Algorithm Variability in Quantification of Epithelial Defect Size in Microbial Keratitis Images.
Algorithm Variability in Quantification of Epithelial Defect Size in Microbial Keratitis Images.
复制标题
DOI：
10.1097/ico.0000000000002258
复制
发表时间：
2020-05
期刊：
影响因子：
2.8
通讯作者：
Woodward MA
中科院分区：
文献类型：
--
作者：
Kriegel MF;Huang J;Ashfaq HA;Niziol LM;Preethi M;Tan H;Tuohy MM;Patel TP;Prajna V;Woodward MA
研究使用半自动图像分析算法从裂隙灯摄影（SLP）图像定量微生物角膜炎（MK）形态时测量变异性的来源。普罗维登招募的MK患者接受SLP以获得其上皮缺损（艾德）的图像。眼睛用荧光素染色，并在蓝光下以低和高放大倍数多次成像。一位戴面具的研究助理选择了3个最佳图像，并且每3次注释以提供对应于艾德和健康角膜的种子区域。该算法返回每个种子图像的艾德区域。排除了无ED和算法失败的眼睛。方差分量估计与随机效应模型和组内相关系数（ICC）估计年级内的可靠性。共拍摄了42名MK参与者的42只眼睛。在排除低质量图像、无ED的眼睛和算法失败后，分析了34名患者的92张图像和274粒种子。平均艾德面积在SEREX或高与低SLP放大率之间没有发现显著差异（所有p>0.5，配对t检验）。图像（0.9%）、放大倍数（0.2%）或种子（0.1%）导致的测量变异性极小。大多数变异性可归因于患者之间艾德尺寸的差异（85.2%）。13.7%的变异无法解释。该算法在同一图像上的多次迭代显示出良好的一致性（ICC=0.98，95%置信区间，0.97-0.99）。图像分析算法显示出良好的可靠性测量艾德面积从SLP图像。大多数测量变异性是由于患者之间的差异，而不是成像设置或用户的算法应用。
To investigate sources of measurement variability when quantifying morphology of microbial keratitis (MK) from slit lamp photography (SLP) images using a semi-automated, image-analysis algorithm. Prospectively enrolled patients with MK underwent SLP to obtain images of their epithelial defects (ED). Eyes were stained with fluorescein and imaged multiple times under blue light, at low and high magnifications. A masked research assistant chose the 3 best images and annotated each 3 times to provide seed regions corresponding to ED and healthy cornea. The algorithm returned ED area for each seeded image. Eyes without EDs and algorithm failures were excluded. Variance components were estimated with a random effects model and intraclass correlation coefficients (ICC) estimated intra-grader reliability. A total of 42 eyes from 42 MK participants were photographed. After excluding poor quality images, eyes with no EDs, and algorithm failures, 34 patients with 92 images and 274 seeds were analyzed. No significant differences in average ED area were found between seedings or high versus low SLP magnifications (all p>0.5, paired t-tests). Minimal measurement variability was due to image (0.9%), magnification (0.2%), or seed (0.1%). Most variability was attributable to differences in ED sizes between patients (85.2%). 13.7% of variability was unexplained. Multiple iterations of the algorithm on the same image showed good consistency (ICC=0.98, 95% confidence interval, 0.97-0.99). Image-analysis algorithms showed good reliably for measuring ED area from SLP images. Most measurement variability was due to between-patient differences, not imaging settings or application of the algorithm by the user.