青光眼影像人工智能深度学习研究现状与展望

doi:10.6040/j.issn.1671-7554.0.2020.1249

Abstract

Abstract:

Glaucoma is a group of heterogeneous neurodegenerative diseases, characterized by the gradual loss of retinal ganglion cells and their axons, and has now become the major reason of irreversible blindness worldwide. Artificial intelligence (AI) is intelligence demonstrated by machines. Deep learning (DL) is a subset of AI based on deep neural networks, and it has made great breakthroughs in medical imaging. In glaucoma imaging, research interests have been increasing on applying DL in fundus photographs and optical coherence tomography (OCT) for glaucomatous optic neuropathy (GON) detection. Promising results show that the incorporation of DL technology in imaging for glaucoma assessment is efficient and accurate, which could potentially address some gaps in the current practice and clinical workflow. However, further research is crucial in tackling some existing challenges, such as setting a standard for ground truth labelling among different studies, visualizing the learning process in the "black box", improving the model generalizability on unseen datasets, developing the DL-powered infrastructure for real-world implementation, establishing a practical clinical workflow, conducting prospective validation and cost-effectiveness analysis. This review summarizes recent studies on the application of AI on glaucoma imaging, discusses the potential clinical impact and future research directions.

Key words: Glaucoma imaging, Artificial intelligence, Optical coherence tomography, Fundus photography, Deep learning

CLC Number:

R574

Carol Y. Cheung,Anran RAN. Artificial intelligence deep learning in glaucoma imaging: current progress and future prospect[J].Journal of Shandong University (Health Sciences), 2020, 58(11): 24-32, 38.

Figures/Tables 5

Fig.1

Table 1

The basic data sets for the development and performance evaluation of deep learning models"

术语	说明
训练集	与调试集和内部验证集来自同一个大的数据集，是基于一定比例划分且互不重合，用于建立深度学习模型。研究者可构建具有不同参数甚至不同网络架构的DL模型来学习训练集中青光眼性视神经改变的相关特征。
调试集	与训练集和内部验证集来自同一个大的数据集，是基于一定比例划分且互不重合，用于实时评估不同模型在训练期间的性能。研究者可根据实时观察的学习曲线进行参数调节，确定训练停止节点，并最终选择最佳DL模型。
内部测试集	与训练集和调试集来自同一个大的数据集，是基于一定比例划分且互不重合，用于评估所选的最佳DL模型在该未知数据集上的性能。临床研究中，研究者需汇报模型在内部测试集上的性能，通常包括受试者工作特征曲线下面积，灵敏度，特异性以及准确性等。
外部测试集	一个或多个独立的数据集，与训练集、测试集和内部验证集来自不同数据集，用于评估上述DL模型在其他未知数据集上的性能，验证其临床通用性。研究者同样需汇报模受试者工作特征曲线下面积、灵敏度、特异性以及准确性等。当DL模型在内部以及外部测试集都具有较好性能时，即证明该模型有良好的通用性。
K-倍交叉验证	主要用于数据量有限时。将整个大的数据集分为k个相等的小数据集，其中(k-1)个合并为训练集用于模型开发，剩下的一个作为测试集评估模型性能。该训练过程重复k次，每次选择其中一个小数据集用作测试集。研究者应汇报所有测试集的平均性能，通常为${\bar x}$±s。
留一法交叉验证	主要用于数据量非常有限时，是一种特殊类型的K-倍交叉验证，此时K等于数据总数(即标注图像的总数)。仅留下一张图像来测试模型性能，其他图像全部用于训练。该训练过程同样重复k次，研究者应汇报所有测试集的平均性能，通常为${\bar x}$±s。

