目的：开发一款可自动校准测试距离的智能手机视力检测APP(WHOeye的iOS版本)，并评估其实用性。方法： WHOeyes在经过验证的视力检测APP “V@home”的基础上新增自动距离校准(automatic distance calibration, ADC)功能。研究招募了3组不同年龄(≤20岁、20~40岁、>40岁)的中国受试者，分别使用糖尿病视网膜病变早期治疗研究(Early Treatment Diabetic Retinopathy Study, ETDRS)视力表和WHOeyes进行远距离和近距离的视力检测。ADC功能用于确定WHOeyes的测试距离。红外测距仪用于确定ETDRS的测试距离以及WHOeyes的实际测试距离。通过问卷调查评估用户满意度。结果：WHOeyes ADC确定的实际测试距离在3个年龄组中均与预期测试距离总体上表现出良好的一致性(P > 0.50)。在远距离和近距离视力检测方面，WHOeyes的准确性与ETDRS相当。WHOeyes与ETDRS之间的平均视力差异范围为–0.084 ~ 0.012 log^MAR，各组的二次加权卡帕系数(quadratic weighted kappa, QWK)均大于0.75。WHOeyes在近距离和远距离视力检测中的重测信度高，平均差异范围为–0.040 ~ 0.004 log^MAR，QWK均大于0.85。问卷调查显示WHOeyes具有较好的用户体验和接受度。结论：与金标准ETDRS视力表方法相比，WHOeyes测试距离较为准确，可以提供准确的远距离和近距离视力测量结果。

Background: To develop and assess usability of a smartphone-based visual acuity (VA) test with an automatic distance calibration (ADC) function, the iOS version of WHOeyes. Methods: The WHOeyes was an upgraded version with a distinct feature of ADC of an existing validated VA testing APP called V@home. Three groups of Chinese participants with different ages (≤20, 20-40, >40 years) were recruited for distance and near VA testing using both an Early Treatment Diabetic Retinopathy Study (ETDRS) chart and the WHOeyes. The ADC function would determine the testing distance. Infrared rangefinder was used to determine the testing distance for the ETDRS, and actual testing distance for the WHOeyes. A questionnaire-based interview was administered to assess satisfaction. Results: The actual testing distance determined by the WHOeyes ADC showed an overall good agreement with the desired testing distance in all three age groups (p > 0.50). Regarding the distance and near VA testing, the accuracy of WHOeyes was equivalent to ETDRS. The mean difference between the WHOeyes and ETDRS ranged from -0.084 to 0.012 log^MAR, and the quadratic weighted kappa (QWK) values were greater than 0.75 across all groups. The test-retest reliability of WHOeyes was high for both near and distance VA, with a mean difference ranging from -0.040 to 0.004 log^MAR and QWK all greater than 0.85. The questionnaire revealed an excellent user experience and acceptance of WHOeyes. Conclusion: WHOeyes could provide accurate measurement of the testing distance as well as the distance and near VA when compared to the gold standard ETDRS chart.

目的: 探讨 ETDRS 对数视力表对儿童视力检查的可重复性及其影响的相关因素。

方法: 在流行病学调查的过程中, 随机使用 ETDRS 对数视力表, 为 250 位裸眼视力低于 0.5 和 98 位视力正常儿童进行裸眼视力重复检查。
结果: 两次视力测量之间差异的均数为0.004log^±0.07; Kappa 分析结果具有很好的一致性(k = 0.71) ; 性别与视力检查一致性无明显相关(P = 0.845) ; 年龄与视力检查一致性有显著相关性(P = 0.019) , 年龄越小视力检查一致性越差; 屈光不正与视力检查一致性也有显著相关性(P = 0.000) , 近视度数在- 1.00D～- 5.00D 之间的儿童视力检查一致性相对差, 而正视眼的视力检查一致性较好。

结论: 结果提示 ETDRS 对数视力表适合儿童视力检查, 建议推广使用。

Purpose: To evaluate repeatability of the ETDRS log MAR visual acuity measurementin children and the relative influence factors.

Methods: The children (n = 348) with visual acuity less than 0.5 ( equal to 0.3 logunit) in either eye,or one tenth in children with normal visual acuity were chosen todo repeatable uncorrected VA measurement with Bland-Altman analysis and Kappa analysis using ETDRS acuity chart.

Results: The mean difference of visual acuity was 0.004log^±0.07. There was a significant repeatability (k = 0.71) between two visual acuity examination. There were significant consistent results both on male and female patients (P = 0.845) . A significant relationship was found between age and VA repeatability(P = 0.019) . The VA repeatability could also be influenced by refractive error (P = 0.000) . The acuity measurement in children with emmetropia (k = 0.82) had a higher repeatability than in the children with myopia (k = 0.66) .

