Access the full text.
Sign up today, get DeepDyve free for 14 days.
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
Hindawi Journal of Healthcare Engineering Volume 2022, Article ID 6726499, 9 pages https://doi.org/10.1155/2022/6726499 Review Article Application of the Full-Width-at-Half-Maximum Image Segmentation Method to Analyse Retinal Vascular Changes in Patients with Diabetic Retinopathy 1 2 3 1 4 1 Bo Lun Xu, Yi Jie Li, Wen Li Zhou, Hai Jing Zhan, Jie Yi Lu, and Yu Hua Tong e Quzhou Aliated Hospital of Wenzhou Medical University, Quzhou People’s Hospital, Quzhou, Zhejiang, China e First Aliated Hospital of Jiangxi Medical College, Shangrao, Jiangxi, China Shanghai Sixth People’s Hospital, Shanghai, China e Second Clinical Medical College, Zhejiang Chinese Medical University, Hangzhou, China Correspondence should be addressed to Yu Hua Tong; yuhuatong@126.com Received 5 April 2022; Revised 29 May 2022; Accepted 14 July 2022; Published 8 August 2022 Academic Editor: Sweta Bhattacharya Copyright © 2022 Bo Lun Xu et al. is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. is study used spectral domain optical coherence tomography (SD-OCT) and full-width-at-half-maximum image segmentation to investigate the morphological changes of retinal blood vessels in patients with diabetic retinopathy (DR). Seventy-Œve patients with type 2 diabetes mellitus (T2DM) without DR and 65 patients with DR were studied. Vascular images of superior temporal region B of the retina were obtained by SD-OCT. e edges of retinal vessels were identiŒed by the full-width-at-half-maximum image segmentation method. e lumen diameter, wall thickness (WT), wall cross-sectional area (WCSA), and wall-to-lumen ratio (WLR) were investigated. We found that, compared with the no diabetic retinopathy (NDR) group, patients in the DR group had an increased retinal arteriolar lumen diameter (RALD), retinal arteriolar outer diameter (RAOD), and WT (128.80 μm versus 104.88 μm; 147.01 μm versus 135.60 μm; 18.29 μm versus 15.26 μm, P < 0.05, respectively). e retinal venular lumen diameter (RVLD), retinal venular outer diameter (RVOD), and venular WT in the DR group were also increased (146.17 μm versus 133.66 μm; 180.20 μm versus 156.43 μm; 17.01 μm versus 11.38 μm, P < 0.05, respectively). e morphological changes in retinal vessels were signiŒcantly correlated with DR stage. In conclusion, in diabetic patients with DR, both retinal arteries and veins are widened and exhibit increased vascular thickness. United States showed that this approach can reduce the 1. Introduction blindness rate from 50% to less than 5%. erefore, early Diabetic retinopathy (DR) is a very common chronic detection of DR and timely and eœective early treatment are complication of diabetes. Approximately 30% of diabetic key to reducing the incidence of visual disability in diabetic patients have DR lesions, and approximately 30% of DR patients. Regular DR screening for diabetic patients is an eœective measure to reduce blindness caused by diabetes. patients will suœer from visual loss due to severe retinopathy. e data from the Early Treatment Diabetic Retinopathy DR is mainly a fundus change due to diabetes-induced Study research group and the Diabetic Retinopathy Study retinal microangiopathy. InsuŸcient oxygen supply to ret- research group showed that regular follow-up, necessary and inal vessels and the local in¡ammatory response are the appropriate vitreous surgery, retinal laser photocoagulation, pathological mechanisms underlying the development of and other interventions can reduce the risk of severe vision DR, and these changes in the intraocular microenvironment loss in 90% of DR patients. If patients with proliferative DR can lead to compensatory dilation of retinal blood vessels. can be evaluated and treated early, the blindness rate of DR For example, irregular dilation of retinal veins (venous will be signiŒcantly reduced. e results of a study in the beading) is a typical clinical feature of severe DR. During DR 2 Journal of Healthcare Engineering pathogenesis or DR progression, in addition to the risk table, and look at the blue light inside the lens. .e retinal factors such as glycosylated haemoglobin concentration, artery and its accompanying veins in superior temporal region B of the right eye were scanned with an SD-OCT blood pressure, and duration of diabetes, recent epidemi- ological studies have also shown that retinal vasodilatation is instrument. Region B was defined as the superior temporal a risk factor for the progression of DR patients. .e dilation region at a distance of 0.5 to 1.0 disc diameter from the optic of retinal veins is a high-risk factor for the pathogenesis of disc edge (Figure 1(a)). First, in terms of morphology, the DR, which may be related to the dysfunction of retinal characteristics of the blood vessels in region B were more vascular endothelial cells, changes in the levels of inflam- consistent