Predictive visual field outcomes after optic chiasm decompressive surgery by retinal vessels parameters using optical coherence tomography angiography

AIM:To evaluate the predictive value of superficial retinal capillary plexus(SRCP)and radial peripapillary capillary(RPC)for visual field recovery after optic cross decompression and compare them with peripapillary nerve fiber layer(pRNFL)and ganglion cell complex(GCC).METHODS:This prospective longitudinal observational study included patients with chiasmal compression due to sellar region mass scheduled for decompressive surgery.Generalized estimating equations were used to compare retinal vessel density and retinal layer thickness preand post-operatively and with healthy controls.Logistic regression models were used to assess the relationship between preoperative GCC,pRNFL,SRCP,and RPC parameters and visual field recovery after surgery.RESULTS:The study included 43 eyes of 24 patients and 48 eyes of 24 healthy controls.Preoperative RPC and SRCP vessel density and pRNFL and GCC thickness were lower than healthy controls and higher than postoperative values.The best predictive GCC and pRNFL models were based on the superior GCC[area under the curve(AUC)=0.866]and the tempo-inferior pRNFL(AUC=0.824),and the best predictive SRCP and RPC models were based on the nasal SRCP(AUC=0.718)and tempo-inferior RPC(AUC=0.825).There was no statistical difference in the predictive value of the superior GCC,tempo-inferior pRNFL,and tempo-inferior RPC(all P>0.05).CONCLUSION:Compression of the optic chiasm by tumors in the saddle area can reduce retinal thickness and blood perfusion.This reduction persists despite the recovery of the visual field after decompression surgery.GCC,pRNFL,and RPC can be used as sensitive predictors of visual field recovery after decompression surgery.

A Deformable Network with Attention Mechanism for Retinal Vessel Segmentation

The intensive application of deep learning in medical image processing has facilitated the advancement of automatic retinal vessel segmentation research.To overcome the limitation that traditional U-shaped vessel segmentation networks fail to extract features in fundus image sufficiently,we propose a novel network(DSeU-net)based on deformable convolution and squeeze excitation residual module.The deformable convolution is utilized to dynamically adjust the receptive field for the feature extraction of retinal vessel.And the squeeze excitation residual module is used to scale the weights of the low-level features so that the network learns the complex relationships of the different feature layers efficiently.We validate the DSeU-net on three public retinal vessel segmentation datasets including DRIVE,CHASEDB1,and STARE,and the experimental results demonstrate the satisfactory segmentation performance of the network.

Quantifying peripapillary vessel density and retinal nerve fibre layer in type 1 diabetic children without clinically detectable retinopathy using OCTA

AIM:To quantify changes in radial peripapillary capillary vessel density(ppVD)and the peripapillary retinal nerve fiber layer(pRNFL)in children with type 1 diabetes without clinical diabetic retinopathy by optical coherence tomography angiography(OCTA),providing a basis for early retinopathy in children with type 1 diabetes.METHODS:This was a retrospective study.A total of 30 patients(3–14y)with type 1 diabetes without clinical diabetic retinopathy(NDR group)were included.A total of 30 age-matched healthy subjects were included as the normal control group(CON group).The HbA1c level in the last 3mo was measured once in the NDR group.The pRNFL thickness and ppVD were automatically measured,and the mean pRNFL and ppVD were calculated in the nasal,inferior,temporal,and superior quadrants.The changes in ppVD and pRNFL in the two groups were analyzed.RESULTS:Compared with CON group,the nasal and superior ppVDs decreased in the NDR group(all P<0.01).The thickness of the nasal pRNFL decreased significantly(P<0.01),while the inferior,temporal and superior pRNFLs slightly decreased but not significant in the NDR group(all P>0.05).Person and Spearman correlation analysis of ppVD and pRNFL thickness in each quadrant of the NDR group showed a positive correlation between nasal and superior(all P<0.01),while inferior and temporal had no significant correlation(all P>0.05).There was no significant correlation between the HbA1c level and ppVD and pRNFL in any quadrant(all P>0.05).There was no significant correlation between the course of diabetes mellitus and ppVD and pRNFL in any quadrant(all P>0.05).CONCLUSION:ppVD and pRNFL decrease in eyes of children with type 1 diabetes before clinically detectable retinopathy and OCTA is helpful for early monitoring.