Table 1

Fig.2

Fig.3

Fig.4

References 58

1	Weinreb RN , Aung T , Medeiros FA . The pathophysiology and treatment of glaucoma: a review[J]. JAMA, 2014, 311 (18): 1901- 1911. doi: 10.1001/jama.2014.3192
2	Tham YC , Li X , Wong TY , et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis[J]. Ophthalmology, 2014, 121 (11): 2081- 2090. doi: 10.1016/j.ophtha.2014.05.013
3	Weinreb RN , Leung CKS , Crowston JG , et al. Primary open-angle glaucoma[J]. Nat Rev Dis Primers, 2016, 2: 16067. doi: 10.1016/S0140-6736(04)16257-0
4	Jonas JB , Aung T , Bourne RR , et al. Glaucoma[J]. Lancet, 2017, 390 (10108): 2183- 2193. doi: 10.1016/S0140-6736(17)31469-1
5	Momont AC , Mills RP . Glaucoma screening: current perspectives and future directions[J]. Semin Ophthalmol, 2013, 28 (3): 185- 190. doi: 10.3109/08820538.2013.771200
6	Schuman JS . Detection and diagnosis of glaucoma: ocular imaging[J]. Invest Ophthalmol Vis Sci, 2012, 53 (5): 2488- 2490. doi: 10.1167/iovs.12-9483k
7	Rossetto JD , Melo LAS , Campos M , et al. Agreement on the evaluation of glaucomatous optic nerve head findings by ophthalmology residents and a glaucoma specialist[J]. Clin Ophthalmol, 2017, 11: 1281- 1284. doi: 10.2147/OPTH.S140225
8	Cheung CY , Leung CK , Lin D , et al. Relationship between retinal nerve fiber layer measurement and signal strength in optical coherence tomography[J]. Ophthalmology, 2008, 115 (8): 1347- 1351. doi: 10.1016/j.ophtha.2007.11.027
9	Cheung CY , Chan N , Leung CK . Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography: impact of signal strength on analysis of the RNFL map[J]. Asia Pac J Ophthalmol (Phila), 2012, 1 (1): 19- 23. doi: 10.1097/APO.0b013e31823e595d
10	Biswas S , Lin C , Leung CK . Evaluation of a myopic normative database for analysis of retinal nerve fiber layer thickness[J]. JAMA Ophthalmol, 2016, 134 (9): 1032- 1039. doi: 10.1001/jamaophthalmol.2016.2343
11	Andresen SL . John McCarthy: father of AI[J]. IEEE Intell Syst, 2002, 17 (5): 84- 85.
12	Simon A DMS , Venkatesan S , Babu DRR . An overview of machine learning and its applications[J]. International Journal of Electrical Sciences & Engineering, 2015, 1 (1): 22- 24.
13	Shinde PP , Shah S . A review of machine learning and deep learning applications[J]. 2018 Fourth International Conference on Computing Communication Control and Automation (Iccubea), Pune, India, 2018, 1- 8. doi: 10.1109/ICCUBEA.2018.8697857
14	Aggarwal CC . Training deep neural networks[M]. Cham: Springer International Publishing, 2018: 105- 167. doi: 10.1007/978-3-319-94463-0_3
15	Becker AS , Mueller M , Stoffel E , et al. Classification of breast cancer form ultrasound imaging using a generic deep learning analysis software: a pilot study[J]. Br J Radiol, 2017, 20170576. doi: 10.1259/bjr.20170576
16	Walsh SLF , Calandriello L , Silva M , et al. Deep learning for classifying fibrotic lung disease on high-resolution computed tomography: a case-cohort study[J]. Lancet Respir Med, 2018, 6 (11): 837- 845. doi: 10.1016/S2213-2600(18)30286-8
17	Brinker TJ , Hekler A , Enk A , et al. A convolutional neural network trained with dermoscopic images performed on par with 145 dermatologists in a clinical melanoma image classification task[J]. Eur J Cancer, 2019, 111: 148- 154. doi: 10.1016/j.ejca.2019.02.005
18	de Fauw J , Ledsam JR , Romera-Paredes B , et al. Clinically applicable deep learning for diagnosis and referral in retinal disease[J]. Nat Med, 2018, 24 (9): 1342- 1350. doi: 10.1038/s41591-018-0107-6
19	Ting DSW , Cheung CY , Lim G , et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes[J]. JAMA, 2017, 318 (22): 2211- 2223. doi: 10.1001/jama.2017.18152
20	Kermany DS , Goldbaum M , Cai W , et al. Identifying medical diagnoses and treatable diseases by image-based deep learning[J]. Cell, 2018, 172 (5): 1122- 1131. doi: 10.1016/j.cell.2018.02.010
