目的：与金标准荧光造影比较，评估免散瞳眼底照相及散瞳眼底照相作为眼底筛查诊断方法的优势。

方法：选取上海市浦东新区4个街道纳入糖尿病慢性病管理档案的患者276例，进行糖尿病视网膜病变(DR)的筛查。在进行视力、裂隙灯检查、屈光检查之后，对所有参检人员分别进行双眼免散瞳眼底数码照相、双眼散瞳眼底数码照相及眼底荧光造影； 由专业人员集中阅片并给出检查报告。可疑眼底病患者，预约相应的专科完成进一步诊治。

结果：共计276例接受眼底病筛查，双眼分别进行免散瞳及散瞳眼底数码照相，共得到1 104张彩色眼底图片，其中1 056张(95.65%)眼底像可用于分析。免散瞳眼底数码照相获得的眼底图片中，可以评估的眼底图片408张，基本可以评估的眼底图片116张，不能评估的眼底图片28张； 散瞳眼底数码照相获得的眼底图片中，可以评估的眼底图片432张，基本可以评估的眼底图片100张，不能评估的眼底图片20张。经卡方检验，两种方式获得的眼底图片质量差异无统计学意义(P>0.05)。与眼底荧光造影(FFA)相对比，以DR Ⅰ期作为临界值，免散瞳眼底数码照相的诊断特异性是95.71%，诊断敏感性是93.56%； 散瞳眼底数码照相的诊断特异性是95.43%，诊断敏感性是98.02%； 两种筛查方法比较差异无统计学意义(P>0.05)； 以DR Ⅱ期作为临界值，免散瞳眼底照相对DR的诊断特异性和敏感性分别为95.35%和93.44%； 散瞳眼底照相对DR的诊断特异性和敏感性分别为95.81%和98.36%； 两种筛查方法比较差异无统计学意义(P>0.05)。

结论：免散瞳眼底数码照相与散瞳眼底数码照相均可用于眼底病筛查诊断。免散瞳眼底数码照相更加简易、快捷，适合大规模筛查时使用。散瞳眼底照相对于疾病的跟踪随访更能提供详细的参考信息。

METHODS: A total of 276 patients which involved in Chronic Diabetes Management Achieves within 4 streets of Pudong District Shanghai, were enrolled for diabetic retinopathy(DR)examination including NMFCS, MFCS and FFA. These DR examinations were proceeded after vision, slit-lamp and dioptroscopy tests, and reported by professionals. For those with suspicious fundus diseases, we would make appointments with specialist for further treatment.

RESULTS: A total of 1104 colorful fundus images, and 1056 images(95.65%)could be used to analyze. There were 408 appreciable images, 116 basically appreciable images and 28 unusable images in 552 NMFCS images. In addition, there were 432 appreciable images, 100 basically appreciable images and 20 unusable images in 552 MFCS images. There was no significant differences between NMFCS and MFCS(P>0.05). Compared with FFA with DRⅠ as the critical value, the specificity of digital photography for NMFCS was 95.71%, the sensitivity was 93.56%; however, MFCS are 95.43% and 98.02%. There was no statistically significant difference between the two screening methods(P>0.05). Compared with FFA with DRⅡ as the critical value, the specificity of digital photography for NMFCS was 95.35% and the sensitivity was 93.44%; however, for MFCS were 95.81% and 98.36%. There was no statistically significant difference between the two screening methods(P>0.05).

CONCLUSION: Both NMFCS and MFCS could be used for the diagnosis and screening for eye diseases. NMFCS is easier and faster for digital photography, which is suitable for mass screening. MFCS is more likely to provide detailed information about the follow-up of the disease.