Segmentation of anterior segment boundaries in swept source OCT images

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.

Cascade spatial and channel-wise multifusion network with criss cross augmentation for corneal segmentation and reconstruction

High-quality 3D corneal reconstruction from AS-OCT images has demonstrated significant potential in computer-aided diagnosis, enabling comprehensive observation of corneal thickness, precise assessment of morphological characteristics, as well as location and quantification of keratitis-affected regions. However, it faces two main challenges: (1) prevalent medical image segmentation networks often struggle to accurately process low-contrast corneal regions, which is a vital pre-processing step for 3D corneal reconstruction, and (2) there are no reconstruction methods that can be directly applied to AS-OCT sequences with 180-degree scanning. To combat these, we propose CSCM-CCA-Net, a simple yet efficient network for accurate corneal segmentation. This network incorporates two key techniques: cascade spatial and channel-wise multifusion (CSCM), which captures intricate contextual interdependencies and effectively extracts low-contrast and obscure corneal features; and criss cross augmentation (CCA), which enhances shape-preserved feature representation to improve segmentation accuracy. Based on the obtained corneal segmentation results, we reconstruct the 3D volume data and generate a topographic map of corneal thickness through corneal image alignment. Additionally, we design a transfer function based on the analysis of intensity histogram and gradient histogram to explore more internal cues for better visualization results. Experimental results on CORNEA benchmark demonstrate the impressive performance of our proposed method in terms of both corneal segmentation and 3D reconstruction. Furthermore, we compare CSCM-CCA-Net with state-of-the-art medical image segmentation approaches using three challenging medical fundus segmentation datasets (DRIVE, CHASEDB1, FIVES), highlighting its superiority in terms of segmentation accuracy. The code and models will be made available at https://github.com/qianguiping/CSCM-CCA-Net.

Central and peripheral scleral lens-induced corneal oedema

PURPOSE

To quantify the magnitude of central and peripheral scleral lens-induced corneal oedema for a range of fluid reservoir thicknesses, and to compare these experimental results with theoretical models of corneal oedema both with and without limbal metabolic support (i.e., the lateral transport of metabolites and the influence of the limbal vasculature).

METHODS

Ten young healthy participants wore scleral lenses (KATT™, Capricornia Contact Lenses) fitted with low (mean 141 μm), medium (482 μm) and high (718 μm) central fluid reservoir thickness values across three separate study visits. The scleral lens thickness, fluid reservoir thickness and stromal corneal oedema were measured using optical coherence tomography. Oedema was quantified across the central (0-2.5 mm from the corneal apex) and peripheral (1.25-3 mm from the scleral spur) cornea. Experimental data were compared with published theoretical models of central to peripheral corneal oedema.

RESULTS

Stromal oedema varied with fluid reservoir thickness (p < 0.001) for both central and peripheral regions. The mean (standard deviation) stromal oedema was greater for the medium (2.08 (1.21)%) and high (2.22 (1.31)%) fluid reservoir thickness conditions compared to the low condition (1.00 (1.01)%) (p ≤ 0.01). Stromal oedema gradually increased from the corneal centre to the periphery by ~0.3% on average (relative increase of 18%), but the change did not reach statistical significance. This trend of increasing, rather than decreasing, oedema towards the limbus is consistent with theoretical modelling of peripheral oedema without metabolic support from the limbus.

CONCLUSIONS

The central and peripheral cornea displayed a similar magnitude of oedema, with increasing levels observed for medium and high fluid reservoir thicknesses. The gradual increase in oedema towards the limbus is consistent with a 'without limbal metabolic support' theoretical model.

SeqCorr-EUNet: A sequence correction dual-flow network for segmentation and quantification of anterior segment OCT image

The accurate segmentation of AS-OCT images is a prerequisite for the morphological details analysis of anterior segment structure and the extraction of clinical biological parameters, which play an essential role in the diagnosis, evaluation, and preoperative prognosis management of many ophthalmic diseases. Manually marking the boundaries of the anterior segment tissue is time-consuming and error-prone, with inherent speckle noise, various artifacts, and some low-quality scanned images further increasing the difficulty of the segmentation task. In this work, we propose a novel model called SeqCorr-EUNet with a dual-flow architecture based on convolutional gated recursive sequence correction for semantic segmentation and quantification of AS-OCT images. An EfficientNet encoder is employed to enhance the intra-slice features extraction ability of semantic segmentation flow. The sequence correction flow based on ConvGRU is introduced to extract inter-slice features from consecutive adjacent slices. Spatio-temporal information is fused to correct the morphological details of pre-segmentation results. And the channel attention gate is inserted into the skip-connection between encoder and decoder to enrich the contextual information and suppress the noise of irrelevant regions. Based on the segmentation results of the anterior segment structures, we achieved automatic extraction of essential clinical parameters, 3D reconstruction of the anterior chamber structure, and measurement of anterior chamber volume. The proposed SeqCorr-EUNet has been evaluated on the public AS-OCT dataset. The experimental results show that our method is competitive compared with the existing methods and significantly improves the segmentation and quantification performance of low-quality imaging structures in AS-OCT images.