21	Hekler A , Utikal JS , Enk AH , et al. Deep learning outperformed 11 pathologists in the classification of histopathological melanoma images[J]. Eur J Cancer, 2019, 118: 91- 96. doi: 10.1016/j.ejca.2019.06.012
22	Maron RC , Weichenthal M , Utikal JS , et al. Systematic outperformance of 112 dermatologists in multiclass skin cancer image classification by convolutional neural networks[J]. Eur J Cancer, 2019, 119: 57- 65. doi: 10.1016/j.ejca.2019.06.013
23	Dascalu A , David EO . Skin cancer detection by deep learning and sound analysis algorithms: a prospective clinical study of an elementary dermoscope[J]. EBioMedicine, 2019, 43: 107- 113. doi: 10.1016/j.ebiom.2019.04.055
24	Al-Antari MA , Al-Masni MA , Kim TS . Deep learning computer-aided diagnosis for breast Lesion in digital mammogram[J]. Adv Exp Med Biol, 2020, 1213: 59- 72. doi: 10.1007/978-3-030-33128-3_4
25	O'Mahony N , Campbell S , Carvalho A , et al. Deep learning vs. traditional computer vision[J]. Adv Intell Syst, 2020, 943: 128- 144.
26	Wang JJ , Ma YL , Zhang LB , et al. Deep learning for smart manufacturing: Methods and applications[J]. J Manuf Syst, 2018, 48: 144- 156. doi: 10.1016/j.jmsy.2018.01.003
27	Rawat W , Wang Z . Deep convolutional neural networks for image classification: a comprehensive review[J]. Neural Comput, 2017, 29 (9): 2352- 449. doi: 10.1162/neco_a_00990
28	Aggarwal CC. Convolutional Neural Networks. In: Neural Networks and Deep Learning[J]. Springer Cham, 2018. doi: 10.1007/978-3-319-94463-0_8.
29	Krizhevsky A , Sutskever I , Hinton GE . ImageNet classification with deep convolutional neural networks[J]. Commun ACM, 2017, 60 (6): 84- 90. doi: 10.1145/3065386
30	LeCun Y , Bengio Y , Hinton G . Deep learning[J]. Nature, 2015, 521 (7553): 436- 444. doi: 10.1038/nature14539
31	Benuwa B , Zhan YZ , Ghansah B , et al. A review of deep machine learning[J]. Int J Eng Res Afr, 2016, 24: 124- 136. doi: 10.4028/www.scientific.net/JERA.24.124
32	Ying X . An overview of overfitting and its solutions[J]. 2018 International conference on computer information science and application technology, 2019, 1168. doi: 10.1088/1742-6596/1168/2/022022
33	Aggarwal CC. Teaching deep learners to generalize. in: neural networks and deep learning[J]. Springer, Cham, 2018. doi: 10.1007/978-3-319-94463-0_4.
34	Zilly J , Buhmann JM , Mahapatra D . Glaucoma detection using entropy sampling and ensemble learning for automatic optic cup and disc segmentation[J]. Comput Med Imag Grap, 2017, 55: 28- 41. doi: 10.1016/j.compmedimag.2016.07.012
35	Yu S , Xiao D , Frost S , et al. Robust optic disc and cup segmentation with deep learning for Glaucoma detection[J]. Comput Med Imaging Graph, 2019, 74: 61- 71. doi: 10.1016/j.compmedimag
36	Sevastopolsky A . Optic disc and cup segmentation methods for Glaucoma detection with modification of U-Net convolutional neural network[J]. Pattern Recognit Image Anal, 2017, 27 (3): 618- 624. doi: 10.1134/S1054661817030269
37	Fu H , Cheng J , Xu Y , et al. Joint optic disc and cup segmentation based on multi-label deep network and polar transformation[J]. IEEE Trans Med Imaging, 2018, 37 (7): 1597- 1605. doi: 10.1109/TMI.2018.2791488
38	Li Z , He Y , Keel S , et al. Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs[J]. Ophthalmology, 2018, 125 (8): 1199- 1206. doi: 10.1016/j.ophtha.2018.01.023
39	Liu H , Li L , Michael Wormstone I , et al. Development and validation of a deep learning system to detect glaucomatous optic neuropathy using fundus photographs[J]. JAMA Ophthalmol, 2019, 137 (12): 1353- 1360. doi: 10.1001/jamaophthalmol.2019.3501
40	Medeiros FA , Jammal AA , Thompson AC . From machine to machine: an oct-trained deep learning algorithm for objective quantification of glaucomatous damage in fundus photographs[J]. Ophthalmology, 2019, 126 (4): 513- 521. doi: 10.1016/j.ophtha.2018.12.033
