Automated fine structure image analysis method for discrimination of diabetic retinopathy stage using conjunctival microvasculature images

A novel algorithm for cardiovascular screening using conjunctival microcirculatory parameters and blood biomarkers

Microvascular haemodynamic alterations are associated with coronary artery disease (CAD). The conjunctival microcirculation can easily be assessed non-invasively. However, the microcirculation of the conjunctiva has not been previously explored in clinical algorithms aimed at identifying patients with CAD. This case–control study involved 66 patients with post-myocardial infarction and 66 gender-matched healthy controls. Haemodynamic properties of the conjunctival microcirculation were assessed with a validated iPhone and slit lamp-based imaging tool. Haemodynamic properties were extracted with semi-automated software and compared between groups. Biomarkers implicated in the development of CAD were assessed in combination with conjunctival microcirculatory parameters. The conjunctival blood vessel parameters and biomarkers were used to derive an algorithm to aid in the screening of patients for CAD. Conjunctival blood velocity measured in combination with the blood biomarkers (N-terminal pro-brain natriuretic peptide and adiponectin) had an area under receiver operator characteristic curve (AUROC) of 0.967, sensitivity 93.0%, specificity 91.5% for CAD. This study demonstrated that the novel algorithm which included a combination of conjunctival blood vessel haemodynamic properties, and blood-based biomarkers could be used as a potential screening tool for CAD and should be validated for potential utility in asymptomatic individuals.

Ultrafast Doppler imaging and ultrasound localization microscopy reveal the complexity of vascular rearrangement in chronic spinal lesion

Acute spinal cord injury (SCI) leads to severe damage to the microvascular network. The process of spontaneous repair is accompanied by formation of new blood vessels; their functionality, however, presumably very important for functional recovery, has never been clearly established, as most studies so far used fixed tissues. Here, combining ultrafast Doppler imaging and ultrasound localization microscopy (ULM) on the same animals, we proceeded at a detailed analysis of structural and functional vascular alterations associated with the establishment of chronic SCI, both at macroscopic and microscopic scales. Using a standardized animal model of SCI, our results demonstrate striking hemodynamic alterations in several subparts of the spinal cord: a reduced blood velocity in the lesion site, and an asymmetrical hypoperfusion caudal but not rostral to the lesion. In addition, the worsening of many evaluated parameters at later time points suggests that the neoformed vascular network is not yet fully operational, and reveals ULM as an efficient in vivo readout for spinal cord vascular alterations. Finally, we show statistical correlations between the diverse biomarkers of vascular dysfunction and SCI severity. The imaging modality developed here will allow evaluating recovery of vascular function over time in pre-clinical models of SCI. Also, used on SCI patients in combination with other quantitative markers of neural tissue damage, it may help classifying lesion severity and predict possible treatment outcomes in patients.

Identifying diabetes from conjunctival images using a novel hierarchical multi-task network

Diabetes can cause microvessel impairment. However, these conjunctival pathological changes are not easily recognized, limiting their potential as independent diagnostic indicators. Therefore, we designed a deep learning model to explore the relationship between conjunctival features and diabetes, and to advance automated identification of diabetes through conjunctival images. Images were collected from patients with type 2 diabetes and healthy volunteers. A hierarchical multi-tasking network model (HMT-Net) was developed using conjunctival images, and the model was systematically evaluated and compared with other algorithms. The sensitivity, specificity, and accuracy of the HMT-Net model to identify diabetes were 78.70%, 69.08%, and 75.15%, respectively. The performance of the HMT-Net model was significantly better than that of ophthalmologists. The model allowed sensitive and rapid discrimination by assessment of conjunctival images and can be potentially useful for identifying diabetes.

Classification of advanced and early stages of diabetic retinopathy from non-diabetic subjects by an ordinary least squares modeling method applied to OCTA images

As the prevalence of diabetic retinopathy (DR) continues to rise, there is a need to develop computer-aided screening methods. The current study reports and validates an ordinary least squares (OLS) method to model optical coherence tomography angiography (OCTA) images and derive OLS parameters for classifying proliferative DR (PDR) and no/mild non-proliferative DR (NPDR) from non-diabetic subjects. OLS parameters were correlated with vessel metrics quantified from OCTA images and were used to determine predicted probabilities of PDR, no/mild NPDR, and non-diabetics. The classification rates of PDR and no/mild NPDR from non-diabetic subjects were 94% and 91%, respectively. The method had excellent predictive ability and was validated. With further development, the method may have potential clinical utility and contribute to image-based computer-aided screening and classification of stages of DR and other ocular and systemic diseases.

