Sharif Amit Kamran PhD , Khondker Fariha Hossain MS , Joshua Ong MD , Ethan Waisberg Mb BCh, BAO , Nasif Zaman MS , Salah A. Baker PhD , Andrew G. Lee MD , Alireza Tavakkoli PhD
{"title":"FA4SANS-GAN:一种新型机器学习生成对抗网络,用于进一步了解航天相关神经-眼综合征(SANS)的眼部变化","authors":"Sharif Amit Kamran PhD , Khondker Fariha Hossain MS , Joshua Ong MD , Ethan Waisberg Mb BCh, BAO , Nasif Zaman MS , Salah A. Baker PhD , Andrew G. Lee MD , Alireza Tavakkoli PhD","doi":"10.1016/j.xops.2024.100493","DOIUrl":null,"url":null,"abstract":"<div><h3>Purpose</h3><p>To provide an automated system for synthesizing fluorescein angiography (FA) images from color fundus photographs for averting risks associated with fluorescein dye and extend its future application to spaceflight associated neuro-ocular syndrome (SANS) detection in spaceflight where resources are limited.</p></div><div><h3>Design</h3><p>Development and validation of a novel conditional generative adversarial network (GAN) trained on limited amount of FA and color fundus images with diabetic retinopathy and control cases.</p></div><div><h3>Participants</h3><p>Color fundus and FA paired images for unique patients were collected from a publicly available study.</p></div><div><h3>Methods</h3><p>FA4SANS-GAN was trained to generate FA images from color fundus photographs using 2 multiscale generators coupled with 2 patch-GAN discriminators. Eight hundred fifty color fundus and FA images were utilized for training by augmenting images from 17 unique patients. The model was evaluated on 56 fluorescein images collected from 14 unique patients. In addition, it was compared with 3 other GAN architectures trained on the same data set. Furthermore, we test the robustness of the models against acquisition noise and retaining structural information when introduced to artificially created biological markers.</p></div><div><h3>Main Outcome Measures</h3><p>For GAN synthesis, metric Fréchet Inception Distance (FID) and Kernel Inception Distance (KID). Also, two 1-sided tests (TOST) based on Welch’s <em>t</em> test for measuring statistical significance.</p></div><div><h3>Results</h3><p>On test FA images, mean FID for FA4SANS-GAN was 39.8 (standard deviation, 9.9), which is better than GANgio model’s mean of 43.2 (standard deviation, 13.7), Pix2PixHD’s mean of 57.3 (standard deviation, 11.5) and Pix2Pix’s mean of 67.5 (standard deviation, 11.7). Similarly for KID, FA4SANS-GAN achieved mean of 0.00278 (standard deviation, 0.00167) which is better than other 3 model’s mean KID of 0.00303 (standard deviation, 0.00216), 0.00609 (standard deviation, 0.00238), 0.00784 (standard deviation, 0.00218). For TOST measurement, FA4SANS-GAN was proven to be statistically significant versus GANgio (<em>P</em> = 0.006); versus Pix2PixHD (<em>P</em> < 0.00001); and versus Pix2Pix (<em>P</em> < 0.00001).</p></div><div><h3>Conclusions</h3><p>Our study has shown FA4SANS-GAN to be statistically significant for 2 GAN synthesis metrics. Moreover, it is robust against acquisition noise, and can retain clear biological markers compared with the other 3 GAN architectures. This deployment of this model can be crucial in the International Space Station for detecting SANS.</p></div><div><h3>Financial Disclosure(s)</h3><p>The authors have no proprietary or commercial interest in any materials discussed in this article.