An automatic approach for cell detection and segmentation of corneal endothelium in specular microscope.

PURPOSE: Specular microscopy is an indispensable tool for clinicians seeking to monitor the corneal endothelium. Automated methods of determining endothelial cell density (ECD) are limited in their ability to analyze images of poor quality. We describe and assess an image processing algorithm to analyze corneal endothelial images. METHODS: A set of corneal endothelial images acquired with a Konan CellChek specular microscope was analyzed using three methods: flex-center, Konan Auto Tracer, and the proposed method. In this technique, the algorithm determines the region of interest, filters the image to differentiate cell boundaries from their interiors, and utilizes stochastic watershed segmentation to draw cell boundaries and assess ECD based on the masked region. We compared ECD measured by the algorithm with manual and automated results from the specular microscope. RESULTS: We analyzed a total of 303 images manually, using the Auto Tracer, and with the proposed image processing method. Relative to manual analysis across all images, the mean error was 0.04% in the proposed method (p = 0.23 for difference) whereas Auto Tracer demonstrated a bias towards overestimation, with a mean error of 5.7% (p = 2.06× 10-8). The relative mean absolute errors were 6.9% and 7.9%, respectively, for the proposed and Auto Tracer. The average time for analysis of each image using the proposed method was 2.5 s. CONCLUSION: We demonstrate a computationally efficient algorithm to analyze corneal endothelial cell density that can be implemented on devices for clinical and research use.

Mobile-CellNet: Automatic Segmentation of Corneal Endothelium Using an Efficient Hybrid Deep Learning Model.

PURPOSE: The corneal endothelium, the innermost layer of the human cornea, exhibits a morphology of predominantly hexagonal cells. These endothelial cells are believed to have limited regeneration capacity, and their density decreases over time. Endothelial cell density (ECD) can therefore be used to measure the health of the corneal endothelium and the overall cornea. In clinical settings, specular microscopes are used to image this layer. Owing to the unavailability of reliable automatic tools, technicians often manually mark the cell centers and borders to measure ECD for such images, a process that is time and resource-consuming. METHODS: In this article, we propose Mobile-CellNet, a novel completely automatic, efficient deep learning-based cell segmentation algorithm to estimate ECD. This uses 2 similar image segmentation models working in parallel along with image postprocessing using classical image processing techniques. We also compare the proposed algorithm with widely used biomedical image segmentation networks U-Net and U-Net++. RESULTS: The proposed technique achieved a mean absolute error of 4.06% for the ECD on the test set, comparable with the error for U-Net of 3.80% ( P = 0.185 for difference), but requiring almost 31 times fewer floating-point operations (FLOPs) and 34 times fewer parameters. CONCLUSIONS: Mobile-CellNet accurately segments corneal endothelial cells and reports ECD and cell morphology efficiently. This can be used to develop tools to analyze specular corneal endothelial images in remote settings.

Alcohol Abuse Is Associated With Alterations in Corneal Endothelial Cell Morphology.

PURPOSE: Alcohol consumption is highly prevalent throughout the world. We sought to detect, in a large sample of cornea donors, whether alcohol abuse is associated with changes in corneal endothelial morphology after accounting for other comorbidities including tobacco use. METHODS: At a single eye bank, 10,322 eyes from a total of 5624 unique donors underwent imaging with a Konan CellChek D specular microscope. Demographic information and medical history were associated with each tissue. Images were analyzed using a standardized protocol for assessment of endothelial cell density, hexagonality, and variation. In this retrospective analysis, a multivariable regression was conducted to assess for an association between alcohol abuse and corneal endothelial metrics. Measurements were averaged across eyes for each donor. Bonferroni corrections were applied to account for multiple comparisons. RESULTS: Among 5624 donors, the mean (standard deviation) endothelial cell density was 2785 (383.0) cells/mm 2 . Indicators of alcohol abuse were present in 1382 donors (24.5%). In a multivariable regression model that included age, sex, tobacco use, history of cataract surgery, and diabetes mellitus, alcohol abuse was associated with a decrease of 60.9 cells/mm 2 [95% confidence interval (CI), -83.0 to -38.7 cells/mm 2 , P = 7.6 × 10 -8 ], an increase in the coefficient of variation by 0.0048 (95% CI, 0.17-0.79, P = 0.002), and a decrease in percent hexagonality by 0.93% (95% CI, -1.3 to -0.6, P = 4.5 × 10 -7 ). CONCLUSIONS: Alcohol abuse is associated with significant alterations to corneal endothelial density and morphology.

Comparison of corneal endothelial cell density and morphology with Optisol GS and Life4C storage media in the eye bank: a 5-year retrospective analysis.

Purpose: Optisol GS and Life4C are corneal storage media used by eye banks worldwide. We sought to determine if either solution was associated with superior corneal endothelial cell density (ECD) or morphology in a large cohort of donor corneas. Methods: From January 2016 through December 2020, 10,316 corneas from 5,624 unique donors were acquired and analyzed at Rocky Mountain Lions Eye Bank. In April 2019, Life4C replaced Optisol GS as the sole storage medium. We compared ECD and morphology before and after April 2019, and excluded corneas processed within the transition period. Univariable and multivariable regression analyses accounted for age, sex, tobacco use, heavy alcohol use, and diabetes. Only right corneas were analyzed to account for the correlation between eyes. Results: Of 5042 right corneas analyzed, 3486 were stored in Optisol GS and 1556 in Life4C. There was no significant difference in ECD across groups (2794 vs. 2793 cells/mm2 in Optisol GS and Life4C, p=0.88). In multivariate analyses, there was no significant difference in corneal ECD (0.6 cells/mm2 higher with Life4C, p=0.96) or hexagonality (0.22% higher with Life4C, p=0.31). However, the coefficient of variation was significantly lower in the Life4C group (-0.0039, p=0.03). After adjustment for above factors, corneas in Life4C demonstrated a 3.1% decreased likelihood of exhibiting CV values greater than 0.40 (p=0.009). Conclusions: This study demonstrates comparable and favorable outcomes using both storage media and confirms their overall efficacy. The decreased CV in Life4C is not of clinically significant magnitude, but merits further research in clinical and long-term settings.

Mobile-PolypNet: Lightweight Colon Polyp Segmentation Network for Low-Resource Settings.

Colon polyps, small clump of cells on the lining of the colon, can lead to colorectal cancer (CRC), one of the leading types of cancer globally. Hence, early detection of these polyps automatically is crucial in the prevention of CRC. The deep learning models proposed for the detection and segmentation of colorectal polyps are resource-consuming. This paper proposes a lightweight deep learning model for colorectal polyp segmentation that achieved state-of-the-art accuracy while significantly reducing the model size and complexity. The proposed deep learning autoencoder model employs a set of state-of-the-art architectural blocks and optimization objective functions to achieve the desired efficiency. The model is trained and tested on five publicly available colorectal polyp segmentation datasets (CVC-ClinicDB, CVC-ColonDB, EndoScene, Kvasir, and ETIS). We also performed ablation testing on the model to test various aspects of the autoencoder architecture. We performed the model evaluation by using most of the common image-segmentation metrics. The backbone model achieved a DICE score of 0.935 on the Kvasir dataset and 0.945 on the CVC-ClinicDB dataset, improving the accuracy by 4.12% and 5.12%, respectively, over the current state-of-the-art network, while using 88 times fewer parameters, 40 times less storage space, and being computationally 17 times more efficient. Our ablation study showed that the addition of ConvSkip in the autoencoder slightly improves the model's performance but it was not significant (p-value = 0.815).