41	Thompson AC , Jammal AA , Medeiros FA . A deep learning algorithm to quantify neuroretinal rim loss from optic disc photographs[J]. Am J Ophthalmol, 2019, 201: 9- 18. doi: 10.1016/j.ajo
42	Asaoka R , Murata H , Hirasawa K , et al. Using Deep Learning and transfer learning to accurately diagnose early-onset glaucoma from macular optical coherence tomography images[J]. Am J Ophthalmol, 2019, 198: 136- 145. doi: 10.1016/j.ajo
43	Yosinski J , Clune J , Bengio Y , et al. How transferable are features in deep neural networks?[J]. Proceedings of the 27th International Conference on Neural Information Processing Systems, 2014, (2): 3320- 3328. doi: 10.5555/2969033.2969197
44	Lee J , Kim YK , Park KH , et al. Diagnosing glaucoma with spectral-domain optical coherence tomography using deep learning classifier[J]. Journal of Glaucoma, 2020, 29 (4): 287- 294. doi: 10.1097/IJG.0000000000001458
45	Thompson AC , Jammal AA , Berchuck SI , et al. Assessment of a segmentation-free deep learning algorithm for diagnosing glaucoma from optical coherence tomography scans[J]. JAMA Ophthalmol, 2020, 138 (4): 333- 339. doi: 10.1001/jamaophthalmol.2019.5983
46	Wang X , Chen H , Ran AR , et al. Towards multi-center Glaucoma OCT image screening with semi-supervised joint structure and function multi-task learning[J]. Med Image Anal, 2020, 63: 101695. doi: 10.1016/j.media
47	Maetschke S , Antony B , Ishikawa H , et al. A feature agnostic approach for glaucoma detection in OCT volumes[J]. PLoS One, 2019, 14 (7): e0219126. doi: 10.1371/journal.pone.0219126
48	Ran AR , Cheung CY , Wang X , et al. Detection of glaucomatous optic neuropathy with spectral-domain optical coherence tomography: a retrospective training and validation deep-learning analysis[J]. Lancet Digit Heal, 2019, 1 (4): 172- 182. doi: 10.1016/S2589-7500(19)30085-8
49	Badano A , Graff CG , Badal A , et al. Evaluation of digital breast tomosynthesis as replacement of full-field digital mammography using an in silico imaging trial[J]. Jama Netw Open, 2018, 1 (7): e185474. doi: 10.1001/jamanetworkopen.2018.5474
50	Cha KH , Petrick N , Pezeshk A , et al. Reducing overfitting of a deep learning breast mass detection algorithm in mammography using synthetic images[J]. Proc SPIE 10950, Medical Imaging 2019: Computer-Aided Diagnosis, 1095004.. doi: 10.1117/12.2512604
51	Goodfellow IJ , Pouget-Abadie J , Mirza M , et al. Generative Adversarial Nets[J]. Proceedings of the 27th International Conference on Neural Information Processing Systems, 2014, (2): 2672- 2680. doi: 10.5555/2969033.2969125
52	Sun Y , Zhou CF , Fu YW , et al. Parasitic Gan for Semi-Supervised Brain Tumor Segmentation[J]. Ieee Image Proc, 2019, 1535- 1539. doi: 10.1109/ICIP.2019.8803073
53	Yang Y , Nan F , Yang P , et al. GAN-Based Semi-Supervised Learning Approach for Clinical Decision Support in Health-IoT Platform[J]. IEEE Access, 2019, 7: 8048- 8057. doi: 10.1109/ACCESS
54	Wang X , Tang FY , Chen H , et al. UD-MIL: Uncertainty-driven deep multiple instance learning for OCT image classification[J]. IEEE J Biomed Health Inform, 2020, 1. doi: 10.1109/jbhi
55	Fu HZ , Li F , Sun X , et al. AGE challenge- angle closure glaucoma evaluation in anterior segment optical coherence tomography[J]. Medical Image Analysis, 2020, (66): 101798. doi: 10.1016/j.media.2020.101798
56	Mariottoni E, Datta S, Dov D, et al. Artificial intelligence mapping of structure to function in glaucoma[J]. 2020, 9(2): 19.
57	Yu S , Xiao D , Frost Shaun , et al. Robust optic disc and cup segmentation with deep learning for glaucoma detection[J]. Computerized Medical Imaging Analysis Graphics, 2019, (74): 61- 71. doi: 10.1016/j.compmedimag
58	Tan NYQ , Friedman DS , Stalmans I , et al. Glaucoma screening: where are we and where do we need to go?[J]. Curr Opin Ophthalmol, 2020, 31 (2): 91- 100. doi: 10.1097/ICU.0000000000000649