</p></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":null,"pages":null},"PeriodicalIF":3.2000,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666914524000290/pdfft?md5=41737a518d5b2ff8293eb883bd6938e7&pid=1-s2.0-S2666914524000290-main.pdf","citationCount":"0","resultStr":"{\"title\":\"FA4SANS-GAN: A Novel Machine Learning Generative Adversarial Network to Further Understand Ophthalmic Changes in Spaceflight Associated Neuro-Ocular Syndrome (SANS)\",\"authors\":\"Sharif Amit Kamran PhD , Khondker Fariha Hossain MS , Joshua Ong MD , Ethan Waisberg Mb BCh, BAO , Nasif Zaman MS , Salah A. Baker PhD , Andrew G. Lee MD , Alireza Tavakkoli PhD\",\"doi\":\"10.1016/j.xops.2024.100493\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><h3>Purpose</h3><p>To provide an automated system for synthesizing fluorescein angiography (FA) images from color fundus photographs for averting risks associated with fluorescein dye and extend its future application to spaceflight associated neuro-ocular syndrome (SANS) detection in spaceflight where resources are limited.</p></div><div><h3>Design</h3><p>Development and validation of a novel conditional generative adversarial network (GAN) trained on limited amount of FA and color fundus images with diabetic retinopathy and control cases.</p></div><div><h3>Participants</h3><p>Color fundus and FA paired images for unique patients were collected from a publicly available study.</p></div><div><h3>Methods</h3><p>FA4SANS-GAN was trained to generate FA images from color fundus photographs using 2 multiscale generators coupled with 2 patch-GAN discriminators. Eight hundred fifty color fundus and FA images were utilized for training by augmenting images from 17 unique patients. The model was evaluated on 56 fluorescein images collected from 14 unique patients. In addition, it was compared with 3 other GAN architectures trained on the same data set. Furthermore, we test the robustness of the models against acquisition noise and retaining structural information when introduced to artificially created biological markers.</p></div><div><h3>Main Outcome Measures</h3><p>For GAN synthesis, metric Fréchet Inception Distance (FID) and Kernel Inception Distance (KID). Also, two 1-sided tests (TOST) based on Welch’s <em>t</em> test for measuring statistical significance.</p></div><div><h3>Results</h3><p>On test FA images, mean FID for FA4SANS-GAN was 39.8 (standard deviation, 9.9), which is better than GANgio model’s mean of 43.2 (standard deviation, 13.7), Pix2PixHD’s mean of 57.3 (standard deviation, 11.5) and Pix2Pix’s mean of 67.5 (standard deviation, 11.7). Similarly for KID, FA4SANS-GAN achieved mean of 0.00278 (standard deviation, 0.00167) which is better than other 3 model’s mean KID of 0.00303 (standard deviation, 0.00216), 0.00609 (standard deviation, 0.00238), 0.00784 (standard deviation, 0.00218). For TOST measurement, FA4SANS-GAN was proven to be statistically significant versus GANgio (<em>P</em> = 0.006); versus Pix2PixHD (<em>P</em> < 0.00001); and versus Pix2Pix (<em>P</em> < 0.00001).</p></div><div><h3>Conclusions</h3><p>Our study has shown FA4SANS-GAN to be statistically significant for 2 GAN synthesis metrics. Moreover, it is robust against acquisition noise, and can retain clear biological markers compared with the other 3 GAN architectures. This deployment of this model can be crucial in the International Space Station for detecting SANS.</p></div><div><h3>Financial Disclosure(s)</h3><p>The authors have no proprietary or commercial interest in any materials discussed in this article.</p></div>\",\"PeriodicalId\":74363,\"journal\":{\"name\":\"Ophthalmology science\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":3.2000,\"publicationDate\":\"2024-02-15\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.sciencedirect.com/science/article/pii/S2666914524000290/pdfft?md5=41737a518d5b2ff8293eb883bd6938e7&pid=1-s2.0-S2666914524000290-main.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Ophthalmology science\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2666914524000290\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"OPHTHALMOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Ophthalmology science","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2666914524000290","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"OPHTHALMOLOGY","Score":null,"Total":0}