with the description of arterioles and venules. matory factors, and hyperglycaemia, and these factors all Second, after the retinal blood vessels emanate from the play important roles in the pathogenesis and progression of centre of the optic disc, arteriovenous crossing and pulsation DR. of the retinal arteries were less frequent in the region B, a In this study, DR, a common eye disease that seriously certain distance from the optic disc, which did not affect the threatens the quality of human life, was studied. We measurement of the diameters of the retinal artery or veins. .e scanning line was adjusted to be as nearly perpendicular quantitatively analysed the lumen diameter (LD) of retinal large vessels in patients with DR by spectral domain optical to the vascular axis as possible. If the vessel had branched coherence tomography (SD-OCT) and explored the corre- before reaching region B, then the part before branching was lation between the LD of retinal large vessels and the scanned. After scanning, the vertical-to-horizontal ratio of pathogenesis and progression of DR in order to provide an the obtained OCT images was first adjusted to 1:1 μm, the accurate, objective, and quantitative method for the early image was magnified to 800%, and then the OCT image was detection and monitoring of DR. saved in BMP format. Only the images that clearly showed the vessel wall were used for the following analysis. At least five clearly scanned OCT images of every vessel were taken 2. Methods for analysis. 2.1. General Data and Grouping. .e study subjects were randomly selected from patients who visited the outpa- tient clinic of Quzhou People’s Hospital from January 2.2.2. Methods of Measuring Retinal Vessels. .e full-width- 2016 to November 2017. .e study was approved by the at-half-maximum (FWHM) method was used to measure Research Ethics Committee of Quzhou People’s Hospital the LD, outer diameter (OD), and blood vessel wall thickness and was conducted in adherence to the tenets of the (WT) of the retinal blood vessels in the OCT images. .e Declaration of Helsinki. Informed consent was obtained OCT images in BMP format were opened in the ImageJ from all participants. Enrolment criteria for the experi- software (National Institutes of Health), and the line tool mental group were as follows: patients above 50 years of was used to draw a vertical line through the middle of the age who were diagnosed with type 2 diabetes mellitus blood vessel to obtain the density curve. .ere were two (T2DM) in the Department of Endocrinology, the pa- concave-upward curves on the density plot, representing the tients or their family members who could clearly describe upper and lower walls of the vessel in the OCTimage. On the a history of diabetes and their diagnostic and treatment left and right sides of each curve, the maximum and min- process, and patients who could cooperate with fundus imum values of the crest and trough were determined using examination after cycloplegia. Exclusion criteria were as the average of three consecutive values, and the median follows: patients with other types of diabetes, value between the maximum and minimum values was hypertension, and previous or current cardiovascular calculated. On each side of the curve, a linear function was disease (excluding cerebrovascular disease), myopia fitted with continuous points with the greatest difference, exceeding −6.0 D, severe refractive medium opacity, and the intersection of this linear function with the hori- intraocular pressure higher than 21 mmHg, history of zontal line of the median value was the position of the edge. glaucoma, ocular ischaemic syndrome, uveitis, ocular Finally, ImageJ software automatically identified the dis- space-occupying lesions, retinal disease, or previous laser tance between the boundary points of the two curves and or phacoemulsification surgery, rheumatic immune calculated the LDs and ODs of the retinal vessels (Figure 1). disease, pulmonary heart disease, severe carotid artery All procedures were performed by the same experienced stenosis, nephropathy, or a history of smoking and ophthalmologist. Because there is a good correlation be- drinking. A total of 140 patients (58 females and 82 tween the measured retinal vessel diameters of the left eye males) were included. .e patients were divided into a no and those of the right eye, only the retinal vessels in the right DR (NDR) group (75 patients) and a DR group (65 eye were measured in each patient. Each vessel of all subjects patients) according to the presence or absence of DR. was measured three times with the FWHM method, and the average value was taken to obtain the retinal arteriolar lumen diameter (RALD), retinal arteriolar outer diameter (RAOD), 2.2. Experimental Procedures retinal venular lumen diameter (RVLD), and retinal venular outer diameter (RVOD). Other morphological parameters 2.2.1. Acquisition of OCT Images of Retinal