Rubeosis faciei diabeticorum is not associated with oxidative stress and skin autofluorescence

Background

Rubeosis faciei diabeticorum is a persistent facial erythema in patients with diabetes mellitus. The actual pathogenesis has not been studied. However, it is speculated to be a cutaneous diabetic microangiopathy.

Objective

Examine the correlation between the severity of facial erythema and the possible causes of microvascular diabetic complications, namely oxidative stress, hyperglycemia, and cutaneous accumulation of advanced glycation end-products .

Methods

Patients diagnosed with Type 2 diabetes mellitus (n = 32) were enrolled in the study. The facial erythema index was measured using the Mexameter MX18; cutaneous accumulation of advanced glycation end-products was estimated by measuring skin auto fluorescence with the AGE Reader (DiagnOptics Technologies B.V. – Groningen, Netherlands). Glycated haemoglobin, total antioxidant status, and malondialdehyde were measured in blood by TBARS assay. The correlation between the selected variables was assessed by Spearman's rank test; p ≤ 0.05 was considered statistically significant.

Results

There was a statistically significant correlation between total antioxidant status and the facial erythema index (ρ = 0.398, p = 0.024). Malondialdehyde, skin autofluorescence, glycated haemoglobin, body mass index, duration of diabetes, and age did not demonstrate statistically significant correlation with the facial erythema index.

Study limitations

This is an observational study. Elevation of total antioxidant status could have been caused by several factors that might have also influenced the development of rubeosis faciei, including hyperbilirubinemia and hyperuricemia.

Conclusions

The results contradicted expectations. Total antioxidant status correlated positively with facial erythema index; however, there was no correlation with oxidative stress and skin autofluorescence. Further investigations should be conducted to reveal the cause of total antioxidant status elevation in patients with rubeosis faciei.

Detection of Subclinical Diabetic Retinopathy by Fine Structure Analysis of Retinal Images

Background and Objective

Diabetic retinopathy (DR) is a major complication of diabetes and the leading cause of blindness among US working-age adults. Detection of subclinical DR is important for disease monitoring and prevention of damage to the retina before occurrence of vision loss. The purpose of this retrospective study is to describe an automated method for discrimination of subclinical DR using fine structure analysis of retinal images.

Methods

Discrimination between nondiabetic control (NC; N = 16) and diabetic without clinical retinopathy (NDR; N = 17) subjects was performed using ordinary least squares regression and Fisher's linear discriminant analysis. A human observer also performed the discrimination by visual inspection of the images.

Results

The discrimination rate for subclinical DR was 88% using the automated method and higher than the rate obtained by a human observer which was 45%.

Conclusions

The method provides sensitive and rapid analysis of retinal images and could be useful in detecting subclinical DR.

Inter-visit variability of conjunctival microvascular hemodynamic measurements in healthy and diabetic retinopathy subjects

Conjunctival microcirculation imaging provides a non-invasive means for detecting hemodynamic alterations due to systemic and ocular diseases. However, reliable longitudinal monitoring of hemodynamic changes due to disease progression requires establishment of measurement variability over time. The purpose of the current study was to determine inter-visit variability of conjunctival microvascular hemodynamic measurements in non-diabetic control (NC, N = 7) and diabetic retinopathy (DR, N = 10) subjects. Conjunctival microvascular imaging was performed during 2 visits, which were 17 ± 12 weeks apart. Images were analyzed to determine vessel diameter (D), axial blood velocity (V), blood flow (Q), wall shear rate (WSR) and wall shear stress (WSS). The inter-visit variability was determined based on mean inter-visit differences. In NC, inter-visit variability of D, V, Q, WSR and WSS were 0.2 ± 0.5 µm, −0.01 ± 0.16 mm/s, −8 ± 46 pl/s, −3 ± 46 s−1 and −0.01 ± 0.10 dyne/cm², respectively. Inter-visit variability of D, V, Q, WSR and WSS were beyond the normal 95% confidence limits in 60%, 20%, 40%, 20% and 20% of DR subjects, respectively. The variability of hemodynamic measurements over time was established in non-diabetic subjects, suggestive of the potential of the method for detecting longitudinal changes due to progression of DR.

A clinical correlation of conjunctival microangiopathy with grades of retinopathy in type 2 diabetes mellitus

BACKGROUND

Conjunctival microangiopathy has been described among diabetics similar to retinal vessel angiopathy. Correlation of these conjunctival microangiopathy changes with retinopathy may form the basis of screening by external examination without expert fundus evaluation.