OCTA characteristics in non-arteritic central retinal artery occlusion and correlation with visual acuity

AIM:To observe the retinal and choroidal circulations in patients with non-arteritic permanent central retinal artery occlusion(NA-CRAO)via optical coherence tomography angiography(OCTA)and analyze their correlation with visual acuity.METHODS:Sixty-two eyes with clinically confirmed acute NA-CRAO were included in the study and divided into:A type(mild n=29),B type(moderate n=27)and C type(severe n=6)based on the degree of visual loss,retinal edema,and arterial blood flow delay in fundus fluorescence angiography(FFA).Contralateral healthy eyes were used as the control group.Best-corrected visual acuity(BCVA),slit lamp microscopy,indirect ophthalmoscopy,fundus color photography,OCTA,and FFA were performed.Spearman’s correlation analysis was used to determine the correlations between retinal and choroidal vessels and visual acuity.RESULTS:There were no statistically significant differences in age,gender,and intraocular pressure among the three types and the control group(P>0.05).Vessel density in deep capillary plexus(VD-DCP)significantly decreased(P<0.05)in all three types of NA-CRAO patients compared to the control group.Vessel density in superficial vascular plexus(VD-SVP)significantly decreased(P<0.05)in type A patients and choriocapillaris flow area significantly decreased(P<0.05)in type B and type C patients compared to the control group;while outer retinal flow areas significantly increased in the type A(P<0.05)and decreased in type C patients(P<0.05).The retinal thickness significantly increased in type C group(P<0.05).The VD-SVP at fovea in the type A was significantly lower than both of type B and C.The VD-SVP at nasal parafovea in type A and B was significantly lower than type C(P<0.05).The logMAR BCVA of type A was significantly better than that of type B and C groups(P<0.05).Spearman’s correlation analysis showed that the logMAR BCVA was positively correlated with VD-SVP at fovea(r=0.679,P=0.031)and nasal parafovea(r=0.826,P=0.013).CONCLUSION:OCTA is valuable for assessing retinal ischemia,and evaluating visual impairment.Deep retinal vasculature is commonly affected in all NA-CRAO types.VDSVPs at fovea and nasal parafovea can serve as reliable markers of visual impairment in NA-CRAO.

SepFE:Separable Fusion Enhanced Network for Retinal Vessel Segmentation

The accurate and automatic segmentation of retinal vessels fromfundus images is critical for the early diagnosis and prevention ofmany eye diseases,such as diabetic retinopathy(DR).Existing retinal vessel segmentation approaches based on convolutional neural networks(CNNs)have achieved remarkable effectiveness.Here,we extend a retinal vessel segmentation model with low complexity and high performance based on U-Net,which is one of the most popular architectures.In view of the excellent work of depth-wise separable convolution,we introduce it to replace the standard convolutional layer.The complexity of the proposed model is reduced by decreasing the number of parameters and calculations required for themodel.To ensure performance while lowering redundant parameters,we integrate the pre-trained MobileNet V2 into the encoder.Then,a feature fusion residual module(FFRM)is designed to facilitate complementary strengths by enhancing the effective fusion between adjacent levels,which alleviates extraneous clutter introduced by direct fusion.Finally,we provide detailed comparisons between the proposed SepFE and U-Net in three retinal image mainstream datasets(DRIVE,STARE,and CHASEDB1).The results show that the number of SepFE parameters is only 3%of U-Net,the Flops are only 8%of U-Net,and better segmentation performance is obtained.The superiority of SepFE is further demonstrated through comparisons with other advanced methods.

