Diabetic retinopathy (DR), one of the most devastating manifestations of diabetes, is a leading cause of blindness among working-age adults. World Health Organization predicts the prevalence of diabetes to increase substantially in the future, leading to an increasing pressure on public health services. In the context of smart healthcare, DR screening has been widely adopted by utilizing fundus imaging with manual input. In this chapter, we review the potential of artificial intelligence enabled automated screening for the detection and classification of DR in the context of the National Health Service. We propose an integrated multimodal approach to enable the combination of manual (through human graders) and automated DR screening. Furthermore, we discuss how this multimodal approach can be enhanced by the integration of additional modalities such as optical coherence tomography. Artificial intelligence in that setting can complement and upskill human graders by acting in an assistive way rather than replacing their role.
|Title of host publication||Innovation in Health Informatics|
|Subtitle of host publication||A Smart Healthcare Primer|
|Number of pages||18|
|Publication status||Published - 2020|