Prediction of OCT Images of Short-Term Response to Anti-VEGF Treatment for Diabetic Macular Edema Using Different Generative Adversarial Networks

18 Pages Posted: 5 Dec 2022

See all articles by Shaopeng Liu

Shaopeng Liu

Guangdong Polytechnic Normal University

Wanlu Hu

Guangdong Polytechnic Normal University

Fabao Xu

Shandong University - Qilu Hospital

Wenjie Chen

People's Hospital of Xiajin

Jie Liu

People's Hospital of Zoucheng

Xuechen Yu

Shandong University - Cheeloo College of Medicine

Zhengfei Wang

Guangzhou University of Chinese Medicine

Zhongwen Li

Wenzhou Medical University

Zhiwen Li

Shandong University - Qilu Hospital

Xueying Yang

Shandong University - Qilu Hospital

Boxuan Song

Shandong University - Qilu Hospital

Shaopeng Wang

Zibo Central Hospital

Kai Wang

Guangdong Polytechnic Normal University

Xinpeng Wang

Guangdong Polytechnic Normal University

Jiaming Hong

Guangzhou University of Chinese Medicine

Li Zhang

Huazhong University of Science and Technology

Jianqiao Li

Shandong University - Qilu Hospital

Abstract

Purpose: This study sought to assess the predictive performance of optical coherence tomography (OCT) images for the response of diabetic macular edema (DME) patients to anti-vascular endothelial growth factor (VEGF) therapy generated from baseline images using generative adversarial networks (GANs).

Methods: Patient information, including clinical and radiologic parameters, was obtained from inpatients at the Ophthalmology Department of Qilu Hospital. 715 and 103 pairs of pre-and post-treatment OCT images of DME patients were included in the training and validation sets, respectively. The post-treatment OCT images were used to assess the validity of the generated images. Six different GAN models (CycleGAN, PairGAN, Pix2pixHD, RegGAN, SPADE, UNIT) were applied to predict the efficacy of anti-VEGF treatment by generating OCT images. Independent screening and evaluation experiments were conducted to validate the quality and comparability of images generated by different GAN models.

Results: OCT images generated f GAN models exhibited high comparability to the real images, especially for edema absorption. RegGAN exhibited the highest prediction accuracy over the CycleGAN, PairGAN, Pix2pixHD, SPADE, and UNIT models. Further analyses were conducted based on the RegGAN. Retinal specialists found it difficult to accurately assess response with most synthetic images (95/103). A mean absolute error of 26.74 ± 21.28 μm was observed for central macular thickness (CMT) between the synthetic and real OCT images.

Conclusion: Different generative adversarial networks have different prognostic efficacy for DME, and RegGAN yielded the best performance in our study. Different GAN models yielded good accuracy in predicting the OCT-based response to anti-VEGF treatment at one month. Overall, the application of GAN models can assist clinicians in prognosis prediction of patients with DME to design better treatment strategies and follow-up schedules.

Note:

Funding Information: This work is partially supported by the Opening Project of Guangdong Province Key Laboratory of Big Data Analysis and Processing at Sun Yat-sen University (202001, 202202), the Natural Science Foundation of Guangdong Province (2019A1515012048), and the Social Sciences Project of Guangzhou University of Chinese Medicine (2021SKYB01).

Declaration of Interests: The authors declare that they have no conflict of interest.

Ethics Approval Statement: Written informed consent was not required, given the retrospective nature of our study, and all images were fully anonymized. Moreover, this study was conducted following the ethical requirements established in the Declaration of Helsinki. The institutional review board of Qilu Hospital also approved this study (Ethical Code: 2021 [068]).

Keywords: Diabetic macular edema, generative adversarial networks, optical coherence tomography, anti-vascular endothelial growth factor, prognostic predictions

undefined

Suggested Citation

Liu, Shaopeng and Hu, Wanlu and Xu, Fabao and Chen, Wenjie and Liu, Jie and Yu, Xuechen and Wang, Zhengfei and Li, Zhongwen and Li, Zhiwen and Yang, Xueying and Song, Boxuan and Wang, Shaopeng and Wang, Kai and Wang, Xinpeng and Hong, Jiaming and Zhang, Li and Li, Jianqiao, Prediction of OCT Images of Short-Term Response to Anti-VEGF Treatment for Diabetic Macular Edema Using Different Generative Adversarial Networks. Available at SSRN: https://ssrn.com/abstract=4273618 or http://dx.doi.org/10.2139/ssrn.4273618

Shaopeng Liu

Guangdong Polytechnic Normal University ( email )

293 Zhongshan Middle Ave
Guangzhou, 510665
China

Wanlu Hu

Guangdong Polytechnic Normal University ( email )

293 Zhongshan Middle Ave
Guangzhou, 510665
China

Fabao Xu (Contact Author)

Shandong University - Qilu Hospital ( email )

Wenjie Chen

People's Hospital of Xiajin ( email )

Jie Liu

People's Hospital of Zoucheng ( email )

Xuechen Yu

Shandong University - Cheeloo College of Medicine ( email )

Zhengfei Wang

Guangzhou University of Chinese Medicine ( email )

Guangzhou
China

Zhongwen Li

Wenzhou Medical University ( email )

China

Zhiwen Li

Shandong University - Qilu Hospital ( email )

Xueying Yang

Shandong University - Qilu Hospital ( email )

Boxuan Song

Shandong University - Qilu Hospital ( email )

Shaopeng Wang

Zibo Central Hospital ( email )

China

Kai Wang

Guangdong Polytechnic Normal University ( email )

293 Zhongshan Middle Ave
Guangzhou, 510665
China

Xinpeng Wang

Guangdong Polytechnic Normal University ( email )

293 Zhongshan Middle Ave
Guangzhou, 510665
China

Jiaming Hong

Guangzhou University of Chinese Medicine ( email )

Guangzhou
China

Li Zhang

Huazhong University of Science and Technology ( email )

Jianqiao Li

Shandong University - Qilu Hospital ( email )

0 References

    0 Citations

      Do you have a job opening that you would like to promote on SSRN?

      Paper statistics

      Downloads
      39
      Abstract Views
      196
      PlumX Metrics