Fractal analysis of fluoroangiographic patterns in anterior ischaemic optic neuropathy and optic neuritis: a pilot study

Vessel Density Features of Optical Coherence Tomography Angiography for Classification of Optic Neuropathies Using Machine Learning

BACKGROUND

To evaluate the classification performance of machine learning based on the 4 vessel density features of peripapillary optical coherence tomography angiography (OCT-A) for classifying healthy, nonarteritic anterior ischemic optic neuropathy (NAION), and optic neuritis (ON) eyes.

METHODS

Forty-five eyes of 45 NAION patients, 32 eyes of 32 ON patients, and 76 eyes of 76 healthy individuals with optic nerve head OCT-A were included. Four vessel density features of OCT-A images were developed using a threshold-based segmentation method and were integrated in 3 models of machine learning classifiers. Classification performances of support vector machine (SVM), random forest, and Gaussian Naive Bayes (GNB) models were evaluated with the area under the receiver-operating-characteristic curve (AUC) and accuracy.

RESULTS

We divided 121 images into a 70% training set and 30% test set. For ON-NAION classification, best results were achieved with 50% threshold, in which 3 classifiers (SVM, RF, and GNB) discriminated ON from NAION with an AUC of 1 and accuracy of 1. For ON-Normal classification, with 100% threshold, SVM and RF classifiers were able to discriminate normal from ON with AUCs of 1 and accuracies of 1. For NAION-normal classification, with 50% threshold, the SVM and RF classified the NAION from normal with AUC and accuracy of 1.

CONCLUSIONS

ML based on the combined peripapillary vessel density features of total vessels and capillaries in the whole image and ring image could provide excellent performance for NAION and ON distinction.

A deep learning system for predicting time to progression of diabetic retinopathy

Diabetic retinopathy (DR) is the leading cause of preventable blindness worldwide. The risk of DR progression is highly variable among different individuals, making it difficult to predict risk and personalize screening intervals. We developed and validated a deep learning system (DeepDR Plus) to predict time to DR progression within 5 years solely from fundus images. First, we used 717,308 fundus images from 179,327 participants with diabetes to pretrain the system. Subsequently, we trained and validated the system with a multiethnic dataset comprising 118,868 images from 29,868 participants with diabetes. For predicting time to DR progression, the system achieved concordance indexes of 0.754-0.846 and integrated Brier scores of 0.153-0.241 for all times up to 5 years. Furthermore, we validated the system in real-world cohorts of participants with diabetes. The integration with clinical workflow could potentially extend the mean screening interval from 12 months to 31.97 months, and the percentage of participants recommended to be screened at 1-5 years was 30.62%, 20.00%, 19.63%, 11.85% and 17.89%, respectively, while delayed detection of progression to vision-threatening DR was 0.18%. Altogether, the DeepDR Plus system could predict individualized risk and time to DR progression over 5 years, potentially allowing personalized screening intervals.

The Leading Causes of Optic Disc Edema Seen in Tertiary Care, Academic Neuro-Ophthalmology Clinics

BACKGROUND

Optic disc edema (ODE) is one of the most common reasons for referral to a neuro-ophthalmologist. There are various causes that require vastly different workup. Thus, differentiating among each cause is important. Our goal was to determine the causes of ODE and various clinical characteristics of consecutive patients with ODE presenting to neuro-ophthalmology clinics.

METHODS

A retrospective review of consecutive patients with ODE over a period of 5 years were included. Fundus photographs were routinely obtained as part of clinical care. Clinical data including retinal nerve fiber layer thickness, best-corrected visual acuity, and visual field mean deviation were retrieved, and patients were grouped by etiology.