Blood Vessels. of retinal vessels were calculated as described [1]. .e for- All OCT images were acquired with an SD-OCT instrument mulas for the retinal artery are as follows: (Heidelberg Engineering, Heidelberg, Germany). .e sub- WT �(RAOD −RALD)/2; WLR �(RAOD −RALD)/RALD; ject was asked to sit down, place the head on the examination Journal of Healthcare Engineering 3 (a) (b) outer diameter 150 lumen diameter 0 20 40 60 80 100 120 140 Distance (pixels) (c) (d) Figure 1: Schematic diagram of measuring retinal blood vessels on OCTimages using the FWHM method. (a) Schematic diagram of region B positioning. .e scanning line runs perpendicular to the axis of a retinal vessel in region B with the red arrow indicating the artery and the blue arrow indicating the vein. (b) .e OCT image clearly showing the cross-sections of the retinal artery (red arrow) and the retinal vein (blue arrow). (c) A line through the centre of the circle drawn between the upper and lower walls of a vessel to produce a greyscale density curve of the vessel wall. (d) .e boundaries of the upper and lower walls of the vessel determined with the FWHM segmentation method (shown by arrows), and the luminal and outer diameters of the vessel are calculated after the boundaries are determined. 2 2 WCSA �3.14 ×(RAOD −RALD )/4. .e formulas for the 2.2.5. DR Staging. .e stage of DR was determined retinal vein were the same as those of the artery, and the according to the staging criteria in the 2014 Chinese corresponding venous values could be substituted into the Guideline for Clinical Diagnosis and Treatment of Dia- formula. .e results of the NDR group and DR group were betic Retinopathy. Stage I was the mild nonproliferative compared. phase, stage II was the moderate nonproliferative phase, stage III was the severe nonproliferative phase, stage IV was the early proliferative phase, stage V was the fibro- 2.2.3. Statistical Analytical Methods. SPSS 21.0 was used. Measurement data are expressed as x± s. Student’s t-test was proliferative phase, and stage VI was the late proliferative phase. used to compare the retinal vessel diameters between the two groups. .e χ test was used to compare count data between 3. Results the groups. Univariate analysis was performed to detect correlations between the structural parameters (LD and the 3.1. Comparison of General Data (Table 1). In the NDR OD, WLR, WT, and WCSA of retinal blood vessels) and group there were 39 males and 36 females, and in the DR other parameters (age, DR stage, and duration of diabetes). group there were 43 males and 22 females (χ test Multivariate linear regression was performed with the LD, P>0.05). .e independent-samples t-test was used for the OD, WLR, WT, and WCSA of retinal blood vessels as de- comparison of age (NDR: 62.40± 10.05; DR: 65.50± 8.41) pendent variables and the DR stage, duration of diabetes, and body mass index (NDR: 25.26± 3.32; DR: and age as independent variables. P values <0.05 were 24.30± 3.32), and the differences were not significant considered statistically significant. (both P>0.05). 2.2.4. Inspection Items. Fundus examination was performed 3.1.1. Disease Staging of DR Patients. In the DR group, 37 under cycloplegia to determine the stage of DR, and the patients had stage I DR, 11 patients had stage II DR, 11 patient’s history of systemic diseases and ocular diseases was patients had stage III DR, four patients had stage IV DR, and collected. two patients had stage V DR. Gray Value 4 Journal of Healthcare Engineering Table 1: General clinical characteristics of patients in the two groups. Number of cases Sex, male (female) Age Body mass index NDR group 75 39 (36) 62.40±10.05 25.26±3.32 DR group 65 43 (22) 65.50±8.41 24.30±3.32 P value 0.090 0.051 0.092 χ or t value 2.875 −1.965 1.699 Note. Body mass index and age were analysed by the independent-samples t-test, and sex was analysed by the χ test. 3.2. Comparison of Retinal Arteries between the NDR and DR (b �3.193, P<0.001) and WT (R �0.274, F �23.817), and Groups (Table 2). .e RALD, RAOD, and WT of the NDR the regression equation was WT �9.997+3.193 ×DR stage. group were 104.88±15.68 μm, 135.60±17.22 μm, and .ere was a linear relationship between age (b �0.002, 15.26±2.56 μm, compared to 128.80±36.00 μm, P � 0.029) and WLR (R �0.074, F �5.006), and the re- 147.01±17.55 μm, and 18.29±5.97 μm in the DR group gression equation was WLR �0.162+0.002 ×age. (t � −5.214, P �0.000; t � −3.872, P � 0.000; t � −3.998, P � 0.000, respectively). .e WLR of the NDR group 3.7. Multivariate Linear Regression Analysis of Retinal Vein (0.29±0.06) was larger than that of the DR group Variables (Table 7). .ere was a linear relationship between (0.28±0.05), but the difference was not significant (t �1.756, DR stage (b �33.287, P<0.001) and RVLD (R �0.346, P � 0.081). .e WCSA (5685.20 ±1255.26 μm ) of the NDR F �18.908), and the regression equation was group was significantly smaller than that (6647.51± RVLD �92.532+20.632 × DR stage. .ere was a linear 1590.52 μm ) of the DR group (t � −3.333, P � 0.001). relationship between DR stage (b �23.819, P<0.001) and RVOD (R �0.364, F �36.012), and the regression equation 3.3. Comparison of Retinal Veins between the NDR and DR was RVOD �118.273+23.819 ×DR stage. .ere was a linear Groups (Table 3). .e RVLD, RVOD, and WT of the NDR relationship between DR stage (b �1.593, P � 0.003) and group were 133.66±21.07 μm, 156.43±23.36 μm, and 2 WT (R �0.136, F �9.916), and the regression equation was 11.38±2.33 μm, respectively, compared to 146.17± WT �12.871+1.593 ×DR stage. .ere was a linear rela- 34.42 μm, 180.20±38.70 μm, and 17.01±4.23 μm in the DR tionship between diabetes duration (b �0.004, P � 0.024) group (t � −2.545, P � 0.012; t � −4.313, P � 0.000; 2 and WLR (R �0.078, F �5.344), and the regression equa- t � −9.529, P � 0.000). .e WLR (0.17 ±0.03) in the NDR tion was WLR �0.208+0.004 ×diabetes duration. .ere was group was significantly smaller than that (0.23±0.05) in the a linear relationship between DR stage (b �2251.506, DR group (t � −7.827, P � 0.000). .e WCSA 2 P<0.001) and WCSA (R �0.321, F �29.814), and the re- (5321.03±1695.19 μm ) in the NDR group was significantly gression equation was WCSA �3108.217+2251.506 ×DR smaller than that (8962.13±3897.34 μm ) in the DR group stage. (t � −6.982, P � 0.000). 4. Analysis and Discussion 3.4. Retinal Artery-Associated Factors (Table 4). RALD was Retinal vessels are the only blood vessels in human body that positively correlated with DR stage (r �0.480; P<0.001). can be directly observed and measured. .ey are small blood WT was positively correlated with DR stage (r �0.524; vessels that are similar in structure and physiological P<0.001). WLR was positively correlated with age function to the terminal blood vessels of important organs (r �0.271; P � 0.029). Other variables were not correlated. such as the heart, brain, and kidney. .erefore, the change in the diameters of retinal vessels can reflect the physiological and pathological changes in other organs, tissues, and their 3.5. Retinal Vein-Associated Factors (Table 5). RVLD was blood vessels to a certain extent, so their lesions have an positively correlated with DR stage (r �0.588; P<0.00); important auxiliary role in the diagnosis of other systemic RVOD was positively correlated with DR stage (r �0.603; diseases. P<0.001). WT was positively correlated with DR stage Diabetes mellitus is a chronic metabolic disease char- (r �0.369; P � 0.003). WLR was positively correlated with acterized by hyperglycaemia. It can be complicated by diabetes duration (r �0.280; P � 0.024). WCSA was chronic lesions of multiple organs such as eyes, kidneys, positively correlated with DR stage (r �0.567; P<0.001). nerves, and blood vessels. DR will cause decreased visual Other variables were not correlated. acuity and even blindness in patients. Retinal vessels can be examined to study the relationship between changes in the 3.6. Multivariate Linear Regression Analysis of Retinal Artery morphological characteristics of retinal vessels and diabetes Variables (Table 6). Among the factors age, DR stage, and [2]. However, there are few methods of retinal vascular diabetes duration, only DR stage (b �17.633, P<0.001) had monitoring. .e morphology of retinal vessels is often a linear relationship with RALD (R �0.231, F �18.908), and observed and followed up with the help of ophthalmoscopy, the regression equation was RALD �82.963+17.633 ×DR fluorescein angiography, colour fundus photography, and stage. .ere was a linear relationship between DR stage other means in clinical practice to track the changes in Journal of Healthcare Engineering 5 Table 2: Parameters of retinal arteries in the NDR and DR groups. RALD (μm) RAOD (μm) WLR (μm) WT (μm) WCSA (μm ) NDR group 104.88±15.68 135.60±17.22 0.29±0.06 15.26±2.56 5824.64±1330.07 DR group 128.80±36.00 147.01±17.55 0.28±0.05 18.29±5.97 6647.51±1590.52 P value <0.001 <0.001 0.081 <0.001 <0.001 t value −5.214 −3.872 1.756 −3.998 −3.333 Table 3: Parameters of retinal veins in the NDR and DR groups. RVLD (μm) RVOD (μm) WLR (μm) WT (μm) WCSA (μm ) NDR group 133.66±21.07 156.43±23.36 0.17±0.03 11.38±2.33 5321.03±1695.19 DR group 146.17±34.42 180.20±38.70 0.23±0.05 17.01±4.23 8962.13±3897.34 P value 0.012 <0.001 <0.001 <0.001 <0.001 t value −2.545 −4.313 −7.827 −9.529 −6.982 Table 4: Correlations between retinal artery-associated factors. RALD RAOD WT WLR WCSA r value 0.031 −0.013 0.161 0.271 0.125 Age P value 0.807 0.915 0.200 0.029 0.320 r value 0.480 −0.139 0.524 0.142 −0.063 DR stage P value <0.001 0.269 <0.001 0.259 0.619 r value −0.087 0.047 −0.184 −0.136 −0.026 Diabetes duration P value 0.489 0.708 0.142 0.281 0.835 Table 5: Correlations between retinal vein-associated factors. RVLD RVOD WT WLR WCSA r value 0.052 0.074 0.127 0.114 0.107 Age P value 0.679 0.556 0.312 0.367 0.395 r value 0.588 0.603 