Application of Artificial Intelligence in Infectious Diseases

Artificial intelligence (AI) is a representative cutting-edge direction in today’s technological development that integrates numerous disciplines and industrial fields and significantly impacts modern science and social production. Medical imaging is one of the most significant AI application areas, and algorithms such as imaging omics, deep learning, and transfer learning have been developed and tested on medical imaging data, forming various applications such as lesion detection, lesion segmentation, lesion characterization, treatment planning, and prognosis prediction. Currently, the training speed and work efficiency of Chinese radiologists are insufficient to deal with the rapidly growing trend of imaging data, and the enormous pressure of imaging diagnosis work and experiential judgement can easily lead to misdiagnosis and missed diagnosis. AI-assisted diagnostic technologies in medical imaging have been applied to diagnose various diseases, such as pulmonary nodules, lung tumours, COVID-19, and drug-sensitive tuberculosis [1]. AI technology can extract massive amounts of data from patient images and deeply mine, analyse, and interpret the data. It can noninvasively analyse imaging features multiple times to obtain more heterogeneous information about the lesion than the human eye. Infectious diseases are a hot research area where AI technology and imaging are combined. The computer-aided diagnostic (CAD) tuberculosis diagnostic system based on AI has progressed. Large-scale clinical trials have shown that using an automatic chest X-ray tuberculosis detection system with AI can improve tuberculosis screening and diagnosis accuracy, and a digital imaging-based tuberculosis screening system can reduce the time for radiologists to read images, greatly improving work efficiency and accuracy in tuberculosis diagnosis, which is of great significance for tuberculosis prevention and treatment.