DT-Net:Joint Dual-Input Transformer and CNN for Retinal Vessel Segmentation

Retinal vessel segmentation in fundus images plays an essential role in the screening,diagnosis,and treatment of many diseases.The acquired fundus images generally have the following problems:uneven illumination,high noise,and complex structure.It makes vessel segmentation very challenging.Previous methods of retinal vascular segmentation mainly use convolutional neural networks on U Network(U-Net)models,and they have many limitations and shortcomings,such as the loss of microvascular details at the end of the vessels.We address the limitations of convolution by introducing the transformer into retinal vessel segmentation.Therefore,we propose a hybrid method for retinal vessel segmentation based on modulated deformable convolution and the transformer,named DT-Net.Firstly,multi-scale image features are extracted by deformable convolution and multi-head selfattention(MHSA).Secondly,image information is recovered,and vessel morphology is refined by the proposed transformer decoder block.Finally,the local prediction results are obtained by the side output layer.The accuracy of the vessel segmentation is improved by the hybrid loss function.Experimental results show that our method obtains good segmentation performance on Specificity(SP),Sensitivity(SE),Accuracy(ACC),Curve(AUC),and F1-score on three publicly available fundus datasets such as DRIVE,STARE,and CHASE_DB1.

Intelligent Machine Learning Enabled Retinal Blood Vessel Segmentation and Classification

Automated segmentation of blood vessels in retinal fundus images is essential for medical image analysis.The segmentation of retinal vessels is assumed to be essential to the progress of the decision support system for initial analysis and treatment of retinal disease.This article develops a new Grasshopper Optimization with Fuzzy Edge Detection based Retinal Blood Vessel Segmentation and Classification(GOFED-RBVSC)model.The proposed GOFED-RBVSC model initially employs contrast enhancement process.Besides,GOAFED approach is employed to detect the edges in the retinal fundus images in which the use of GOA adjusts the membership functions.The ORB(Oriented FAST and Rotated BRIEF)feature extractor is exploited to generate feature vectors.Finally,Improved Conditional Variational Auto Encoder(ICAVE)is utilized for retinal image classification,shows the novelty of the work.The performance validation of the GOFEDRBVSC model is tested using benchmark dataset,and the comparative study highlighted the betterment of the GOFED-RBVSC model over the recent approaches.

Identification of retinal thickness and blood flow in agerelated macular degeneration with reticular pseudodrusen

AIM:To investigate thickness characteristics and vascular plexuses in retinas with reticular pseudodrusen(RPD)as an early detection strategy for age-related macular degeneration(AMD).METHODS:This retrospective study included 24 subjects(33 eyes)with RPD and 25 heathy control subjects(34 eyes).The superficial capillary plexus(SCP)and the deep capillary plexus(DCP)of the retinal posterior poles were investigated with optical coherence tomography angiography(OCTA).Retinal thicknesses and vessel densities were analyzed statistically.RESULTS:The general retinal thicknesses of RPD eyes were significantly decreased(95%CI-14.080,-0.655;P=0.032).The vessel densities of DCP in RPD eyes were significantly increased in the global(95%CI 1.067,7.312;P=0.027),parafoveal(95%CI 0.417,5.241;P=0.022),and perifoveal(95%CI 0.181,6.842;P=0.039)quadrants.However,the vessel densities of the SCP were rarely increased in the eyes with RPD.CONCLUSION:The thinning of retinas in the RPD group suggests a reduction in the number of cells.Additionally,the increased vessel density of the DCP in retinas with RPD indicates a greater demand for blood supply,possibly due to the hypoxia induced RPD compensation caused by RPD in the outer retina.This study highlights the pathological risks associated with RPD and emphasizes the importance of early intervention to retard the progression of AMD.

Analysis of retinal arteriolar and venular parameters in primary open angle glaucoma

