Comparison of semi-automated center-dot and fully automated endothelial cell analyses from specular microscopy images

Overestimation of corneal endothelial cell density by automated method in glaucomatous eyes with impaired corneal endothelial cells

PURPOSE

To determine between-method differences in corneal endothelial cell parameters using center and automated methods of non-contact specular microscopy (CellCheck software of Konan, Inc.) in glaucomatous eyes.

METHODS

We analyzed the central corneal endothelial cell density (ECD) of 245 glaucomatous eyes using center (ECD-Ce) and automated methods (ECD-Au). Based on the ECD-Ce, we allocated subjects to Groups 1 to 10 (at 250 cells/mm2 intervals) and evaluated the ECD, coefficient of variation in cell area (CV), and percentage of hexagonal cells (HEX).

RESULTS

There was a close correlation (r = 0.91) between the ECD values measured using both methods. However, ECD-Au were significantly higher than those measured by the center method when ECD-Ce was less than 2500 (in Groups 1 to 8; P < 0.001 to P = 0.006). The regression equation of (ECD-Au-ECD-Ce) = 1028-0.397*ECD-Ce shows greater deviation in eyes with lower ECD, and this difference became 0 when ECD -Ce was 2593 cells/mm2. None of the 44 subjects with an ECD-Ce of < 1000 cells/mm2 recorded an ECD-Au < 1000 cells/mm2. Compared with the center method, the automated method had higher and lower median CV and HEX values, respectively (P < 0.001). The between-method differences in both CV and HEX were negatively correlated with ECD-Ce (r = -0.49, P < 0.001 and r = -0.25, P < 0.001, respectively).

CONCLUSION

The automated method of the CellCheck software overestimates ECD in eyes with lower ECD values and may overlook risk of corneal decompensation.

Corneal endothelial cell density in patients receiving chemotherapy

PURPOSE

This study aimed to determine if the corneal endothelium was affected by chemotherapy.

METHODS

Chemotherapy patients were recruited to undergo specular microscopy before treatment and again at 1- and 2-year follow-up visits. One eye per patient, per follow-up, was selected for comparison to baseline.

RESULTS

Forty-six volunteers completed baseline and at least one follow-up assessment. From 51 eyes, there was no significant change in endothelial cell density for 41 eyes assessed at one year (M_D = 0.73%, 95% CI -1.33 to 2.78%) and 18 eyes at two years (M_D = 0.31%, 95% CI -3.53 to 4.15%).

CONCLUSION

Although other studies have shown that chemotherapy can adversely affect the corneal epithelium, this study showed no measurable change in endothelial cell density.

Unbiased corneal tissue analysis using Gabor-domain optical coherence microscopy and machine learning for automatic segmentation of corneal endothelial cells

SIGNIFICANCE

An accurate, automated, and unbiased cell counting procedure is needed for tissue selection for corneal transplantation.

AIM

To improve accuracy and reduce bias in endothelial cell density (ECD) quantification by combining Gabor-domain optical coherence microscopy (GDOCM) for three-dimensional, wide field-of-view (1  mm2) corneal imaging and machine learning for automatic delineation of endothelial cell boundaries.

APPROACH

Human corneas stored in viewing chambers were imaged over a wide field-of-view with GDOCM without contacting the specimens. Numerical methods were applied to compensate for the natural curvature of the cornea and produce an image of the flattened endothelium. A convolutional neural network (CNN) was trained to automatically delineate the cell boundaries using 180 manually annotated images from six corneas. Ten additional corneas were imaged with GDOCM and compared with specular microscopy (SM) to determine performance of the combined GDOCM and CNN to achieve automated endothelial counts relative to current procedural standards.

RESULTS

Cells could be imaged over a larger area with GDOCM than SM, and more cells could be delineated via automatic cell segmentation than via manual methods. ECD obtained from automatic cell segmentation of GDOCM images yielded a correlation of 0.94 (p  <  0.001) with the manual segmentation on the same images, and correlation of 0.91 (p  <  0.001) with the corresponding manually counted SM results.

CONCLUSIONS

Automated endothelial cell counting on GDOCM images with large field of view eliminates selection bias and reduces sampling error, which both affect the gold standard of manual counting on SM images.

Importance of standardizing the number of cells measured for coefficient of variation (COV) estimates of corneal endothelial cell area values as relevant to contact lens wear

PURPOSE

To assess the impact of using different numbers of cells in calculations of the coefficient of variation (COV) value for normal and polymegethous endothelia METHODS: Four sets of 20 non-contact specular microscope images obtained from Caucasian individuals were assessed, and categorized according to the extent of polymegethism, i.e. grade 0 (none), grade 1 (mild), grade 2 (moderate) and grade 3 (substantial). Cell areas were measured manually and then values for between 2 and 100 cells were progressively added and averaged to generate COV estimates. These were then assessed in terms of their relative values (as percentages +/- SD) in relation to the value obtained on over 100 cells.

RESULTS

For the 4 sets of endothelia with group-mean COV values of 25.8, 33.1, 45.1 and 56.8%, the reliability of the COV estimates realized asymptotic values of close to ±1.0, ±2.7, ±3.6 and ±5.0% with 90 cells, but with greater uncertainty with few numbers of cells, e.g. only to within ±3.4, ±5.3, ±6.1 and ±7.9% with 75 cells.

CONCLUSIONS

COV estimates for the corneal endothelium are dependent on the number of cells used in the calculations. It is recommended that every effort should be made to not only assess 75-100 cells per endothelial image, but that this number should be the same or very similar for all endothelial images in a particular data set so that the uncertainly (or estimated proportional error) in the estimates is balanced.

Improved Interchangeability with Different Corneal Specular Microscopes for Quantitative Endothelial Cell Analysis

Introduction

During our clinical practice and research, we encountered an interchangeability problem when using the SP-2000P and SP-3000P TopCon corneal specular microscopes (CSMs) (TopCon Medical Systems, Tokyo, Japan) regarding the endothelial cell count (ECC). We describe a method to improve interchangeability between these CSMs.

Methods

Five consecutive good-quality endothelial cell photographs were obtained in 22 eyes of 11 subjects. An ECC comparison between the two CSMs was performed after (I) gauging and calibration by the manufacturer, (II) adjustment of the magnification, (III) correction after external horizontal and vertical calibration.

Results

There was a statistically significant difference between the ECC of the SP-2000P and SP-3000P at the start. The SP-2000P counted an average of 500 cells/mm² more than the SP-3000P (p=0.00). After correction for magnification and determining a correction factor based on external calibration, the difference between the ECC of the SP-2000P and the SP-3000P was then found to be 0.4 cells/mm² and was not statistically significant (p=0.98).

Discussion

We propose a method for improving interchangeability, which involves checking magnification settings, re-checking magnification calibration with external calibration devices, and then calculating correction factors. This method can be applied to various specular or confocal microscopes and their associated endothelial cell analysis software packages to be able to keep performing precise endothelial cell counts and to enable comparison of ECCs when a CSM needs to be replaced or when results from different microscopes need to be compared.