Conclusions: The ETDRS visual acuity chart can provide a repeatable measure of visual acuity in children. It is recommended for clinic examination of children. 

目的：探索基于眼底彩照和人工智能构建冠心病智能诊断系统的可行性。方法：于2013—2014年收集广东省人民医院530例患者共2117张眼底彩照，其中冠心病217例共909张眼底彩照。根据患者有无冠心病的情况进行标记，使用Inception-V3深度卷积神经网络训练人工智能模型，随后使用验证数据判断模型的准确率。计算深度卷积网络模型的准确性、一致率、敏感性、特异性和受试者工作特性曲线下面积(area under the curve，AUC)。结果：在2117张眼底彩照中，1903张用于模型训练，214张用于模型的性能评估。在测试集中，该算法的准确性为98.1%，一致率为98.6%，敏感性为100.0%，特异性为96.7%，AUC为0.988(95%CI：0.974~1.000)。结论：眼底彩照联合人工智能技术可精准判定冠心病，该模型具备较高的敏感性和特异性，但须进一步增加样本量，使用大样本量数据验证该模型，排除过拟合的可能性。

Objective: To explore the feasibility of developing a deep learning algorithm for detecting coronary heart diseases based on fundus color photography and artificial intelligence (AI). Methods: A total of 2 117 fundus  color photographs were taken from 530 patients in Guangdong Provincial People’s Hospital from 2013 to 2014,including 909 fundus color photographs from 217 patients with coronary heart disease (CHD). According to whether the patient had coronary heart disease or not, the Inception-V3 depth convolution neural network was used to train the deep learning model, and then the validation data were used to judge the accuracy of the model. The accuracy, consistency rate, sensitivity and specificity of the deep convolution network model and the area under the working characteristic curve (AUC) were calculated. Results: Among the 2 117 fundus color photographs, 1 903 were used for model training, and 214 were used to test the accuracy of the model. In the test dataset, the accuracy of the algorithm was 98.1%, the consistency rate was 98.6%, the sensitivity was 100.0%, and the specificity was 96.7%. The AUC was 0.988 (95% CI, 0.974–1.000). Conclusion: The combination of fundus color photography and artificial intelligence can achieve the accurate diagnosis of the coronary heart disease, and the model has high sensitivity and specificity. However, future studies are warranted to validate our model and exclude the possibility of over-fitting.

目的：探索基于眼底彩照和人工智能构建冠心病智能诊断系统的可行性。方法：于2013—2014年收集广东省人民医院530例患者共2 117张眼底彩照，其中冠心病217例共909张眼底彩照。根据患者有无冠心病的情况进行标记，使用Inception-V3深度卷积神经网络训练人工智能模型，随后使用验证数据判断模型的准确率。计算深度卷积网络模型的准确性、一致率、敏感性、特异性和受试者工作特性曲线下面积(area under the curve，AUC)。结果：在2 117张眼底彩照中，1 903张用于模型训练，214张用于模型的性能评估。在测试集中，该算法的准确性为98.1%，一致率为98.6%，敏感性为100.0%，特异性为96.7%，AUC为0.988(95%CI：0.974~1.000)。结论：眼底彩照联合人工智能技术可精准判定冠心病，该模型具备较高的敏感性和特异性，但须进一步增加样本量，使用大样本量数据验证该模型，排除过拟合的可能性。

Objective: To explore the feasibility of developing a deep learning algorithm for detecting coronary heart diseases based on fundus color photography and artificial intelligence (AI). Methods: A total of 2 117 fundus color photographs were taken from 530 patients in Guangdong Provincial People’s Hospital from 2013 to 2014, including 909 fundus color photographs from 217 patients with coronary heart disease (CHD). According to whether the patient had coronary heart disease or not, the Inception-V3 depth convolution neural network was used to train the deep learning model, and then the validation data were used to judge the accuracy of the model. The accuracy, consistency rate, sensitivity and specificity of the deep convolution network model and the area under the working characteristic curve (AUC) were calculated. Results: Among the 2 117 fundus color photographs, 1 903 were used for model training, and 214 were used to test the accuracy of the model. In the test dataset, the accuracy of the algorithm was 98.1%, the consistency rate was 98.6%, the sensitivity was 100.0%, and the specificity was 96.7%. The AUC was 0.988 (95% CI, 0.974–1.000). Conclusion: The combination of fundus color photography and artificial intelligence can achieve the accurate diagnosis of the coronary heart disease, and the model has high sensitivity and specificity. However, future studies are warranted to validate our model and exclude the possibility of over-fitting.