0.369 −0.090 0.567 DR stage P value <0.001 <0.001 0.003 0.474 <0.001 r value −0.075 −0.017 0.226 0.280 0.092 Diabetes duration P value 0.552 0.891 0.070 0.024 0.464 Table 6: Correlations between retinal artery-associated factors (multivariate regression). Equation R F b P value RALD �82.963+17.633 ×DR stage 0.231 18.908 17.633 <0.001 WT �9.997+3.193 ×DR stage 0.274 23.817 3.193 <0.001 WLR �0.162+0.002 ×age 0.074 5.006 0.002 0.029 Table 7: Correlations between retinal vein-associated factors (multivariate regression). Equation R F b P value RVLD �92.532+20.632 ×DR stage 0.346 18.908 33.287 <0.001 RVOD �118.273+23.819 ×DR stage 0.364 36.012 23.819 <0.001 WT �12.871+1.593 ×DR stage 0.136 9.916 1.593 0.003 WLR �0.208+0.004 ×diabetes duration 0.078 5.344 0.004 0.024 WCSA �3108.217+2251.506 ×DR stage 0.321 29.814 2251.506 <0.001 arterial and venous diameters, and tools for more accurate blood vessel measurement software was used to obtain a and objective quantitative detection are lacking. At present, fundus photograph centred on the optic disc. Blood vessels many large institutions that conduct epidemiological in- in the annular area of 0.5D–1.0D from the edge of the optic vestigations obtain retinal blood vessel-related parameters disc were imaged, and the six largest arterioles and venules using fundus photography. In this study, semiautomatic were used to measure the retinal blood vessels. .e relative 6 Journal of Healthcare Engineering patients were much larger than those in non-DR patients, central retinal artery equivalent (CRAE) and central retinal vein equivalent (CRVE) data were obtained by conversion. and that they were all linearly correlated with DR stage. .e more severe the DR, the larger the diameters of the retinal Recently, researchers reported the use of deep learning models to automatically measure retinal vessel diameters in vein and the thicker the vessel wall. However, there was no fundus photographs [3]. Others utilize a system called correlation between the duration of diabetes and retinal EyeArt, a cloud-based automated AI eye screening tech- vessel diameter. In a recent development and validation nology that can effectively help endocrinologists, diabetol- study of a deep learning model involving more than 70,000 ogists, and general practitioners address the growing colour fundus photographs, researchers found that higher concerns regarding DR screening and monitoring [4]. glycated-haemoglobin levels were associated with a wider CRVE, which is consistent with our findings [3]. After the SD-OCT is an important imaging examination system for retinal tomography analysis. It is mainly used for occurrence of DR, if the retinal vein diameter is enlarged, this may predict the progression of DR. Clinically, it is quantitative and qualitative analysis of retinal tissue struc- ture. It has the advantages of clear imaging, high accuracy, generally accepted that patients with a long duration of diabetes and poorly controlled fasting blood glucose are capacity for dynamic continuous tomography analysis, and no invasive damage [5]. At the same time, SD-OCT has eye- more likely to develop DR and experience DR progression, tracking function and can lock onto the blood vessels to be reflecting the inseparable relationships between the diameter measured during eye rotation. .erefore, it is suitable to be of retinal veins, fasting blood glucose level, and diabetes used as a tool for analysing retinal vascular structure and duration. It is certain that, in patients with T2DM, changes obtaining tomographic images of blood vessels, which are in retinal vascular diameter, especially venous diameter, are more intuitive and convenient and can accurately locate and a warning sign for the occurrence and progression of mi- crovascular complications in T2DM, and changes in venous repeat the detection. In recent years, microdensitometry has been widely used for such purposes as the measurement and diameter are the earliest microcirculatory abnormalities in patients with T2DM. evaluation of retinal blood vessels in fundus photography [6,7]. .e more widely used method is the FWHM algo- Although the likelihood of concurrent DR and severe progression increase with a longer diabetes duration, the rithm, which is a segmentation method based on the grey value of the image. It is often applied to the identification diabetes duration is not completely equivalent to that of DR, and measurement of the edges of CT and magnetic reso- and our multivariate linear regression analysis also excluded nance images [8]. It can more quickly and stably determine the correlation between diabetes duration and DR. .is blood vessel boundaries and is less sensitive to the inter- study also provides more detailed and practical reference ference of noise points and adjacent tissues [9]. .erefore, indices of WT and WCSA. Because the impact of vessel the difference and innovation in this study compared with pulsation