involvement in diagnosis and management of retinal disease are significantly associated with socioeconomic backgrounds and sex. Additional studies are required to further investigate these determinants of robot receptiveness to ensure acceptance and compliance with treatment with these new technologies.

Artificial intelligence in retinal image analysis: Development, advances, and challenges

Modern advances in diagnostic technologies offer the potential for unprecedented insight into ophthalmic conditions relating to the retina. We discuss the current landscape of artificial intelligence in retina with respect to screening, diagnosis, and monitoring of retinal pathologies such as diabetic retinopathy, diabetic macular edema, central serous chorioretinopathy, and age-related macular degeneration. We review the methods used in these models and evaluate their performance in both research and clinical contexts and discuss potential future directions for investigation, use of multiple imaging modalities in artificial intelligence algorithms, and challenges in the application of artificial intelligence in retinal pathologies.

Validation of an Automated Artificial Intelligence Algorithm for the Quantification of Major OCT Parameters in Diabetic Macular Edema

Artificial intelligence (AI) and deep learning (DL)-based systems have gained wide interest in macular disorders, including diabetic macular edema (DME). This paper aims to validate an AI algorithm for identifying and quantifying different major optical coherence tomography (OCT) biomarkers in DME eyes by comparing the algorithm to human expert manual examination. Intraretinal (IRF) and subretinal fluid (SRF) detection and volumes, external limiting-membrane (ELM) and ellipsoid zone (EZ) integrity, and hyperreflective retina foci (HRF) quantification were analyzed. Three-hundred three DME eyes were included. The mean central subfield thickness was 386.5 ± 130.2 µm. IRF was present in all eyes and confirmed by AI software. The agreement (kappa value) (95% confidence interval) for SRF presence and ELM and EZ interruption were 0.831 (0.738–0.924), 0.934 (0.886–0.982), and 0.936 (0.894–0.977), respectively. The accuracy of the automatic quantification of IRF, SRF, ELM, and EZ ranged between 94.7% and 95.7%, while accuracy of quality parameters ranged between 99.0% (OCT layer segmentation) and 100.0% (fovea centering). The Intraclass Correlation Coefficient between clinical and automated HRF count was excellent (0.97). This AI algorithm provides a reliable and reproducible assessment of the most relevant OCT biomarkers in DME. It may allow clinicians to routinely identify and quantify these parameters, offering an objective way of diagnosing and following DME eyes.