RESULTS

A total of 654 patients (n = 462 or 70.6% women and 192 or 29.4% men) with ODE were included with a mean age of 41.2 ± 17.9 years. Female patients were significantly younger than male patients (mean age female participants: 38.7 years, male participants: 47.6; P < 0.001). The top 5 most common causes of ODE in our clinics were idiopathic intracranial hypertension (IIH; 351/654 or 53.7%), nonarteritic anterior ischemic optic neuropathy (NAION; 116/654 or 17.4%), non-IIH papilledema (71/654 or 10.9%), optic neuritis (ON; 46/654 or 7.0%), and uveitis (17/654 or 2.6%). When considering female-only patients, the top 3 causes of ODE were IIH, non-IIH papilledema, and NAION. Among male-only patients, the top 3 causes were NAION, IIH, and non-IIH. Among the top 5 causes of ODE, visual acuity was the worst at presentation in the ON group and the best in the patients with IIH. The Humphrey mean deviation was the worst in ON/NAION groups and best in IIH group. The ODE was most severe in patients with non-IIH papilledema and least severe in ON group. Non-IIH patients with papilledema were not significantly different in visual acuity and visual field parameters at presentation compared with patients with IIH papilledema. Patients with papilledema (both IIH and non-IIH etiologies) had significantly better visual function at presentation compared with other top causes of disc edema (P < 0.001).

CONCLUSIONS

The most common cause of ODE seen in neuro-ophthalmology clinics was IIH, and these patients were more likely to present with preserved visual function. The higher prevalence of IIH was likely why most patients with ODE were women. Visual function at presentation was not able to differentiate the specific cause of papilledema; however, it was an important differentiating factor for all papilledema cases compared with all other causes of ODE including ON and NAION.

A spectrum of retinal vasculature measures and coronary artery disease

BACKGROUND AND AIMS

We aimed to comprehensively describe a spectrum of retinal vessel measures including fractal dimension (D_f) and their associations with indices of coronary artery disease (CAD) extent and severity, as well as hypertension and diabetes.

METHODS

The Australian Heart Eye Study (AHES) is an observational study that surveyed 1680 participants presenting to a tertiary referral hospital for the evaluation of potential CAD by coronary angiography. A range of newer retinal vessel geometric measures (D_f, curvature tortuosity, and branching angle) were quantified from retinal photographs using semi-automated software, the Singapore 'I' Vessel Assessment (SIVA) tool. A combined retinal score was constructed, aiming to assess the joint effect of multiple retinal vessel parameters on CAD, comprising of those variables that were most strongly significant in multivariate analysis - D_f, arteriolar curvature tortuosity, and retinal arteriolar calibre. CAD was objectively quantified using a range of measures obtained from coronary angiography.

RESULTS

A total of 1187 participants had complete data on retinal vessel measurements and coronary vessel evaluation. Retinal vascular D_f and curvature tortuosity decreased with increasing age; women had significantly lower D_f than men (p<0.003). Straighter retinal vessels were associated with CAD extent and Gensini scores in multivariable analysis (p<0.02). Accounting for media opacity by sub-group analysis in pseudophakic patients, the combined retinal score was associated with stenosis greater than 50% in any coronary artery segment (vessel score) and obstructive coronary stenosis in all three main coronary arteries (segment score) (p = 0.01). Lower D_f and narrower arteriolar branching angle were associated with CAD vessel score (p<0.03). In sex-stratified multivariate analyses, straighter arterioles were associated with greater odds of CAD in men, and narrower venular branching angle was associated with CAD in women.

CONCLUSIONS

A range of retinal vessel measures were associated with CAD extent and severity. A sparser retinal microvascular network (smaller D_f) was associated with older age and female gender. After accounting for the impact of media opacity, retinal vessel measures were associated with more diffuse and severe CAD.

