Deep Learning for Personalized Medicine and Precision Healthcare

Abstract

The combination of artificial intelligence (AI) and precision medicine could potentially change the way medicine is practiced. In this scenario, advances in precision medicine technology will enable healthcare providers to identify the phenotype of each patient who has a unique β-cell response to treatment, as well as other clinical requirements related to their disease. Using AI, sophisticated computers and techniques can help generate insights and create intelligent systems, thereby giving increased decision-making abilities to healthcare professionals. The latest scientific literature shows that as translational research into this combination of precision medicine and AI becomes more widespread, many of the challenges of implementing precision medicine might be addressed. Specifically, this means considering nongenomic, as well as genomic, factors; patient symptoms; clinical history; and lifestyle data in the complete evaluation and management of each patient, which will provide greater diagnostic accuracy and prognostic ability. One of the important causes of diabetic retinopathy (DR) is a common complication of diabetes and has affected many millions of people with diabetes throughout the world. DR, like many complications of diabetes is progressive and typically goes undetected until it reaches a stage of nonpreventable blindness. This chapter describes an investigation of the implementation of state-of-the-art deep learning methods in the detection and classification of diabetic retinopathy (DR). This chapter ends with some promising further research topics, including explainable artificial intelligence (AI) to strengthen trust as a mechanism, federated learning as a mechanism to facilitate data privacy, and edge AI as a mechanism to implement rapid on-site screening in underserved healthcare domains.