Corneal Endothelial Cell Morphometry and Corneal Thickness in Diabetic Contact Lens Wearers

Deep learning-based automated diagnosis of fungal keratitis with in vivo confocal microscopy images

Background

The aim of this study was to develop an intelligent system based on a deep learning algorithm for automatically diagnosing fungal keratitis (FK) in in vivo confocal microscopy (IVCM) images.

Methods

A total of 2,088 IVCM images were included in the training dataset. The positive group consisted of 688 images with fungal hyphae, and the negative group included 1,400 images without fungal hyphae. A total of 535 images in the testing dataset were not included in the training dataset. Deep Residual Learning for Image Recognition (ResNet) was used to build the intelligent system for diagnosing FK automatically. The system was verified by external validation in the testing dataset using the area under the receiver operating characteristic curve (AUC), accuracy, specificity and sensitivity.

Results

In the testing dataset, 515 images were diagnosed correctly and 20 were misdiagnosed (including 6 with fungal hyphae and 14 without). The system achieved an AUC of 0.9875 with an accuracy of 0.9626 in detecting fungal hyphae. The sensitivity of the system was 0.9186, with a specificity of 0.9834. When 349 diabetic patients were included in the training dataset, 501 images were diagnosed correctly and thirty-four were misdiagnosed (including 4 with fungal hyphae and 30 without). The AUC of the system was 0.9769. The accuracy, specificity and sensitivity were 0.9364, 0.9889 and 0.8256, respectively.

Conclusions

The intelligent system based on a deep learning algorithm exhibited satisfactory diagnostic performance and effectively classified FK in various IVCM images. The context of this deep learning automated diagnostic system can be extended to other types of keratitis.

Diabetes and contact lens wear

The literature suggests that diabetic patients may have altered tear chemistry and tear secretion as well as structural and functional changes to the corneal epithelium, endothelium and nerves. These factors, together with a reported increased incidence of corneal infection, suggest that diabetic patients may be particularly susceptible to developing ocular complications during contact lens wear. Reports of contact lens-induced complications in diabetic patients do exist, although a number of these reports concern patients with advanced diabetic eye disease using lenses on an extended wear basis. Over the past decade or so, there have been published studies documenting the response of the diabetic eye to more modern contact lens modalities. The results of these studies suggest that contact lenses can be a viable mode of refractive correction for diabetic patients. Furthermore, new research suggests that the measurement of tear glucose concentration could, in future, be used to monitor metabolic control non-invasively in diabetic patients. This could be carried out using contact lenses manufactured from hydrogel polymers embedded with glucose-sensing agents or nanoscale digital electronic technology. The purpose of this paper is to review the literature on the anterior ocular manifestations of diabetes, particularly that pertaining to contact lens wear.

Central corneal thickness and corneal endothelial cell changes caused by contact lens use in diabetic patients

PURPOSE

To analyze the effects of soft contact lenses on central corneal thickness and morphologic characteristics of the corneal endothelium in diabetic patients.

MATERIALS AND METHODS

Ultrasound pachymetry and noncontact specular microscopy were performed on 26 diabetic patients who regularly use soft contact lenses (group 1), 27 diabetic patients who do not use soft contact lenses (group 2) and 30 normal subjects (group 3). We compared the values in each group using the Mann-Whitney test.

RESULTS

The central cornea was found to be thicker in diabetic patients, both those who use and do not use contact lenses, than in the normal control group. The central corneal thickness was significantly higher in group 1 (564.73 ± 35.41 μm) and group 2 (555.76 ± 45.96 μm) than in the control group (534.05 ± 27.02 μm), but there was no statistically significant difference between groups 1 and 2. Endothelial cell density was significantly different between the groups, and was smallest in the group of diabetic patients using contact lenses. The coefficient of variation of cell size was significantly higher and the percentage of hexagonal cells was significantly lower in contact lens using diabetic patients than in non-contact lens using diabetic patients and in the control group.

CONCLUSION

Central corneal thickness and endothelial cell density is more affected by diabetes mellitus, and corneal endothelial cell morphology is more affected by contact lens use, when compared with normal subjects.

Profile of central corneal thickness in diabetics with and without dry eye in a Saudi population

PURPOSE

The aim of this study was to determine the impact of diabetes in patients with and without dry eyes (DE) on central corneal thickness (CCT) measurements.

METHODS

Eighty-six subjects (51 diabetics and 35 controls) participated in the study. Ultrasound pachymetry was used to measure the CCT, whereas tear break-up time (TBUT) and Schirmer tests were conducted for tear assessments. The participants were divided into group 1 (diabetics without DE), group 2 (diabetics with DE), and group 3 (nondiabetics controls). The measurements were compared using analysis of variance.

RESULTS

The mean (95% confidence interval [CI]) of CCT was 610 (599 to 620), 601 (582 to 618), and 583 (576 to 589) mum, respectively, in diabetics without DE, with DE, and the control groups. There was significant difference in mean CCT between diabetics without DE and controls (P < 0.0001), but no significant difference in diabetics with DE versus controls (P > 0.05). The mean (95% CI) values for TBUT and Schirmer tests were 14 (13 to 15), 6 (4 to 6), 14 (14 to 15) seconds, and 21 (19 to 23), 7 (6 to 8), and 20 (19 to 21) mm, respectively, for diabetics without DE, with DE, and the nondiabetic groups.