Traumatic brain injury results in acute rarefication of the vascular network

The role of the cerebrovascular network and its acute response to TBI is poorly defined and emerging evidence suggests that cerebrovascular reactivity is altered. We explored how cortical vessels are physically altered following TBI using a newly developed technique, vessel painting. We tested our hypothesis that a focal moderate TBI results in global decrements to structural aspects of the vasculature. Rats (naïve, sham-operated, TBI) underwent a moderate controlled cortical impact. Animals underwent vessel painting perfusion to label the entire cortex at 1 day post TBI followed by whole brain axial and coronal images using a wide-field fluorescence microscope. Cortical vessel network characteristics were analyzed for classical angiographic features (junctions, lengths) wherein we observed significant global (both hemispheres) reductions in vessel junctions and vessel lengths of 33% and 22%, respectively. Biological complexity can be quantified using fractal geometric features where we observed that fractal measures were also reduced significantly by 33%, 16% and 13% for kurtosis, peak value frequency and skewness, respectively. Acutely after TBI there is a reduction in vascular network and vascular complexity that are exacerbated at the lesion site and provide structural evidence for the bilateral hemodynamic alterations that have been reported in patients after TBI.

Entropy of corneal nerve fibers distribution observed by laser scanning confocal microscopy: A noninvasive quantitative method to characterize the corneal innervation in Sjogren's syndrome patients

Sjogren's syndrome (SS) is a progressive autoimmune condition mainly affecting the salivary and lacrimal glands with an incidence of primary SS between 1/100 and 1/1,000. SS implies an alteration in the epithelium and subepithelium innervation, with consequent reduction of corneal sensitivity. It is necessary to have noninvasive quantitative methods to characterize the status of the corneal nerve fibers of the patients in order to choose and follow the best therapy. Entropy (information dimension) of the nerve corneal fibers distribution observed by confocal microscopy was evaluated in patients with primary SS (n = 30, 6 males, 24 females, 21-81 years), diagnosed by biopsy of salivary gland and blood tests and in sex- age-matched healthy subjects (n = 12). Corneal nerve fiber density, Langerhans cell count, and cell density in the nerve plexus images were also evaluated. In selected patients salivary gland atrophy degree was also evaluated. Nerve corneal distribution observed by confocal microscopy is fractal. Entropy of the corneal nerve distribution statistically distinguishes between SS patients and healthy subjects: patients present a lower value of information dimension of the corneal nerve fibers distribution than healthy individuals (P < 0.001). Percentage of grouped cases classified by entropy according to the subjects (selected patients vs. healthy) showed a 100% sensitivity and 96% specificity, P < 0.0001 with a low value of coefficient of variation among the individuals (6-7 times lower than the other morphometric indexes). Entropy correlated with the severity of the disease (salivary gland atrophy degree, P < 0.01). Evaluation of entropy of the corneal nerve distribution observed by a laser confocal microscopy appears to quantitatively and noninvasively characterize an aspect of the SS patients in relation to the recognition of an impairment of their ocular surface, giving us for the first time a method to objectively and precisely characterize the corneal innervation status in the SS patients.

Age-related rarefaction in retinal vasculature is not linear

The fractal dimension is a global measure of complexity and is useful for quantifying anatomical structures, including the retinal vascular network. A previous study found a linear declining trend with aging on the retinal vascular fractal dimension (D_F); however, it was limited to the older population (49 years and older). This study aimed to investigate the possible models of the fractal dimension changes from young to old subjects (10-73 years). A total of 215 right-eye retinal samples, including those of 119 (55%) women and 96 (45%) men, were selected. The retinal vessels were segmented using computer-assisted software, and non-vessel fragments were deleted. The fractal dimension was measured based on the log-log plot of the number of grids versus the size. The retinal vascular D_F was analyzed to determine changes with increasing age. Finally, the data were fitted to three polynomial models. All three models are statistically significant (Linear: R² = 0.1270, 213 d.f., p < 0.001, Quadratic: R² = 0.1536, 212 d.f., p < 0.001, Cubic: R² = 0.1529, 211 d.f., p < 0.001). The quadratic regression is significantly better than the linear regression (p < 0.001); however, the increase in R² from the quadratic model to the cubic model is not significant (p = 0.97). These results suggest that the decreasing trend of the fractal dimension associated with aging is better explained by the quadratic model than by the linear and cubic models in a sample with a broader age spectrum.

Retinal vascular fractal dimension and its relationship with cardiovascular and ocular risk factors

PURPOSE

To examine the influence of a range of cardiovascular risk factors and ocular conditions on retinal vascular fractal dimension in the Singapore Malay Eye Study.