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

Comments

Recommended 10

[1]	SUO Dongyang, SHEN Fei, GUO Hao, LIU Lichang, YANG Huimin, YANG Xiangdong. Expression and mechanism of Tim-3 in animal model of drug-induced acute kidney injury[J]. Journal of Shandong University (Health Sciences), 2020, 1(7): 1 -6 .
[2]	MA Qingyuan, PU Peidong, HAN Fei, WANG Chao, ZHU Zhoujun, WANG Weishan, SHI Chenhui. Effect of miR-27b-3p regulating SMAD1 on osteosarcoma cell proliferation, migration and invasion[J]. Journal of Shandong University (Health Sciences), 2020, 1(7): 32 -37 .
[3]	LI Ning, LI Juan, XIE Yan, LI Peilong, WANG Yunshan, DU Lutao, WANG Chuanxin. Expression of LncRNA AL109955.1 in 80 cases of colorectal cancer and its effect on cell proliferation, migration and invasion[J]. Journal of Shandong University (Health Sciences), 2020, 1(7): 38 -46 .
[4]	XU Yuxiang, LIU Yudong, ZHANG Peng, DUAN Ruisheng. A retrospective analysis of risk factors of cerebral microbleeds in 101 patients with cerebral small vessel disease[J]. Journal of Shandong University (Health Sciences), 2020, 1(7): 67 -71 .
[5]	LUO Xin, HE Bing, NIE Qingsheng, HOU Zhenbo, DONG Jun, LI Yuhua, ZENG Xiangqin, LIU Wei, KONG Demin, CAO Jinfeng. Comparison of the value of mono-exponential mode and diffusion kurtosis imaging mode in grading clear cell renal cell carcinoma using magnetic resonance diffusion-weighted imaging[J]. Journal of Shandong University (Health Sciences), 2020, 1(7): 89 -95 .
[6]	Di ZHANG,Meng YU,Xia LIU. A general overview of neuromodulation[J]. Journal of Shandong University (Health Sciences), 2020, 1(8): 50 -60 .
[7]	Wei ZHANG,Wenhao TAN,Yibin LI. Locmotion control of quadruped robot based on deep reinforcement learning: review and prospect[J]. Journal of Shandong University (Health Sciences), 2020, 1(8): 61 -66 .
[8]	Anjing CHEN,Xun ZHANG. The new strategies of targeting SUMOylation in the treatment of glioma[J]. Journal of Shandong University (Health Sciences), 2020, 1(8): 88 -94 .
[9]	ZHANG Hongbin, ZHAO Hanhui, WANG Suxia, ZHOU Peng, HE Qingqing, WANG Yanqun, DING Weiping, LIU Gang. Perioperative observation and postoperative risk factors of severe hypocalcemia after parathyroidectomy: a report of 303 cases[J]. Journal of Shandong University (Health Sciences), 2020, 1(9): 14 -20 .
[10]	SUN Yifeng, GAO Yu, LIANG Yongyuan, GAO Yang. Expression of CPLX2 and its in vitro effects on the proliferation migration and invasion of hepatocellular carcinoma cells[J]. Journal of Shandong University (Health Sciences), 2020, 1(9): 34 -39 .

Artificial intelligence deep learning in glaucoma imaging: current progress and future prospect

RichHTML

PDF (PC)

Abstract

Cite this article

share this article

Figures/Tables 5

References 58

Related Articles 8

Metrics

Comments

Recommended 10

[1]	Ju LIU,Qiang WU,Luyue YU,Fengming LIN. Brain tumor image segmentation based on deep learning techniques [J]. Journal of Shandong University (Health Sciences), 2020, 1(8): 42-49, 73.
[2]	Xingang LI,Xin ZHANG,Anjing CHEN. The latest advances in human brain projects [J]. Journal of Shandong University (Health Sciences), 2020, 1(8): 5-9, 21.
[3]	Jizong ZHAO. Neurosurgery plays a key role in brain science research [J]. Journal of Shandong University (Health Sciences), 2020, 1(8): 1-4.
[4]	QU Yi, ZHANG Huankai, SONG Xian, CHU Baorui. Research progress of artificial intelligence diagnosis system in retinal diseases [J]. Journal of Shandong University (Health Sciences), 2020, 58(11): 39-44.
[5]	Zongyuan GE,Wanji HE,Lie JU,Xuan YAO,Lin WANG,Yelin HUANG,Zhiwen YANG,Jianhao XIONG,Yining BAO,Ming LI,Bing ZHANG,Xin ZHAO. New developments in ophthalmic AI algorithms [J]. Journal of Shandong University (Health Sciences), 2020, 58(11): 17-23.
[6]	Haotian LIN,Longhui LI,Jingjing CHEN. Research progress of artificial intelligence in childhood eye diseases [J]. Journal of Shandong University (Health Sciences), 2020, 58(11): 11-16.
[7]	Mingguang HE,Chi LIU,Zhixi LI. Applying artificial intelligence in ophthalmic real-world practice: opportunities and challenges [J]. Journal of Shandong University (Health Sciences), 2020, 58(11): 1-10.
[8]	王宇雷,曲毅,周芳,冯进波. A diode laserinduced choroidal neovascularization model in Brown Norway rats: an experimental study [J]. JOURNAL OF SHANDONG UNIVERSITY (HEALTH SCIENCES), 2007, 45(2): 163-167.