on the vessel wall is less than the impact of vessel pulsation on vessel diameter, the measures such as WT and other studies were that we used the FWHM segmentation method to obtain truly accurate LDs and ODs of retinal WCSA have higher reproducibility and higher research value. Some studies suggest that inflammatory response blood vessels on SD-OCT images. Additionally, other structural values, such as WT, wall cross-sectional area leads to venous distension and increases WT [10], further (WCSA), and wall-to-lumen ratio (WLR), were obtained on increasing WCSA, which is consistent with the results of this SD-OCT images. Analysing SD-OCT data is more reason- study. Moreover, retinal blood circulation disorders in di- able and accurate for detecting vascular structural changes abetic patients often damage the function of vascular en- caused by the long-term microcirculatory hypoxia and the dothelial cells, raise the expression of nitric oxide in stimulation of inflammatory factors in DR patients. For endothelial cells, mediate retinal flash stimulation through example, in the early stage of DR, when the diameter of nitric oxide [11, 12], and thus participate in the regulatory retinal vessels remains unchanged, there is still a possibility process of vasodilatation. Some research also suggests that hyperglycaemia can cause retinal hypoxia and lactate ac- of damage to endothelial cell function, leading to changes in WT and increases in the LD of blood vessels. However, the cumulation, and venous dilation is a compensatory mech- anism to increase blood supply to the retina [13]. Due to the SD-OCT instrument cannot automatically locate the retinal region B, so, in this experiment, a transparent film marked limitations of the present experiments, the sensitivity of WT, with auxiliary lines was used to help the examiner locate that RVOD, and RVLD and the timing of their changes could not region (Figure 1(a)) to improve the operability and scien- be compared, though they should be in the future. tificity of the experiment. In this experiment, the RAOD, RALD, WT, and WCSA In some early studies, it was found that DR development of retinal arteries in DR patients were significantly greater widened the retinal veins. However, due to technical limi- than those in non-DR patients. Moreover, RALD and WT were linearly correlated with DR stage. Cheung et al. [14] tations, it was difficult to quantitatively assess the specificity of such changes. .is finding has not been listed as an found that retinal arteries widened with increasing blood glucose and glycosylated haemoglobin and the progression observation indicator for DR at home or abroad, and only the relatively easily observed venous beading can be used as of DR. Islam et al. [15] suggested that retinal artery diameter was not associated with the progression of DR, which was one of the markers for venous abnormalities and non- proliferative DR. In this study, it was found that the RVOD inconsistent with the above findings. However, some re- and RVLD of retinal veins in DR patients were larger than searchers have found that the retinal artery of DR patients those in the NDR group, that WLR, WT, and WCSA in DR became narrower [16]. .erefore, the correlations between Journal of Healthcare Engineering 7 (a) (b) Figure 2: (a) .e border of the retinal artery wall (red arrow) clearly identified. (b) .e right retinal vein wall nearly overlaps with the retinal edge (blue arrow) and cannot be distinguished. diabetes and retinal artery remain controversial. .is is between these measures were significant. .ese data have perhaps due to the different races of the studied populations, important research value and expand the ways in which the the individual-level differences, and different sample sizes in relationship between DR and retinal vessel diameters can be different studies. studied. In patients with T2DM, fasting blood glucose level is a Wong et al. [19] showed that, in patients with T2DM, factor influencing the retinal vein diameter. When the long-term stimulation to the vascular wall by factors such as hyperglycaemia, lactate accumulation, insulin resistance, fasting blood glucose level increases, the retinal vein di- ameter shows a tendency to dilate; however, the fasting and haemodynamic changes caused vascular wall remod- elling. According to whether WCSA is enlarged or not, the blood glucose level does not affect the retinal artery di- ameter. .is may be related to the structure and plasticity of remodelling can be divided into inward eutrophic remod- the retinal artery and vein as well as the sensitivity of elling and inward hypertrophic remodelling [20–23]. In vitro measurement methods. Clinically, retinal vein dilation is experiments by Endemann et al. [24] showed that eutrophic more likely to be observed when fasting blood glucose is remodelling is a common form of structural change in re- elevated. In patients with T2DM, blood