DESIGN

Population-based cross-sectional study.

METHODS

Fractal analysis of the retinal vessels is a method to quantify the global geometric complexity of the retinal vasculature. Retinal vascular fractal dimension (D(f)) and caliber were measured from retinal photographs using a computer-assisted program. D(f) and arteriolar caliber were combined to form a retinal vascular optimality score (ranging from 0 to 3). Data on cardiovascular and ocular factors were collected from all participants based on a standardized protocol.

RESULTS

Two thousand nine hundred thirteen (88.8% of 3280 participants) persons had retinal photographs of sufficient quality for the measurement. The mean D(f) was 1.405 (standard deviation, 0.046; interquartile range, 1.243 to 1.542). In the multiple linear regression analysis, after controlling for gender, serum glucose, intraocular pressure, anterior chamber depth, and retinal vascular caliber, smaller D(f) was associated independently with older age (standardized regression coefficient [sβ] = -0.311; P < .001), higher mean arterial blood pressure (sβ = -0.085; P < .001), a more myopic spherical equivalent (sβ = 0.152; P < .001), and presence of cataract (sβ = -0.107; P < .001). Retinal vascular optimality score was associated significantly with higher mean arterial blood pressure (P > .001 for trend).

CONCLUSIONS

Age, blood pressure, refractive error, and lens opacity had significant influence on retinal vascular fractal measurements. A new score of retinal vascular optimality combining fractals and caliber showed strong association with blood pressure. Quantitative analysis of retinal vasculature therefore may provide additional information on microvascular architecture and optimality.

Lens opacity and refractive influences on the measurement of retinal vascular fractal dimension

PURPOSE

To examine the influence of lens opacity and refraction on the measurement of retinal vascular fractal dimension (Df).

METHODS

  Optic disc photographs (right eyes) of 3654 baseline Blue Mountains Eye Study participants (aged 49-97) were digitized. Retinal vascular Df was quantified using a computer-based program. Summated severity scores for nuclear, cortical and posterior subcapsular (PSC) cataract were assessed from lens photographs. Refraction data were converted to spherical equivalent refraction (SER), as sum spherical plus 0.5 cylinder power. Axial length was measured at 10-year follow-up examinations using an IOL master.

RESULTS

  Mean Df of the retinal vasculature was 1.444±0.023 for 2859 eligible participants. Increasing lens opacity scores were associated with significant reduction in Df (β=-0.0030, p<0.0001). Both cortical and PSC cataract involving central lens area were associated with reduced Df, after controlling for confounding factors (p(trend) ≤0.0105). Increasing myopia severity was associated with reduced Df after adjusting for lens opacity scores and other confounders (p(trend) <0.0001). The slope of Df decrease per SER reduction was 0.0040 in eyes with SER≤-4D, compared to -0.0016 in eyes with SER>-4D. For axial length quintiles, there were no significant differences in mean Df in all groups except a reduction in the fifth quintile (axial length ≥24.15mm) (all p<0.05).

CONCLUSION

  Ocular media opacity independently influenced retinal vascular Df measurement, but we found no evidence supporting any refractive axial magnification effect on this measure. Myopic refraction ≤-4D was associated with a reduction in Df, suggesting rarefaction of retinal vasculature associated with high myopia.

Retinal fractals and acute lacunar stroke

This study aimed to determine whether retinal fractal dimension, a quantitative measure of microvascular branching complexity and density, is associated with lacunar stroke. A total of 392 patients presenting with acute ischemic stroke had retinal fractal dimension measured from digital photographs, and lacunar infarct ascertained from brain imaging. After adjusting for age, gender, and vascular risk factors, higher retinal fractal dimension (highest vs lowest quartile and per standard deviation increase) was independently and positively associated with lacunar stroke (odds ratio [OR], 4.27; 95% confidence interval [CI], 1.49-12.17 and OR, 1.85; 95% CI, 1.20-2.84, respectively). Increased retinal microvascular complexity and density is associated with lacunar stroke.