METHODS: Firstly, color fundus images from the publicly available database DRIVE were converted from RGB to grayscale. To enhance the contrast of the dark objects (blood vessels) against the background, the dot product of the grayscale image with itself was generated. To rectify the variation in contrast, we used a 5 × 5 window filter on each pixel. Based on 5 regional features, 1 intensity feature and 2 Hessian features per scale using 9 scales, we extracted a total of 24 features. A linear minimum squared error (LMSE) classifier was trained to classify each pixel into a vessel or non-vessel pixel.
RESULTS: The DRIVE dataset provided 20 training and 20 test color fundus images. The proposed algorithm achieves a sensitivity of 72.05% with 94.79% accuracy.
CONCLUSIONS: Our proposed algorithm achieved higher accuracy (0.9206) at the peripapillary region, where the ocular manifestations in the microvasculature due to glaucoma, central retinal vein occlusion, etc. are most obvious. This supports the proposed algorithm as a strong candidate for automated vessel segmentation.
METHODS: This was a cross-sectional study involving 166 children aged 6 to 12 years old in Malaysia. Ocular examination, biometry, retinal photography, blood pressure and body mass index measurement were performed. Participants were divided into two groups; obese and non-obese. Retinal vascular parameters were measured using validated software.
RESULTS: Mean age was 9.58 years. Approximately 51.2% were obese. Obese children had significantly narrower retinal arteriolar caliber (F(1,159) = 6.862, p = 0.010), lower arteriovenous ratio (F(1,159) = 17.412, p < 0.001), higher venular fractal dimension (F(1,159) = 4.313, p = 0.039) and higher venular curvature tortuosity (F(1,158) = 5.166, p = 0.024) than non-obese children, after adjustment for age, gender, blood pressure and axial length.
CONCLUSIONS: Obese children have abnormal retinal vascular geometry. These findings suggest that childhood obesity is characterized by early microvascular abnormalities that precede development of overt disease. Further research is warranted to determine if these parameters represent viable biomarkers for risk stratification in obesity.
METHODS: This was a cross-sectional, hospital-based study involving 86 Malay girls aged 6 to 12 years old in Hospital Universiti Sains Malaysia from 2015-2016. Ocular examination, refraction, biometry, retinal photography, and anthropometric measurements were performed. The central retinal arteriolar equivalent (CRAE), central retinal venular equivalent (CRVE) and overall fractal dimension (Df) were measured using validated computer-based methods (Singapore I vessel analyzer, SIVA version 3.0, Singapore). The associations of ocular biometry and CRAE, CRVE and Df were analyzed using multivariable analysis.
RESULTS: The mean CRAE, CRVE and Df in Malay girls were 171.40 (14.40) um, 248.02 (16.95) um and 1.42 (0.05) respectively. Each 1 mm increase in axial length was associated with a reduction of 4.25 um in the CRAE (p = 0.03) and a reduction of 0.02 in the Df (p = 0.02), after adjustment for age, blood pressure and body mass index. No association was observed between axial length and CRVE. Anterior chamber depth and corneal curvature had no association with CRAE, CRVE or Df.
CONCLUSION: Axial length affects retinal vessel measurements. Narrower retinal arterioles and reduced retinal fractal dimension were observed in Malay girls with longer axial lengths.
METHODOLOGY: We performed a cross-sectional cohort study on healthy subjects and patients with glaucoma. The AngioVue Enhanced Microvascular Imaging System was used to capture the optic nerve head and macula images during one visit. En face segment images of the macular and optic disc were studied in layers. Microvascular density of the optic nerve head and macula were quantified by the number of pixels measured by a novel in-house developed software. Areas under the receiver operating characteristic curves (AUROC) were used to determine the accuracy of differentiating between glaucoma and healthy subjects.
RESULTS: A total of 24 (32 eyes) glaucoma subjects (57.5±9.5-y old) and 29 (58 eyes) age-matched controls (51.17±13.5-y old) were recruited. Optic disc and macula scans were performed showing a greater mean vessel density (VD) in healthy compared with glaucoma subjects. The control group had higher VD than the glaucoma group at the en face segmented layers of the optic disc (optic nerve head: 0.209±0.05 vs. 0.110±0.048, P<0.001; vitreoretinal interface: 0.086±0.045 vs. 0.052±0.034, P=0.001; radial peripapillary capillary: 0.146±0.040 vs. 0.053±0.036, P<0.001; and choroid: 0.228±0.074 vs. 0.165±0.062, P<0.001). Similarly, the VD at the macula was also greater in controls than glaucoma patients (superficial retina capillary plexus: 0.115±0.016 vs. 0.088±0.027, P<0.001; deep retina capillary plexus: 0.233±0.027 vs. 0.136±0.073, P<0.001; outer retinal capillary plexus: 0.190±0.057 vs. 0.136±0.105, P=0.036; and choriocapillaris: 0.225±0.053 vs. 0.153±0.068, P<0.001. The AUROC was highest for optic disc radial peripapillary capillary (0.96), followed by nerve head (0.92) and optic disc choroid (0.76). At the macula, the AUROC was highest for deep retina (0.86), followed by choroid (0.84), superficial retina (0.81), and outer retina (0.72).
CONCLUSIONS: Microvascular density of the optic disc and macula in glaucoma patients was reduced compared with healthy controls. VD of both optic disc and macula had a high diagnostic ability in differentiating healthy and glaucoma eyes.
METHODS: Cross sectional observational cohort study. Subjects with normal eyes were recruited. Two sets of optical coherence tomography angiography images of macula and optic nerve head were acquired during one visit. Novel in-house developed software was used to count the pixels in each images and to compute the microvessel density of the macula and optic disc. Data were analysed to determine the measurement repeatability.
RESULTS: A total of 176 eyes from 88 consecutive normal subjects were recruited. For macular images, the mean vessel density at superficial retina, deep retina, outer retina and choriocapillaries segment was OD 0.113 and OS 0.111, OD 0.239 and OS 0.230, OD 0.179 and OS 0.164, OD 0.237 and OS 0.215 respectively. For optic disc images, mean vessel density at vitreoretinal interface, radial peripapillary capillary, superficial nerve head and disc segment at the level of choroid were OD 0.084 and OS 0.085, OD 0.140 and OS 0.138, OD 0.216 and OS 0.209, OD 0.227 and OS 0.236 respectively. The measurement repeatability tests showed that the coefficient of variation of macular scans, for right and left eyes, ranged from 6.4 to 31.1% and 5.3 to 59.4%. Likewise, the coefficient of variation of optic disc scans, for right and left eyes, ranged from 14.3 to 77.4% and 13.5 to 75.3%.
CONCLUSIONS: Optical coherence tomography angiography is a useful modality to visualise the microvasculature plexus of macula and optic nerve head. The vessel density measurement of macular scan by mean of optical coherence tomography angiography demonstrated good repeatability. The optic disc scan, on the other hand, showed a higher coefficient of variation indicating a lower measurement repeatability than macular scan. Interpretation of optical coherence tomography angiography should take into account test-retest repeatability of the imaging system.
TRIAL REGISTRATION: National Healthcare Group Domain Specific Review Board ( NHG DSRB ) Singapore. DSRB Reference: 2015/00301.