pressure influences sistance vessels in patients with essential hypertension, while the retinal artery diameter, and the retinal artery diameter hypertrophic remodelling is common in hypertensive pa- tends to narrow with increasing blood pressure. Some re- tients with diabetes. In this study, the experimental results searchers have given antihypertensive treatment to diabetic obtained through noninvasive SD-OCT on humans are consistent with the results of those in vitro experiments. patients with hypertension to control their blood pressure to a good level and have found that the diameter of the retinal .is study has several limitations. First, because the retinal vessels are close to the surface of the retina, it is difficult to artery will gradually increase after treatment. .e above conclusions suggest a strong correlation between retinal distinguish the upper wall of the blood vessel from the tissue, artery diameter and blood pressure. resulting in a larger outer diameter of the blood vessel and a In clinical practice, retinal artery stenosis is generally thicker vessel wall (Figure 2). Second, the FWHM method can considered an early indicator of vascular damage due to age, reduce the measurement error. However, the method is not hypertension, and other cardiovascular diseases, and fully automated; there is a possibility of error, and the steps are structural changes to the retinal artery can provide an early very cumbersome and inefficient. .ird,our samplesize is small warning about the health or disease of microvessels. In other compared to other large-sample epidemiological studies. words, if retinal arteriolar stenosis occurs in patients with Fourth, dynamic changes occur in retinal blood vessel mor- phology, requiring long-term follow-up. At present, we are T2DM, it may indicate high blood pressure, progression of diabetes, or vascular damage, suggesting the possibility of jointly developing retinal blood vessel measurement software based on deep learning, which can increase the accuracy of damage caused by complications such as DR and diabetic nephropathy [17, 18]. Strict blood pressure control can blood vessel mass measurement, thus improving the database. significantly reduce the macrovascular and microvascular complications in patients with T2DM. Moreover, it is easier 5. Conclusion to control the blood pressure stably than to control the blood glucose stably, and it is more effective to prevent and treat .e LDs and ODs of retinal arteries and veins in patients chronic complications of T2DM by controlling blood with DR are widened, the vessel wall is thickened, and the pressure than by controlling blood glucose. .e conclusions cross-sectional area of the vessel wall is increased. .e LDs drawn by recent studies on the changes in retinal artery and ODs of retinal veins and vessel WT in patients with DR diameter in diabetic patients are still controversial, and a are linearly correlated with DR stage and gradually widen larger sample size and further investigation are needed. with the progression of DR. Changes in the diameters of However, we provide some new arterial parameters, such as retinal arteries and veins in patients with DR are not cor- WTand WCSA, in this study, and the intergroup differences related with the duration of diabetes. 8 Journal of Healthcare Engineering [6] T. Y. Wong, M. D. Knudtson, R. Klein, B. E. Klein, Ethical Approval S. M. Meuer, and L. D. Hubbard, “Computer-assisted mea- surement of retinal vessel diameters in the Beaver Dam Eye .e Ethics Committee of Quzhou People’s Hospital, Zhe- Study: methodology, correlation between eyes, and effect of jiang, China, approved the study. Image acquisition, pro- refractive errors,” Ophthalmology, vol. 111, no. 6, cessing, and analysis were performed according to the tenets pp. 1183–1190, 2004. of the Declaration of Helsinki. [7] H. E. Moss, G. Treadwell, J. Wanek, S. DeLeon, and M. Shahidi, “Retinal vessel diameter assessment in papil- Consent ledema by semi-automated analysis of SLO images: feasibility and reliability,” Investigative Ophthalmology & Visual Science, Before study enrolment, a consent form was obtained from vol. 55, no. 4, pp. 2049–2054, 2014. all participants. [8] M. A. Merkx, J. O. Bescos, ´ L. Geerts, E. M. H. Bosboom, F. N. van de Vosse, and M. Breeuwer, “Accuracy and precision of vessel area assessment: manual versus automatic lumen Disclosure delineation based on full-width at half-maximum,” Journal of Magnetic Resonance Imaging, vol. 36, no. 5, pp. 1186–1193, .is preprint has been presented on Research Square https:// www.researchsquare.com/article/rs-579300/v1. [9] Y. H. Tong, T. P. Zhu, Z. L. Zhao, H. J. Zhan, F. Z. Jiang, and H. L. Lian, “Retinal Arteriolar Morphometry based on full Conflicts of Interest width at half Maximum Analysis of Spectral-Domain optical Coherence Tomography images,” PLoS One, vol. 10, no. 12, .e authors declare that they have no conflicts of