AIM: To measure the retinal vessels of primary open angle glaucoma(POAG) patients on spectral domain optical coherence tomography(SD-OCT) with a full-width at half-maximum(FWHM) algorithm to better explore their structural changes in the pathogenesis of POAG.METHODS: In this retrospective case-control study, the right eyes of 32 patients with POAG and 30 healthy individuals were routinely selected.Images of the supratemporal and infratemporal retinal vessels in the B zones were obtained by SD-OCT, and the edges of the vessels were identified by the FWHM method.The internal and external diameters, wall thickness(WT), wall cross-sectional area(WCSA) and wall-to-lumen ratio(WLR) of the blood vessels were studied.RESULTS: Compared with the healthy control group, the POAG group showed a significantly reduced retinal arteriolar outer diameter(RAOD), retinal arteriolar lumen diameter(RALD) and WSCA in the supratemporal(124.22±12.42 vs 138.32±10.73 μm, 96.09±11.09 vs 108.53±9.89 μm,and 4762.02 ± 913.51 vs 5785.75 ± 114 8.28 μm^(2), respectively, all P<0.05) and infratemporal regions(125.01±15.55 vs 141.57±10.77 μm, 96.27±13.29 vs 110.83 ± 10.99 μm, and 4925.56 ± 1302.88 vs 6087.78±1061.55 μm^(2), all P<0.05).The arteriolar WT and WLR were not significantly different between the POAG and control groups, nor were the retinal venular outer diameter(RVOD), retinal venular lumen diameter(RVLD) or venular WT in the supratemporal or infratemporal region.There was a positive correlation between the arteriolar parameters and visual function.CONCLUSION: In POAG, narrowing of the supratemporal and infratemporal arterioles and a significant reduction in the WSCA is observed, while the arteriolar WT and WLR do not change.Among the venular parameters, the external diameter, internal diameter, WT, WLR, and WSCA of the venules are not affected.

Effects of two different doses of intravitreal bevacizumab on subfoveal choroidal thickness and retinal vessel diameter in branch retinal vein occlusion

AIM： To investigate the effects of two different doses of intravitreal bevacizumab on subfoveal choroidal thickness （SFChT） and retinal vessel diameter in patients with branch retinal vein occlusion. METHODS： An interventional, restrospective study of 41 eyes of 41 patients who had completed 12mo of follow-up, divided into group 1 （1.25 mg of bevacizumab, 21 eyes of 21 patients） and group 2 （2.5 mg of bevacizumab, 20 eyes of 21 patients）. Complete ophthalmic examination, fluorescein angiography, enhanced depth imaging optical coherence tomography and measurement of retinal vessel diameter with IVAN software were performed at baseline and follow-up. RESULTS： The SFChT changed from 279.1 （165-431） μm at baseline to 277.0 （149-413） μm at 12mo in group 1 （P= 0.086）, and from 301.4 （212-483） μm to 300.3 （199-514） μm in group 2 （P=0.076）. The central retinal arteriolar equivalent （CRAE） changed from 128.8 ±11.2 μm at baseline to 134.5±8.4 μm at 12mo in group 1, and from 134.6±9.0 μm to 131.4±12.7 μm in group 2 （P =0.767）. The central retinal venular equivalent （CRVE） changed from 204.1±24.4 μm at baseline to 196.3±28.2 μm at 12mo in group 1, and from 205.8±16.3 μm to 194.8±18.2 μm in group 2 （P=0.019）. The mean central macular thickness （P〈0.05） and average best-corrected visual acuity （BCVA; P〈0.05） improved in both groups CONCLUSION： Changes in the SFChT are not statistically significant and not different according to the doses of bevacizumab. The CRAE did not show significant change, however, the CRVE showed significant decrease regardless of the dose.

Characterization of human retinal vessel arborisation in normal and amblyopic eyes using multifractal analysis

AIMTo characterize the human retinal vessel arborisation in normal and amblyopic eyes using multifractal geometry and lacunarity parameters.METHODSMultifractal analysis using a box counting algorithm was carried out for a set of 12 segmented and skeletonized human retinal images, corresponding to both normal (6 images) and amblyopia states of the retina (6 images).RESULTSIt was found that the microvascular geometry of the human retina network represents geometrical multifractals, characterized through subsets of regions having different scaling properties that are not evident in the fractal analysis. Multifractal analysis of the amblyopia images (segmented and skeletonized versions) show a higher average of the generalized dimensions (Dq) for q=0, 1, 2 indicating a higher degree of the tree-dimensional complexity associated with the human retinal microvasculature network whereas images of healthy subjects show a lower value of generalized dimensions indicating normal complexity of biostructure. On the other hand, the lacunarity analysis of the amblyopia images (segmented and skeletonized versions) show a lower average of the lacunarity parameter Λ than the corresponding values for normal images (segmented and skeletonized versions).CONCLUSIONThe multifractal and lacunarity analysis may be used as a non-invasive predictive complementary tool to distinguish amblyopic subjects from healthy subjects and hence this technique could be used for an early diagnosis of patients with amblyopia.

