Choroidal Thickness And Choroidal Vessel Density In Nonexudative Age-related Macular Degeneration

Objective: To analyze the relationship between choroidal thickness and the distribution of large choroidal blood vessels in non-exudative AMD eyes with and with reticular pseudodrusen(RPD)using widefield swept source optical coherence tomography(SS-OCT)imaging.Methods:(1)A total of 209 patients(307 eyes)with the diagnosis of early,intermediate or late non-exudative AMD were prospectively enrolled in an SS-OCT imaging study at Bascom Palmer Eye institute between August 2013 and August 2014.On the same day,all participants underwent conventional multimodal imaging including color fundus,fundus autoflorescence(FAF)and infrared reflectance(IR)with a field of view between 30° and 35°,as well as imaging with a 100 kHz SS-OCT instrument with a central wavelength of 1,050 nm.Select all the eyes with the RPD and without RPD based on a consensus evaluation of whether RPD were present or absent based on conventional fundus images which included color,FAF,IR reflectance images,and en face OCT iamges generated using slabs with boundaries from 35 to 55 microns above the RPE.(2)We randomly chose 25 non-exudative AMD eyes without RPD(Group 1)and 25 non-exudative AMD eyes with RPD(Group 2)for this study.A fully automated algorithm was used to segment all scans for generating automated choroid thickness maps.Five eyes from each group were randomly selected for manual segmentation,in order to validate the performance of the choroidal segmentation algorithm.We defined an Early Treatment Diabetic Retinopathy Study(ETDRS)-like target centered on the fovea and divided the en face image into nine subfields(central,interior inferonasal,interior superonasal,interior superotemporal,interior inferotemporal,outer infernasal,outer superonasal,outer superotemporal,and outer inferotemporal).The average choroid thickness of all the subfields were measured on the automated choroid thickness maps.(3)We used the automated segmentation algorithm generated the en face OCT choroidal slabs for the 50 eyes in both groups.Its upper boundary lay 40 ?m below Bruch's membrane,and its lower boundary was the choroid/sclera interface.Using intensity thresholding in a graphics software,one well-trained researcher generated binary maps representing the distribution of the large choroid vessels.The same ETDRS-like target were used to quantify choroid vessel density,which were computed from the binary vessel maps as the number of black pixel(vessels)in a given region divided the total number of pixels in that region.Results:(1)The average difference between the manual segmented and automated segmented choroid thickness maps was 13.7 ?m.Although large variance of choroidal thickness parterns was found from eye to eye or from the different regions of each eye,choroidal thickness was typically thinner in non-exudative AMD eyes with RPD(Group 2)compared with non-exudative eyes without RPD(Group 1)(213.45?m VS.255.60?m,P<0.001).Except in outer inferotemporal subfield(217.36?m VS.237.62?m,P=0.157),in all other eight subfields,the choroidal thickness of Group 2 was significant thinner than that of Group 1(center: 236.57?m VS.292.46?m,P=0.007;interior inferonasal: 195.96 ? m VS.242.29 ? m,P=0.008;interior superonasal: 207.91?m VS.264.75?m,P=0.003;interior superortemporal: 244.63?m VS.296.45?m,P=0.003;interior inferotemporal: 225.00?m VS.265.48?m,P=0.018;outer inferotemporal : 160.59 ? m VS.185.54 ? m,P=0.023;outer superonasal: 182.91?m VS.225.58?m,P<0.001;outer superotemporal: 250.14?m VS.290.23?m,P=0.003).(2)Generally,the average choroidal vessel density was significantly lower in the eyes with RPD(Group1: 55% VS.Group 2: 45%,P<0.001).Only in the center(53% VS.73%,P<0.001),interior superonasal(40% VS.58%,P=0.002),and outer superonasal(44% VS.52%,P=0.01),the mean values of large choroidal vessel density of Group 2 were significantly smaller than Group 1.In the other six subfileds – interior inferonasal(36% VS.48%,P=0.05),interior superotemporal(55% VS.64%,P=0.06),interior inferotemporal(44% VS.51%,P=0.29),outer inferonasal(31% VS.32%,P=0.75),outer superotemporal(59% VS.67%,P=0.10),outer inferonasal(44% VS.44%,P=0.97)the average choroidal thickness of Group 2 was lower than Group 1,but the difference didn't reach statistical significance.There was a significant and stable correlation between choroidal thickness and large choroidal vessel density across two groups,overall in the target grid(Group 1: r=0.77,Group 2: r=0.70)and in the central subfield(Group 1: r=0.75,Group 2: r=0.80).There was no significant difference between the slope of the linear regressions in the two groups(P =0.66).Conclusions:(1)Our fully automated choroidal segmentation algorithm was able to capture the different patterns of choroidal thickness over a wide area.(2)Large variabilities were found inside each choroidal thickness map and in the choroidal thickness maps of different eyes.(3)In spite of both average choroidal thickness and vessel density were significantly lower in eyes with RPD,choroidal thickness has a clear relationship with the density of large choroid vessels in our sample,irrespective of the presence or absence of RPD.