METHODS

Conjunctival vessels widths and tortuous segment length of conjunctival vessels of 96 patients with type 2 diabetes mellitus were photographed and measured by the Zeiss Fundus camera Visupac software. The measurements were correlated with retinopathy grade in those eyes.

RESULTS

The mean conjunctival vessel width was 40.61 μ (SD 17.25) with a uniform increase from 34.4 μ (SD 8.70) in mild NPDR to 53.50 μ (SD 33.45) in the PDR group which was statistically significant (p < 0.01). The tortuous conjunctival vessel segment length increased from 711.51 μ (SD 83.90) in the mild NPDR group to 921.94 μ (SD 129.26) in those with PDR (p < 0.01). Vessel width greater than 80 μ was seen only in PDR and tortuosity values greater than 900 μ were seen in severe grades (severe NPDR and PDR). Both conjunctival vessel width and tortuosity showed a positive statistical correlation with increasing severity of retinopathy (r = 0.386, r² = 0.149 and r = 0.645, r² = 0.415).

CONCLUSION

A positive correlation was seen between conjunctival vessel width and tortuosity with severity of retinopathy. Widths over 80 μ and tortuous segment length over 900 μ are suggestive of severe grades of retinopathy.

Assessment of Conjunctival Microvascular Hemodynamics in Stages of Diabetic Microvasculopathy

Diabetes impairs the microcirculation and function of various vital tissues throughout the body. The conjunctival microcirculation can be non-invasively imaged and thus enables assessment of microvascular hemodynamics. In this study, alterations in conjunctival microvascular hemodynamics were quantitatively assessed at stages of increasing diabetic microvasculopathy based on diabetic retinopathy (DR). Subjects were categorized into non-diabetic control (C, N = 34), no clinically visible DR (NDR, N = 47), non-proliferative DR (NPDR, N = 45), and proliferative DR (PDR, N = 35). Conjunctival hemodynamic descriptors, namely vessel diameter (D), blood velocity (V), blood flow (Q), wall shear rate (WSR), and wall shear stress (WSS) were measured in arterioles and venules, and compared between DR and C subjects using generalized linear mixed models. In arterioles, V, WSR, and WSS were lower in NDR (P ≤ 0.01). V was lower in NDR than NPDR and PDR subjects (P ≤ 0.02). In venules, D was higher in NDR and NPDR (P ≤ 0.03), while V was lower in PDR (P = 0.04). Venular V and Q were higher in NPDR than PDR subjects (P ≤ 0.04). WSR and WSS were lower in all stages of DR (P ≤ 0.05), suggestive of the potential of WSS as a marker of diabetic microvasculopathy. Quantitative assessment of conjunctival hemodynamics can potentially be useful for evaluation of diabetic microvasculopathy.

Distribution of Different Sized Ocular Surface Vessels in Diabetics and Normal Individuals

Purpose:

To compare the distribution of different sized vessels using digital photographs of the ocular surface of diabetic and normal individuals.

Methods:

In this cross-sectional study, red-free conjunctival photographs of diabetic and normal individuals, aged 30-60 years, were taken under defined conditions and analyzed using a Radon transform-based algorithm for vascular segmentation. The image areas occupied by vessels (AOV) of different diameters were calculated. The main outcome measure was the distribution curve of mean AOV of different sized vessels. Secondary outcome measures included total AOV and standard deviation (SD) of AOV of different sized vessels.

Results:

Two hundred and sixty-eight diabetic patients and 297 normal (control) individuals were included, differing in age (45.50 ± 5.19 vs. 40.38 ± 6.19 years, P < 0.001), systolic (126.37 ± 20.25 vs. 119.21 ± 15.81 mmHg, P < 0.001) and diastolic (78.14 ± 14.21 vs. 67.54 ± 11.46 mmHg, P < 0.001) blood pressures. The distribution curves of mean AOV differed between patients and controls (smaller AOV for larger vessels in patients; P < 0.001) as well as between patients without retinopathy and those with non-proliferative diabetic retinopathy (NPDR); with larger AOV for smaller vessels in NPDR (P < 0.001). Controlling for the effect of confounders, patients had a smaller total AOV, larger total SD of AOV, and a more skewed distribution curve of vessels compared to controls.

Conclusion:

Presence of diabetes mellitus is associated with contraction of larger vessels in the conjunctiva. Smaller vessels dilate with diabetic retinopathy. These findings may be useful in the photographic screening of diabetes mellitus and retinopathy.