AIM in Endocrinology

This chapter focuses on applications of artificial intelligence (AI) in endocrinology. Endocrinology is the field of medicine that relates to the endocrine system that is consisted of endocrine glands, hormones, target organs, and feedback loop to maintain metabolic homeostasis. Endocrinology covers a broad range of health-related issues from common diseases such as diabetes mellitus, osteoporosis, and hypothyroidism to the rare diseases such as Cushing disease and acromegaly. The principle of endocrine system is based on feedback loop mediated by hormones, which makes it ideal to deploy AI system by both mechanistic (deductive) and statistical (inductive) modeling. In this chapter, we will review the current AI applications in major domains of endocrinology including diabetes mellitus, bone and mineral disorders, thyroid disorders, and pituitary and adrenal disorders. Each domain has unique tasks that may be improved by AI application. Supervised learning is the most commonly used algorithm, with increasing trend to apply unsupervised or reinforcement learning. To apply AI in screening, diagnosis, risk prediction, and treatment decision of endocrine disorders, various data sources (electronic medical records, medical images, laboratory tests) are currently being used in combination or separately. Finally, we conclude with a perspective for endocrinologists regarding current AI applications in endocrinology fields.