共聚焦激光后照法眼底成像对不同类型糖尿病黄斑水肿的检出分析

Analysis of the positive rate of scanning laser ophthalmoscope in the retromode for different types of diabetic macular edema

目的观察共聚焦激光后照法(RM-SLO)眼底成像对不同类型糖尿病黄斑水肿(DME)的检出率,初步分析其与黄斑中心凹厚度(CMT)、黄斑体积的相关性。方法2021年3~5月于温州医科大学附属眼视光医院杭州院区检查确诊的DME患者40例65只眼纳入研究。所有患眼均行最佳矫正视力(BCVA)、光相干断层扫描(OCT)、RM-SLO眼底成像检查;同时行荧光素眼底血管造影(FFA)检查47只眼。采用超广角激光眼底成像系统Mirante SLO行RM-SLO眼底成像检查,包括右侧、左侧后照法(RMDR、RMDL)眼底成像。RMDR、RMDL眼底成像中一项及以上能识别DME则定义为RM-SLO能识别DME。采用OCT设备自带软件测量CMT、距中心凹6 mm处黄斑体积。根据DME在OCT、FFA图像上的特征分别分为弥漫性视网膜增厚(DRT)型、黄斑囊样水肿(CME)型、浆液性视网膜脱离(SRD)型和局灶性渗漏型、弥漫性渗漏型、弥漫性囊性渗漏型。观察RMDR、RMDL眼底成像在不同分型DME中的检出情况以及检出DME的一致性。能否被RM-SLO识别DME者之间CMT、黄斑体积比较采用非参数检验。RM-SLO检出情况与DME类型、CMT、黄斑体积的相关性以及BCVA与DME类型、CMT、黄斑体积相关性行Spearman相关性分析。结果65只眼中,RMDR、RMDL眼底成像检出DME分别为46(70.77%,46/65)、48(73.85%,48/65)只眼;RMDR、RMDL眼底成像对DME检出的一致性较好(Kappa值=0.770,P<0.001)。RMDR、RMDL眼底成像对DRT型、CME型、SRD型的检出率分别为42.11%(8/19)、57.89%(11/19),77.78%(28/36)、77.78%(28/36)和100.00%(10/10)、90.00%(9/10);对局灶性渗漏型、弥漫性渗漏型、弥漫性囊性渗漏型的检出率分别为68.75%(11/16)、62.50%(10/16),68.00%(17/25)、76.00%(19/25)和100.00%(6/6)、100.00%(6/6)。Spearman相关性分析结果显示,RM-SLO能否识别DME与CMT、OCT分型呈显著相关(r=0.310、0.365,P=0.120、0.003);与FFA分型、黄斑体积无相关性(r=0.113、0.117,P=0.449、0.352)。BCVA与CMT及黄斑体积呈显著相关(r=0.307、0.269,P=0.013、0.030);与OCT、FFA分型均无相关性(r=0.051、0.175,P=0.684、0.240)。结论RM-SLO眼底成像中RMDR、RMDL诊断DME的一致性较好;对DRT型及CMT较小的DME不易识别。

Objective:To explore the positive rate of scanning laser ophthalmoscope in the retromode(RM-SLO)in different types of diabetic macular edema(DME),and to analyze its correlation with foveal thickness(CMT)and macular volume.Methods:From March to May 2021,40 patients(65 eyes)were diagnosed as DME by fundus examination combined with optical coherence tomography(OCT)in Affiliated Eye Hospital of Wenzhou Medical University at Hangzhou were included in the study.All eyes underwent best corrected visual acuity(BCVA),OCT and RM-SLO fundus imaging examinations,47 eyes underwent fluorescein fundus angiography(FFA)examination.RM-SLO fundus imaging examinations were performed with Mirante SLO,including retro mode illumination deviated right(RMDR)and retro mode illumination deviated left(RMDL).If one or more of the RMDR and RMDL of the examined patient can identify macular edema,RM-SLO was considered to be able to identify macular edema.The macular volume at CMT and 6 mm from the fovea was measured by OCT software.DME were divided into 3 types based on OCT images:diffuse retinal thinkening(DRT)type;cystoid macular edema(CME)type;serous retinal detachment(SRD)type,focal leakage type,diffuse leakage type and diffuse cystic leakage type.The consistency of RMDR and RMDL in the diagnosis of DME in RM-SLO fundus imaging was evaluated,as well as their positive rate in different classifications of DME.The correlation between the detection of macular edema by RM-SLO and the DME type,CMT and foveal volume,and the correlation between BCVA and edema type,CMT and macular volume were analyzed.Results:Among 65 eyes,the positive rates of RMDR and RMDL fundus imaging to detect DME were 46(70.77%,46/65)and 48(73.85%,48/65),respectively.There was good consistency in identifying DME(Kappa value=0.770;P<0.001).The positive rates of RMDR and RMDL fundus imaging DRT,CME and SRD type of DME were 42.11%(8/19),57.89%(11/19),77.78%(28/36),77.78%(28/36),100.00%(10/10),90.00%(9/10),respectively.In the FFA classification of them,the positive rates of focal leakage,diffuse leakage and diffuse cystic leakage were 68.75%(11/16),62.50%(10/16),68.00%(17/25),76.00%(19/25),100.00%(6/6),100.00%(6/6),respectively.The results of Spearman correlation analysis showed that whether RM-SLO could identify DME was associated with CMT and OCT classification(r=0.310,0.365;P=0.120,0.003);there was no correlation between FFA classification and macular volume(r=0.113,0.117;P=0.449,0.352).BCVA was correlated with CMT and macular volume(r=0.307,0.269;P=0.013,0.030),however,there was no significant correlation with OCT type,angiographic type(r=0.051,0.175;P=0.684,0.240).Conclusion:The diagnostic agreement of DME are good between RMDR,RMDL of RM-SLO image.DME of DRT type and patients with smaller CMT in OCT are difficult to identified by RM-SLO fundus imaging.