CONCLUSIONS

DE affects the CCT in diabetic subjects, having a predilection for lower values in those exhibiting DE. In addition, the results also support the view that diabetics tend to present with higher CCT values when compared with nondiabetic patients.

Clinical applications of corneal confocal microscopy

Corneal confocal microscopy is a novel clinical technique for the study of corneal cellular structure. It provides images which are comparable to in-vitro histochemical techniques delineating corneal epithelium, Bowman's layer, stroma, Descemet's membrane and the corneal endothelium. Because, corneal confocal microscopy is a non invasive technique for in vivo imaging of the living cornea it has huge clinical potential to investigate numerous corneal diseases. Thus far it has been used in the detection and management of pathologic and infectious conditions, corneal dystrophies and ecstasies, monitoring contact lens induced corneal changes and for pre and post surgical evaluation (PRK, LASIK and LASEK, flap evaluations and Radial Keratotomy), and penetrating keratoplasty. Most recently it has been used as a surrogate for peripheral nerve damage in a variety of peripheral neuropathies and may have potential in acting as a surrogate marker for endothelial abnormalities.

Differences in corneal thickness and corneal endothelium related to duration in diabetes

PURPOSE

This study evaluated the differences of corneal thickness and corneal endothelial morphology in diabetes compared with age-matched, healthy control subjects; in addition, we tested for correlation according to the duration of diabetes.

METHODS

Ultrasound pachymetry and noncontact specular microscopy were performed on 200 patients with diabetes and 100 control subjects. We compared the values for diabetics and normal persons with ANACOVA to adjust for age. Moreover, we examined the correlation between the subject parameters and the duration of diabetes by using a partial correlation coefficient that controlled for age.

RESULTS

The diabetic subjects had thicker corneas, less cell density and hexagonality, and more irregular cell size of the corneal endothelium than did the controls (P < 0.05). Central corneal thickness and the coefficient of variation for cell size were significantly higher for diabetes of over 10 years' duration than for diabetes of under 10 years' duration (P < 0.05). The endothelial cell density and percentage of hexagonal cells were lower for diabetes of over 10 years' duration than for diabetes of under 10 years' (P > 0.05). Central corneal thickness was correlated with duration of diabetes (P < 0.05), but corneal endothelial morphology was not (P < 0.05).

CONCLUSIONS

Those patients with diabetic duration of over 10 years have more corneal morphological abnormalities, especially the coefficient of variation in cell size, compared with the normal subjects. The central corneal thickness was significantly correlated with diabetic duration after controlling for age.

The relationship among race, iris color, central corneal thickness, and intraocular pressure

PURPOSE

Central corneal thickness (CCT) influences applanation intraocular pressure (IOP) measurement. The present study sought to determine whether iris color might represent a qualitative surrogate for CCT or race, and therefore differential risk for elevated IOP and, consequently, developing glaucoma.

METHODS

Eligible patients included those with best-corrected visual acuity (BCVA) better than 20/40 and who had not worn contact lenses within 24 hours. Exclusion criteria were prior ophthalmic surgery, topical ocular or systemic medication that would influence IOP, previous ocular inflammatory conditions, or current treatment for ophthalmic treatment. Data collection included demographic (name, date of birth, race), BCVA, and iris color. Iris color was judged according to a purpose-developed chart (white: blue, green, brown or black: brown) and patients were assigned to one of four groups. Goldmann applanation tonometry and pachymetry measurements were performed consecutively. To attain a power of 90% to find a difference of 40 microm with alpha < 0.05, we examined at least 14 subjects (28 eyes) for each group.

RESULTS

Comparing pachymetry measurements among iris colors revealed no statistically significant difference among the three groups of whites: blue (552 microm), green (552 microm), and brown (562 microm). The same held true when comparing IOP and CCT-adjusted IOP with iris color: blue-15.2, 15.1, green-15.4, 15.2, and brown-14.7, 14.0. When comparing CCT between whites and blacks, CCT was significantly thinner in blacks (533 microm), whether evaluating all whites (555 microm, p = 0.03) or comparing only the brown-iris white group with the black group (562 microm vs. 533 microm, p = 0.03). Mean CCT-adjusted IOP was barely significantly different between whites (14.8) and blacks (16.7) (p = 0.04).

CONCLUSION

These results suggest that iris color is not associated with CCT and apparently iris color does not influence measured IOP. We were able to establish a relationship between race and IOP when adjusting IOP for CCT. Our data show a significantly higher CCT-adjusted IOP for blacks than whites demonstrating a racial difference in CCT-corrected IOP.

Corneal Hydration Control in Contact Lens Wearers With Diabetes Mellitus

PURPOSE

The purpose of this study was to assess the acute swelling and recovery response to contact lens-induced corneal hypoxia in diabetic patients who wear contact lenses.

METHODS

A thick, low water content, soft contact lens was fitted under a light patch to one eye of 23 diabetic patients and one eye of 23 nondiabetic control subjects in a single-masked, controlled clinical study. After 2.5 hours an ultrasonic pachymeter was used to measure the induced corneal edema and the rate at which the cornea recovered to baseline thickness.

RESULTS

The induced corneal edema was significantly less for the diabetic patients compared with the nondiabetic control subjects (p = 0.004). The corneas of the diabetic patients were found to recover from the induced edema at a similar rate to the corneas of the nondiabetic control subjects (p = 0.28).

CONCLUSIONS

Diabetes mellitus alters aspects of corneal hydration control.