Retinal microvascular and microstructural alterations in the diagnosis of meibomian gland dysfunction in severely obese population:a new approach

AIM:To study retinal microvascular and microstructural alterations in meibomian gland dysfunction(MGD)in severely obese population using optical coherence tomography angiography(OCTA).METHODS:Twelve MGD patients with severely obese population(PAT group;24 eyes)and 12 healthy controls(HC group;24 eyes)were recruited.OCTA images were segmented into five[superior(S),nasal(N),inferior(I),temporal(T),and central foveal(C)]or nine[inner superior(IS),outer superior(OS),inner nasal(IN),outer nasal(ON),inner inferior(II),outer inferior(OI),inner temporal(IT),outer temporal(OT),and C]subregions.The superficial vessel density(SVD),retinal thickness(RT),foveal avascular zone(FAZ)parameters,and retinal volume were measured.RESULTS:Visual acuity was significantly different between two groups(0.8±0.17 in PAT group vs 0.2±0.06 in HC group).SVD was significantly lower in PATs in N,T,OS,IN,OT,and ON.The area under the receiver operating characteristic curve(AUC)for T was 0.961[95% confidence interval(CI):0.908 to 1.000],for OS was 0.962(95%CI:0.915 to 1.000).RT was significantly lower in PATs in IS,OS,OI,OT,ON,IT,IN,and II.AUC for OT was 0.935(95%CI:0.870 to 0.999),for IS was 0.915(95%CI:0.838 to 0.992).Angiography results showed significantly lower area and perimeter of FAZ,SVD of the inner retina and both retinal volume and the average volume thickness in the PAT group.CONCLUSION:Vision may be affected in patients with MGD due to changes in retinal microvessels and microstructures.These changes detected by OCTA may be a potential marker for diagnosing MGD in severe obesity.

Effects of obstructive sleep apnea on retinal microvasculature

AIM:To detect retinal microvascular variations in obstructive sleep apnea syndrome patients.METHODS:This prospective,observational case-control study included healthy controls and patients with mild,moderate,and severe obstructive sleep apnea syndrome.Vascular parameters,foveal avascular area,and flow areas in macula-centered,6.00×6.00 mm2 scan size optical coherence tomography angiography images were compared.RESULTS:The control group had the highest whole image,parafoveal,and perifoveal vessel density among the groups in both superficial and the deep capillary plexus(all P<0.05).Rapid eye movement sleep apnoea-hypopnoea index was reversely correlated with whole(Rho=-0.195,P=0.034),parafoveal(Rho=-0.242,P=0.008),perifoveal(Rho=-0.187,P=0.045)vessel density in the superficial capillary plexus,and whole(Rho=-0.186,P=0.046),parafoveal(Rho=-0.260,P=0.004),perifoveal(Rho=-0.189,P=0.043)vessel density in the deep capillary plexus,though the mean and non-rapid eye movement sleep apnoeahypopnoea index related with only parafoveal vessel density in the superficial capillary plexus(Rho=-0.213,P=0.020;Rho=-0.191,P=0.038)and the deep capillary plexus(Rho=-0.254,P=0.005;Rho=-0.194,P=0.035).CONCLUSION:This study shows decreased vessel density and its reverse correlation with the apnoea-hypopnoea index in patients with obstructive sleep apnea syndrome.

Peripapillary vessel density and retinal nerve fiber layer thickness changes in early diabetes retinopathy

