Deep Learning-Based Model for Automatic Assessment of Anterior Angle Chamber in Ultrasound Biomicroscopy

A Deep Learning Model for Automatically Quantifying the Anterior Segment in Ultrasound Biomicroscopy Images of Implantable Collamer Lens Candidates

OBJECTIVE

This study aimed to develop and evaluate a deep learning-based model that could automatically measure anterior segment (AS) parameters on preoperative ultrasound biomicroscopy (UBM) images of implantable Collamer lens (ICL) surgery candidates.

METHODS

A total of 1164 panoramic UBM images were preoperatively obtained from 321 patients who received ICL surgery in the Eye Center of Renmin Hospital of Wuhan University (Wuhan, China) to develop an imaging database. First, the UNet++ network was utilized to segment AS tissues automatically, such as corneal lens and iris. In addition, image processing techniques and geometric localization algorithms were developed to automatically identify the anatomical landmarks (ALs) of pupil diameter (PD), anterior chamber depth (ACD), angle-to-angle distance (ATA), and sulcus-to-sulcus distance (STS). Based on the results of the latter two processes, PD, ACD, ATA, and STS can be measured. Meanwhile, an external dataset of 294 images from Huangshi Aier Eye Hospital was employed to further assess the model's performance in other center. Lastly, a subset of 100 random images from the external test set was chosen to compare the performance of the model with senior experts.

RESULTS

Whether in the internal test dataset or external test dataset, using manual labeling as the reference standard, the models achieved a mean Dice coefficient exceeding 0.880. Additionally, the intra-class correlation coefficients (ICCs) of ALs' coordinates were all greater than 0.947, and the percentage of Euclidean distance distribution of ALs within 250 μm was over 95.24%.While the ICCs for PD, ACD, ATA, and STS were greater than 0.957, furthermore, the average relative error (ARE) of PD, ACD, ATA, and STS were below 2.41%. In terms of human versus machine performance, the ICCs between the measurements performed by the model and those by senior experts were all greater than 0.931.

CONCLUSION

A deep learning-based model could measure AS parameters using UBM images of ICL candidates, and exhibited a performance similar to that of a senior ophthalmologist.

Artificial intelligence and big data integration in anterior segment imaging for glaucoma

The integration of artificial intelligence (AI) and big data in anterior segment (AS) imaging represents a transformative approach to glaucoma diagnosis and management. This article explores various AS imaging techniques, such as AS optical coherence tomography, ultrasound biomicroscopy, and goniophotography, highlighting their roles in identifying angle-closure diseases. The review focuses on advancements in AI, including machine learning and deep learning, which enhance image analysis and automate complex processes in glaucoma care, and provides current evidence on the performance and clinical applications of these technologies. In addition, the article discusses the integration of big data, detailing its potential to revolutionize medical imaging by enabling comprehensive data analysis, fostering enhanced clinical decision-making, and facilitating personalized treatment strategies. In this article, we address the challenges of standardizing and integrating diverse data sets and suggest that future collaborations and technological advancements could substantially improve the management and research of glaucoma. This synthesis of current evidence and new technologies emphasizes their clinical relevance, offering insights into their potential to change traditional approaches to glaucoma evaluation and care.

Correlation Between Anterior Chamber Angle Status and Limbal Stem Cell Deficiency in Primary Angle-Closure Glaucoma

PURPOSE

To investigate the correlation between the opening and closing states of anterior chamber angle (ACA) and the density of limbal epithelial basal cells (LEBCs) in subjects with primary angle-closure glaucoma (PACG).

DESIGN

Cross-sectional observational study.

METHODS

A total of 54 eyes of 29 patients diagnosed with PACG were included in the study. Fifty-four eyes from normal subjects were included as control. Automatic evaluation system for ultrasound biomicroscopy images of anterior chamber angle was used to assist ophthalmologists in identifying the opening or closing state of ACA, and the in vivo confocal microscopy (IVCM) was used to evaluate the density of LEBCs in different directions.

RESULTS

(1) The average density of LEBCs in the superior, inferior, nasal, and temporal limbus of the eyes in the PACG group was lower than that in the control group, and this pattern did not align with the density distribution observed in the control group. (2) In the early, moderate and advanced PACG, the density of LEBCs corresponding to the closed angle was lower than that in the control group (P < .05). Compared with the density of LEBCs corresponding to the closed angle and the open angle, the closed angle of PACG in the early, moderate and advanced stages was less than that in the open angle (P < .05 in the early and moderate stages; advanced stage P > .05). (3) The basal cell density was processed by dimensionless analysis. In the data calculated by averaging and minimizing, both closed angle dimensionless values were smaller than the open angle (P < .05). (4) Comparative analysis was conducted among the normal, open-angle, and closed-angle conditions in the superior, inferior, nasal, and temporal limbus. In the early stage of PACG, significant differences were observed in 4 limbal regions (P < .05), while in the moderate PACG stage, this difference was noted in 3 limbal regions (P < .05). In advanced PACG, 2 limbal regions exhibited significant differences (P < .05). These findings suggest that during the early PACG stage, angle closure is the predominant influencing factor on LEBCs density, while in the advanced stage, the decrease in density is attributed to a combination of angle closure and the natural progression of the disease.

CONCLUSIONS

There is a significant correlation between anterior chamber angle status and LEBCs. Advanced PACG and angle closure should be highly suspected of the occurrence of limbal stem cell deficiency (LSCD).

[Application of artificial intelligence in glaucoma. Part 1. Neural networks and deep learning in glaucoma screening and diagnosis]

This article reviews literature on the use of artificial intelligence (AI) for screening, diagnosis, monitoring and treatment of glaucoma. The first part of the review provides information how AI methods improve the effectiveness of glaucoma screening, presents the technologies using deep learning, including neural networks, for the analysis of big data obtained by methods of ocular imaging (fundus imaging, optical coherence tomography of the anterior and posterior eye segments, digital gonioscopy, ultrasound biomicroscopy, etc.), including a multimodal approach. The results found in the reviewed literature are contradictory, indicating that improvement of the AI models requires further research and a standardized approach. The use of neural networks for timely detection of glaucoma based on multimodal imaging will reduce the risk of blindness associated with glaucoma.