Ultra-Widefield Fundus Imaging for Diabetic Retinopathy

Diabetic retinopathy (DR), a common and specific complication of diabetes mellitus, is one of the leading causes of preventable blindness among working-aged people [1]. An estimated 103 million adults worldwide were affected by DR in 2020, and the number of people with DR is projected to rise to 130 million in 2030 and 161 million in 2045 [2]. DR can be classified into five categories: no apparent retinopathy, mild non-proliferative diabetic retinopathy (NPDR), moderate NPDR, severe NPDR, and proliferative diabetic retinopathy (PDR) according to the International Clinical Diabetic Retinopathy (ICDR) Severity Scale [3]. In addition, referable DR (RDR) was defined as moderate NPDR or worse, including diabetic macular edema (DME) [4]. Prompt screening, timely referral, and early treatment are widely accepted as consensus for preventing visual loss [5]. Standard colour fundus photography (CFP) that captures macular and optic nerve with a field view of 30 to 50-degree is the gold standard photograph method for the detection of DR [6], and deep learning methods for detection of DR using CFP have gradually matured [7, 8]. However, ultra-widefield (UWF) fundus images, which have a wide range up to 200-degree view of the retina, have emerged as an alternative photograph method for DR management, which not only has agreement with the standard 7-field of the Early Treatment Diabetic Retinopathy Study photos but decreases the rate of ungradable images compared to CFP [9, 10]. Furthermore, UWF fundus images allow the identification of predominantly peripheral lesions (PPL), presenting in 30%-40% of eyes with DR and suggesting a more severe DR level in 11% of eyes [11]. However, the classification of UWF fundus images is time-consuming and labor-intensive, requiring significant effort from human graders, and studies of using computer-aided system for effective analysis of UWF fundus images are limited. Aiming to advance the state-of-the-art in automatic DR analysis from UWF fundus images, we organize the ultra-widefield fundus imaging for diabetic retinopathy challenge. The challenge encourages researchers to develop algorithms for different tasks in DR analysis using UWF fundus images regarding UWF fundus images quality assessment, classification of DRD, and the classification of DME. On the one hand, the images quality assessment task ensures that the images used for classification are of sufficient quality. On the other hand, the classification of DR and DME provide the foundation for automatic analysis that can help the management of DR patients. This challenge serves as an important milestone in the analysis of DR using ultra-widefield images. We hope that the challenge will drive innovation in automatic medical image analysis, propelling advancements in the field.

References
1. Cheung, N., P. Mitchell, and T.Y. Wong, Diabetic retinopathy. Lancet, 2010. 376(9735): p. 124-36.
2. Teo, Z.L., et al., Global Prevalence of Diabetic Retinopathy and Projection of Burden through 2045: Systematic Review and Meta-analysis. Ophthalmology, 2021. 128(11): p. 1580-1591.
3. Wilkinson, C.P., et al., Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology, 2003. 110(9): p. 1677-82.
4. Bellemo, V., et al., Artificial intelligence using deep learning to screen for referable and vision-threatening diabetic retinopathy in Africa: a clinical validation study. Lancet Digit Health, 2019. 1(1): p. e35-e44. 
5. Ting, D.S., G.C. Cheung, and T.Y. Wong, Diabetic retinopathy: global prevalence, major risk factors, screening practices and public health challenges: a review. Clin Exp Ophthalmol, 2016. 44(4): p. 260-77.
6. Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology, 1991. 98(5 Suppl): p. 786-806.
7. Ting, D.S.W., 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. Jama, 2017. 318(22): p. 2211-2223.
8. Lin, D., et al., Application of Comprehensive Artificial intelligence Retinal Expert (CARE) system: a national real-world evidence study. Lancet Digit Health, 2021. 3(8): p. e486-e495.
9. Sun, J.K. and L.P. Aiello, The Future of Ultrawide Field Imaging for Diabetic Retinopathy: Pondering the Retinal Periphery. JAMA Ophthalmology, 2016. 134(3): p. 247-248.
10. Silva, P.S., et al., Identification of Diabetic Retinopathy and Ungradable Image Rate with Ultrawide Field Imaging in a National Teleophthalmology Program. Ophthalmology, 2016. 123(6): p. 1360-7.
11. Aiello, L.P., et al., Comparison of Early Treatment Diabetic Retinopathy Study Standard 7-Field Imaging With Ultrawide-Field Imaging for Determining Severity of Diabetic Retinopathy. JAMA Ophthalmology, 2019. 137(1): p. 65-73.