·AIM:To perform a quantitative analysis of the peripapillary vessel density(VD)and retinal nerve fiber layer(RNFL)thickness changes in patients with early stage of diabetic retinopathy(DR).·METHODS:In this case-control study,swept-source optical coherence tomography angiography(SS-OCTA)imaging was used to examine diabetic and age-matched healthy subjects.The optic disc HD 6×6 mm^(2) blood flow imaging scan mode was selected.Automatic software was used to measure the peripapillary VD,capillary vessel density(CVD),and RNLF in an optic nerve head(ONH)filed based on the Garway-Heath map.In addition,the correlation between peripapillary VD,CVD,and RNFL was further investigated.·RESULTS:The cohort consisted of 32 healthy individuals and 72 patients with diabetes(34 eyes with no DR and 38 eyes with mild-moderate NPDR).Peripapillary VD decreased in the mild-moderate NPDR group compared to the control group in most regions(P<0.05).Peripapillary CVD and RNFL thickness were significantly lower in the mild-moderate NPDR group in the superior temporal(ST)quadrants(P=0.018.P=0.030).In the correlation analysis of each region,the RNFL thickness in the NS region was positively correlated with the peripapillary VD and CVD(r=0.233,P=0.05;r=0.288.P=0.015).In the TI region,the RNFL thickness was positively correlated with the peripapillary CVD(r=0.237,P=0.047).·CONCLUSION:The measurement based on the ONH topographic map may be helpful in detecting functional and structural impairments in DR.The peripapillary VD,CVD and RNFL decrease in early DR,and the RNFL thickness altered in association with the CVD or/and VD in some regions.

Retinal thickness and vascular parameters using optical coherence tomography in Alzheimer's disease:a meta-analysis

Examining the retinal tissue has the potential to provide a unique method and technique to quantify Alzheimer’s disease-related changes in participants at various stages of the disease.In this metaanalysis,we aimed to investigate the association of various optical coherence tomography parameters with Alzheimer’s disease and whether retinal measurements can be used to diffe rentiate between Alzheimer’s disease and control subjects.Scientific databases including Google Schola r,Web of Science,and PubMed were systematically searched for published articles that evaluated retinal nerve fiber layer thickness and retinal microvascular network in Alzheimer’s disease and control subjects.Seventy-three studies(5850 participants,including 2249 Alzheimer’s disease patients and 3601controls) were included in this meta-analysis.Relative to controls,Alzheimer’s disease patients had a significantly lower global retinal nerve fiber layer thickness(standardized mean difference [SMD]=-0.79,95% confidence intervals [CI]:-1.03 to-0.54,P <0.00001) as well as each quadrant being thinner in Alzheimer’s disease versus controls.Regarding macular paramete rs,values measured by optical coherence tomography were significantly lower in Alzheimer’s disease than controls for macular thickness(pooled SMD:-0.44,95% CI:-0.67 to-0.20,P=0.0003),foveal thickness(pooled SMD=-0.39,95% CI:-0.58 to-0.19,P <0.0001),ganglion cell inner plexiform layer(SMD=-1.26,95% CI:-2.24 to-0.27,P=0.01) and macular volume(pooled SMD=-0.41,95% CI-0.76 to-0.07,P=0.02).Analysis using optical coherence tomography angiography parameters revealed mixed results between Alzheimer’s disease and controls.Superficial vessel density(pooled SMD=-0.42,95% CI:-0.68 to-0.17,P=0.0001) and deep vessel density(pooled SMD=-0.46,95% CI:-0.75 to-0.18,P=0.001) were found to be thinner in Alzheimer’s disease patients whereas the foveal avascular zone(SMD=0.84,95% CI:0.17-1.51,P=0.01) was larger in controls.Vascular density and thickness of various retinal laye rs were decreased in Alzheimer’s disease patients compared to controls.Our results provide evidence for optical coherence tomography technology having the potential to detect retinal and microvascular changes in patients diagnosed with Alzheimer’s disease and aid in monito ring and early diagnosis methods.

MIA-UNet:Multi-Scale Iterative Aggregation U-Network for Retinal Vessel Segmentation