引用次数: 0
摘要
目的提供一种从彩色眼底照片合成荧光素血管造影(FA)图像的自动化系统,以避免与荧光素染料相关的风险,并将其未来应用扩展到资源有限的太空飞行相关神经眼综合征(SANS)检测中。方法使用 2 个多尺度生成器和 2 个补丁-GAN 判别器训练 FA4SAN-GAN 从彩色眼底照片生成 FA 图像。通过增强 17 位独特患者的图像,利用 8500 张彩色眼底和 FA 图像进行了训练。该模型在从 14 名患者处收集的 56 张荧光素图像上进行了评估。此外,该模型还与在相同数据集上训练的其他 3 个 GAN 架构进行了比较。此外,我们还测试了模型对采集噪声的鲁棒性,以及在引入人工创建的生物标记时对结构信息的保留情况。主要结果测量对于 GAN 合成,指标为弗雷谢特起始距离(FID)和核起始距离(KID)。结果在测试的 FA 图像上,FA4SANS-GAN 的平均 FID 为 39.8(标准差,9.9),优于 GANgio 模型的平均值 43.2(标准差,13.7)、Pix2PixHD 的平均值 57.3(标准差,11.5)和 Pix2Pix 的平均值 67.5(标准差,11.7)。同样,在 KID 方面,FA4SANS-GAN 的平均值为 0.00278(标准偏差为 0.00167),优于其他三个模型的平均 KID 值 0.00303(标准偏差为 0.00216)、0.00609(标准偏差为 0.00238)和 0.00784(标准偏差为 0.00218)。在 TOST 测量方面,FA4SANS-GAN 与 GANgio(P = 0.006)、Pix2PixHD(P < 0.00001)和 Pix2Pix(P < 0.00001)相比具有显著的统计学意义。此外,与其他 3 种 GAN 架构相比,FA4SANS-GAN 对采集噪声具有鲁棒性,并能保留清晰的生物标记。在国际空间站部署该模型对检测 SANS 至关重要。
FA4SANS-GAN: A Novel Machine Learning Generative Adversarial Network to Further Understand Ophthalmic Changes in Spaceflight Associated Neuro-Ocular Syndrome (SANS)
Purpose
To provide an automated system for synthesizing fluorescein angiography (FA) images from color fundus photographs for averting risks associated with fluorescein dye and extend its future application to spaceflight associated neuro-ocular syndrome (SANS) detection in spaceflight where resources are limited.
Design
Development and validation of a novel conditional generative adversarial network (GAN) trained on limited amount of FA and color fundus images with diabetic retinopathy and control cases.
Participants
Color fundus and FA paired images for unique patients were collected from a publicly available study.
Methods
FA4SANS-GAN was trained to generate FA images from color fundus photographs using 2 multiscale generators coupled with 2 patch-GAN discriminators. Eight hundred fifty color fundus and FA images were utilized for training by augmenting images from 17 unique patients. The model was evaluated on 56 fluorescein images collected from 14 unique patients. In addition, it was compared with 3 other GAN architectures trained on the same data set. Furthermore, we test the robustness of the models against acquisition noise and retaining structural information when introduced to artificially created biological markers.
Main Outcome Measures
For GAN synthesis, metric Fréchet Inception Distance (FID) and Kernel Inception Distance (KID). Also, two 1-sided tests (TOST) based on Welch’s t test for measuring statistical significance.
Results
On test FA images, mean FID for FA4SANS-GAN was 39.8 (standard deviation, 9.9), which is better than GANgio model’s mean of 43.2 (standard deviation, 13.7), Pix2PixHD’s mean of 57.3 (standard deviation, 11.5) and Pix2Pix’s mean of 67.5 (standard deviation, 11.7). Similarly for KID, FA4SANS-GAN achieved mean of 0.00278 (standard deviation, 0.00167) which is better than other 3 model’s mean KID of 0.00303 (standard deviation, 0.00216), 0.00609 (standard deviation, 0.00238), 0.00784 (standard deviation, 0.00218). For TOST measurement, FA4SANS-GAN was proven to be statistically significant versus GANgio (P = 0.006); versus Pix2PixHD (P < 0.00001); and versus Pix2Pix (P < 0.00001).
Conclusions
Our study has shown FA4SANS-GAN to be statistically significant for 2 GAN synthesis metrics. Moreover, it is robust against acquisition noise, and can retain clear biological markers compared with the other 3 GAN architectures. This deployment of this model can be crucial in the International Space Station for detecting SANS.
Financial Disclosure(s)
The authors have no proprietary or commercial interest in any materials discussed in this article.