LONGL-Net: temporal correlation structure guided deep learning model to predict longitudinal age-related macular degeneration severity.
LONGL-Net: temporal correlation structure guided deep learning model to predict longitudinal age-related macular degeneration severity.
复制标题
DOI：
10.1093/pnasnexus/pgab003
复制
发表时间：
2022-03
期刊：
影响因子：
--
通讯作者：
Chen W
中科院分区：
文献类型：
--
作者：
Ganjdanesh A;Zhang J;Chew EY;Ding Y;Huang H;Chen W
视网膜相关性黄斑变性（AMD）是发达国家致盲的主要原因，到2040年其患病率将增加到2.88亿人。因此，自动分级和预测方法对于识别晚期AMD的易感受试者并使临床医生能够开始对他们采取预防措施非常有益。在临床上，AMD严重程度通过视网膜的彩色眼底照片（CFP）来量化，并且提出了许多基于机器学习的方法来对AMD严重程度进行分级。然而，很少有模型被开发来预测纵向进展状态，即基于当前CFP预测未来晚期AMD风险，这在临床上更有趣。在本文中，我们提出了一种新的基于深度学习的分类模型（LONGL-Net），可以同时对当前CFP进行分级并预测纵向结果，即受试者在未来时间点是否会处于晚期AMD。我们设计了一种新的时间相关结构引导的生成对抗网络模型，该模型学习连续时间点CFPs中时间变化的相互关系，并通过预测未来CFPs中的AMD症状来为分类器的决策提供可解释性。我们使用了来自4，628名参与者的约30，000张CFP图像，这些图像来自与眼相关的眼病研究。我们的分类器在同时对当前时间点的AMD状况进行分级和预测受试者在未来时间点的晚期AMD进展的3类分类问题上显示平均0.905（95%CI：0.886-0.922）AUC和0.762（95%CI：0.733-0.792）准确度。我们在英国生物银行数据集上进一步验证了我们的模型，我们的模型在300张CFP图像的分级中显示出平均0.905的准确性和0.797的灵敏度。
Age-related macular degeneration (AMD) is the principal cause of blindness in developed countries, and its prevalence will increase to 288 million people in 2040. Therefore, automated grading and prediction methods can be highly beneficial for recognizing susceptible subjects to late-AMD and enabling clinicians to start preventive actions for them. Clinically, AMD severity is quantified by Color Fundus Photographs (CFP) of the retina, and many machine-learning-based methods are proposed for grading AMD severity. However, few models were developed to predict the longitudinal progression status, i.e. predicting future late-AMD risk based on the current CFP, which is more clinically interesting. In this paper, we propose a new deep-learning-based classification model (LONGL-Net) that can simultaneously grade the current CFP and predict the longitudinal outcome, i.e. whether the subject will be in late-AMD in the future time-point. We design a new temporal-correlation-structure-guided Generative Adversarial Network model that learns the interrelations of temporal changes in CFPs in consecutive time-points and provides interpretability for the classifier's decisions by forecasting AMD symptoms in the future CFPs. We used about 30,000 CFP images from 4,628 participants in the Age-Related Eye Disease Study. Our classifier showed average 0.905 (95% CI: 0.886–0.922) AUC and 0.762 (95% CI: 0.733–0.792) accuracy on the 3-class classification problem of simultaneously grading current time-point's AMD condition and predicting late AMD progression of subjects in the future time-point. We further validated our model on the UK Biobank dataset, where our model showed average 0.905 accuracy and 0.797 sensitivity in grading 300 CFP images.
登录
查看更多内容
影响因子：
--
作者：
Kassoff, A;Kassoff, J;Chew, EY
通讯作者：
Chew, EY
DOI：
10.1146/annurev-genom-090413-025610
发表时间：
2014
影响因子：
8.7
作者：
Fritsche LG;Fariss RN;Stambolian D;Abecasis GR;Curcio CA;Swaroop A
通讯作者：
Swaroop A
影响因子：
13.7
作者：
Keenan, Tiarnan D.;Dharssi, Shazia;Chew, Emily Y.
通讯作者：
Chew, Emily Y.
DOI：
10.1073/pnas.1717139115
发表时间：
2018-03-27
影响因子：
11.1
作者：
Mobadersany P;Yousefi S;Amgad M;Gutman DA;Barnholtz-Sloan JS;Velázquez Vega JE;Brat DJ;Cooper LAD
通讯作者：
Cooper LAD
影响因子：
4.2
作者：
Holz, Frank G.;Bindewald-Wittich, Almut;Schmitz-Valckenberg, Steffen
通讯作者：
Schmitz-Valckenberg, Steffen