As an important part of the new generation of information technology,the Internet of Things(IoT)has been widely concerned and regarded as an enabling technology of the next generation of health care system.The fundus photography equipment is connected to the cloud platform through the IoT,so as to realize the realtime uploading of fundus images and the rapid issuance of diagnostic suggestions by artificial intelligence.At the same time,important security and privacy issues have emerged.The data uploaded to the cloud platform involves more personal attributes,health status and medical application data of patients.Once leaked,abused or improperly disclosed,personal information security will be violated.Therefore,it is important to address the security and privacy issues of massive medical and healthcare equipment connecting to the infrastructure of IoT healthcare and health systems.To meet this challenge,we propose MIA-UNet,a multi-scale iterative aggregation U-network,which aims to achieve accurate and efficient retinal vessel segmentation for ophthalmic auxiliary diagnosis while ensuring that the network has low computational complexity to adapt to mobile terminals.In this way,users do not need to upload the data to the cloud platform,and can analyze and process the fundus images on their own mobile terminals,thus eliminating the leakage of personal information.Specifically,the interconnection between encoder and decoder,as well as the internal connection between decoder subnetworks in classic U-Net are redefined and redesigned.Furthermore,we propose a hybrid loss function to smooth the gradient and deal with the imbalance between foreground and background.Compared with the UNet,the segmentation performance of the proposed network is significantly improved on the premise that the number of parameters is only increased by 2%.When applied to three publicly available datasets:DRIVE,STARE and CHASE DB1,the proposed network achieves the accuracy/F1-score of 96.33%/84.34%,97.12%/83.17%and 97.06%/84.10%,respectively.The experimental results show that the MIA-UNet is superior to the state-of-the-art methods.

Evaluation of macular vessel density changes after vitrectomy with silicone oil tamponade in patients with rhegmatogenous retinal detachment

AIM:To evaluate macular microvasculature changes in eyes after pars plana vitrectomy(PPV)and intraocular silicone oil(SO)tamponade for macula-off rhegmatogenous retinal detachment(RRD)using optical coherence tomography angiography(OCTA).METHODS:Totally 19 eyes(19 patients)with maculaoff RRD who underwent PPV and intraocular SO tamponade were retrospectively reviewed.The parafoveal superficial capillary plexus(SCP)vessel density(VD),deep capillary plexus(DCP)VD,choriocapillaris plexus(CCP)VD,and foveal macular thickness were evaluated using OCTA throughout 16 wk postoperatively.The values of healthy fellow eyes were used as control.RESULTS:The parafoveal SCP,DCP,and CCP VDs were significant increased over time in RRD eyes during the 12 wk postoperatively,then decreased at 16 wk postoperatively(all P<0.01).The ratios of RRD eyes and fellow healthy eyes(r/f ratios)of the SCP and DCP VDs were lower than those of the CCP VD postoperatively(all P<0.05).There were not significant differences in the r/f ratios between SCP and DCP VDs postoperatively(all P>0.05).CONCLUSION:The parafoveal SCP,DCP,and CCP VDs gradually recover over time after PPV surgery with SO tamponade.Long-time SO tamponade might decrease postoperative macular VDs.Compared to parafoveal CCP VD,the parafoveal SCP and DCP VDs were more vulnerable in RRD eyes postoperatively.

Modified Anam-Net Based Lightweight Deep Learning Model for Retinal Vessel Segmentation

The accurate segmentation of retinal vessels is a challenging taskdue to the presence of various pathologies as well as the low-contrast ofthin vessels and non-uniform illumination. In recent years, encoder-decodernetworks have achieved outstanding performance in retinal vessel segmentation at the cost of high computational complexity. To address the aforementioned challenges and to reduce the computational complexity, we proposea lightweight convolutional neural network (CNN)-based encoder-decoderdeep learning model for accurate retinal vessels segmentation. The proposeddeep learning model consists of encoder-decoder architecture along withbottleneck layers that consist of depth-wise squeezing, followed by fullconvolution, and finally depth-wise stretching. The inspiration for the proposed model is taken from the recently developed Anam-Net model, whichwas tested on CT images for COVID-19 identification. For our lightweightmodel, we used a stack of two 3 × 3 convolution layers (without spatialpooling in between) instead of a single 3 × 3 convolution layer as proposedin Anam-Net to increase the receptive field and to reduce the trainableparameters. The proposed method includes fewer filters in all convolutionallayers than the original Anam-Net and does not have an increasing numberof filters for decreasing resolution. These modifications do not compromiseon the segmentation accuracy, but they do make the architecture significantlylighter in terms of the number of trainable parameters and computation time.The proposed architecture has comparatively fewer parameters (1.01M) thanAnam-Net (4.47M), U-Net (31.05M), SegNet (29.50M), and most of the otherrecent works. The proposed model does not require any problem-specificpre- or post-processing, nor does it rely on handcrafted features. In addition,the attribute of being efficient in terms of segmentation accuracy as well aslightweight makes the proposed method a suitable candidate to be used in thescreening platforms at the point of care. We evaluated our proposed modelon open-access datasets namely, DRIVE, STARE, and CHASE_DB. Theexperimental results show that the proposed model outperforms several stateof-the-art methods, such as U-Net and its variants, fully convolutional network (FCN), SegNet, CCNet, ResWNet, residual connection-based encoderdecoder network (RCED-Net), and scale-space approx. network (SSANet) in terms of {dice coefficient, sensitivity (SN), accuracy (ACC), and the areaunder the ROC curve (AUC)} with the scores of {0.8184, 0.8561, 0.9669, and0.9868} on the DRIVE dataset, the scores of {0.8233, 0.8581, 0.9726, and0.9901} on the STARE dataset, and the scores of {0.8138, 0.8604, 0.9752,and 0.9906} on the CHASE_DB dataset. Additionally, we perform crosstraining experiments on the DRIVE and STARE datasets. The result of thisexperiment indicates the generalization ability and robustness of the proposedmodel.