interest. Article ID e0144437, 2015. [10] R. Klein, B. E. Klein, M. D. Knudtson, T. Y. Wong, and Authors’ Contributions M. Y. Tsai, “Are inflammatory factors related to retinal vessel caliber? .e Beaver Dam Eye Study,” Archives of Ophthal- .e first two authors contributed equally to this study. B. L. mology, vol. 124, no. 1, pp. 87–94, 2006. Xu and W. L. Zhou designed the study and the experiments [11] A. Mandecka, J. Dawczynski, M. Blum et al., “Influence of flickering light on the retinal vessels in diabetic patients,” and drafted the manuscript. Y. H. Tong revised the study Diabetes Care, vol. 30, no. 12, pp. 3048–3052, 2007. protocol. B. L. Xu, Y. J. Li, and W. L. Zhou participated in the [12] T. T. Nguyen, N. Cheung, and T. Y. Wong, “Influence of acquisition, analysis, and interpretation of data. H. J. Zhan flickering light on the retinal vessels in diabetic patients: and J. Y. Lu obtained the SD-OCTresults of the patients. All response to Mandecka et al,” Diabetes Care, vol. 31, no. 6, authors have read and approved the final manuscript. p. e51, 2008. [13] E. Saldıvar, P. Cabrales, A. G. Tsai, and M. Intaglietta, “Mi- Acknowledgments crocirculatory changes during chronic adaptation to hyp- oxia,” American Journal of Physiology - Heart and Circulatory .is study was supported by funding from Zhejiang Prov- Physiology, vol. 285, no. 5, pp. H2064–H2071, 2003. ince Public Welfare Technology Application Research [14] N. Cheung, S. L. Rogers, K. C. Donaghue, A. J. Jenkins, Project (GF18H12003). G. Tikellis, and T. Y. Wong, “Retinal arteriolar dilation predicts retinopathy in adolescents with type 1 diabetes,” Diabetes Care, vol. 31, no. 9, pp. 1842–1846, 2008. References [15] F. M. A. Islam, T. T. Nguyen, J. J. Wang et al., “Quantitative retinal vascular calibre changes in diabetes and retinopathy: [1] M. Ritt, J. M. Harazny, C. Ott et al., “Analysis of retinal the Singapore Malay eye study,” Eye, vol. 23, no. 8, arteriolar structure in never-treated patients with essential pp. 1719–1724, 2009. hypertension,” Journal of Hypertension, vol. 26, no. 7, [16] X. Wang and D. Liu, “[Retinal vessel diameter variation pp. 1427–1434, 2008. analysis in diabetic patients],” Zhonghua Yan Ke Za Zhi, [2] C. Sun, J. J. Wang, D. A. Mackey, and T. Y. Wong, “Retinal vol. 52, no. 5, pp. 358–361, 2016. vascular caliber: systemic, environmental, and genetic asso- [17] Z. Z. Wang, Y. P. Wang, and Y. Y. Cheng, “.e occurrence ciations,” Survey of Ophthalmology, vol. 54, no. 1, pp. 74–95, and prevention for diabetic retinopathy in community-based survey,” Chinese Medicine, vol. 05, no. 7, pp. 607-608, 2010. [3] C. Y. Cheung, D. Xu, C. Y. Cheng et al., “A deep-learning [18] Y. L. Guo, “Color Doppler flow imaging analysis of ocular system for the assessment of cardiovascular disease risk via the hemodynamics in patients with diabetic retinopathy,” Chin. measurement of retinal-vessel calibre,” Nature Biomedical J. Prim. Med. Pharm.vol. 13, no. 5, pp. 826-827, 2006. Engineering, vol. 5, no. 6, pp. 498–508, 2020. [19] T. Y. Wong, A. Kamineni, R. Klein et al., “Quantitative retinal [4] M. Bhaskaranand, C. Ramachandra, S. Bhat et al., “.e Value of Automated Diabetic Retinopathy Screening with the venular caliber and risk of cardiovascular disease in older persons: the cardiovascular health study,” Archives of Internal EyeArt System: A Study of More .an 100, 000 Consecutive Encounters from People with Diabetes,” Diabetes Technology Medicine, vol. 166, no. 21, pp. 2388–2394, 2006. [20] A. M. Heagerty, C. Aalkjaer, S. J. Bund, N. Korsgaard, and & erapeutics, vol. 21, no. 11, pp. 635–643, 2019. [5] M. Ruggeri, H. Wehbe, S. Jiao et al., “In vivo three-dimen- M. J. Mulvany, “Small artery structure in hypertension. Dual processes of remodeling and growth,” Hypertension, vol. 21, sional high-resolution imaging of rodent retina with spectral- domain optical coherence tomography,” Investigative Oph- no. 4, pp. 391–397, 1993. [21] E. L. Schiffrin, L. Y. Deng, and P. Larochelle, “Morphology of thalmology & Visual Science, vol. 48, no. 4, pp. 1808–1814, 2007. resistance arteries and comparison of effects of Journal of Healthcare Engineering 9 vasoconstrictors in mild essential hypertensive patients,” Clinical and Investigative Medicine, vol.16, no. 3, pp.177–186, [22] M. J. Mulvany, G. L. Baumbach, C. Aalkjaer et al., “Vascular remodeling,” Hypertension, vol. 28, no. 3, pp. 505-506, 1996. [23] E. L. Schiffrin, “Remodeling of resistance arteries in essential hypertension and effects of antihypertensive treatment,” American Journal of Hypertension, vol. 17, no. 12, pp. 1192– 1200, 2004. [24] D. H. Endemann, Q. Pu, C. De Ciuceis et al., “Persistent remodeling of resistance arteries in type 2 diabetic patients on antihypertensive treatment,” Hypertension, vol. 43, no. 2, pp. 399–404, 2004.
Journal of Healthcare Engineering – Hindawi Publishing Corporation
Published: Aug 8, 2022
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.