Changes in retinal vessel diameters in migraine patients during attack-free period

AIM：To evaluate the retinal vessel diameters in patients with migraine by optical coherence tomography（OCT）.METHODS：In this cross-sectional study,124 eyes of 62 patients with a diagnosis of unilateral migraine during attack-free period and 42 age-and sex-matched control subjects were included. Migraine patients were divided into the ≤2 migraine attacks per month group and the ≥5 migraine attacks per month group. All subjects underwent complete ophthalmological and neurological examinations before measurements. Retinal vessel diameters and choroidal thickness were examined with the Spectralis OCT.RESULTS：The mean diameters of the arteries in the eyes on the headache side of control group,≥5 migraine attacks per month and ≤2 migraine attacks per month group at 480 μm from the optic disk（Raster 3）were 119.54±46.69,136.68±25.93 and 119.34±31.75 μm respectively with a steady decline to 105.57±32.15,118.18±31.87 and 108.05±38.77 μm at 1440 μm（Raster 7）,the last measurement point,respectively. The retinal artery diameter measurements were significantly increased in ≥5 migraine attacks per month patients at four out of five measured points compared to control group（P〈0.05）. There were no statistical differences at any of the points of vein measurements. The choroidal thickness measurements were significantly decreased in ≥5 migraine attacks per month patients at all measured points compared to control group（P〈0.05）.CONCLUSION：The retinal artery diameter is found to increase significantly and the choroidal thickness is found to decrease in the eyes on the headache side in ≥5 migraine attacks per month patients compared to control group.

Changes of the peripapillary vascular parameters inpremature infants without retinopathy of prematurityusing U-net segmentation

AIM:To quantitatively assess the changes in mean vascular tortuosity(mVT)and mean vascular width(mVW)around the optic disc and their correlation with gestational age(GA)and birth weight(BW)in premature infants without retinopathy of prematurity(ROP).METHODS:A single-center retrospective study included a total of 133(133 eyes)premature infants[mean corrected gestational age(CGA)43.6wk]without ROP as the premature group and 130(130 eyes)CGA-matched fullterm infants as the control group.The peripapillary mVT and mVW were quantitatively measured using computerassisted techniques.RESULTS:Premature infants had significantly higher mVT(P=0.0032)and lower mVW(P=0.0086)by 2.68(10^(4) cm^(-3))and 1.85μm,respectively.Subgroup analysis with GA showed significant differences(P=0.0244)in mVT between the early preterm and middle to late preterm groups,but the differences between mVW were not significant(P=0.6652).The results of the multiple linear regression model showed a significant negative correlation between GA and BW with mVT after adjusting sex and CGA(P=0.0211 and P=0.0006,respectively).For each day increase in GA at birth,mVT decreased by 0.1281(10^(4) cm^(-3))and for each 1 g increase in BW,mVT decreased by 0.006(10^(4) cm^(-3)).However,GA(P=0.9402)and BW(P=0.7275)were not significantly correlated with mVW.CONCLUSION:Preterm birth significantly affects the peripapillary vascular parameters that indicate higher mVT and narrower mVW in premature infants without ROP.Alterations in these parameters may provide new insights into the pathogenesis of ocular vascular disease.