デジタルツイン・AIモデルと治療予測

抄録

Digital twins are a technology that uses data collected from real–world objects and systems to recreate identical conditions in a virtual space. In the medical field, digital twins, also known as “human digital twins,” digitize and recreate individual patients' physiological conditions, lifestyle, and medical information in a virtual space. These digital twins are expected to realize the ultimate in personalized medicine, including ultra–personalized/precision medicine, disease onset prediction, and optimal drug and treatment recommendations, with high effectiveness and minimal side effects. Digital twins are also being explored in the field of neuroscience, including the treatment of Alzheimer disease (AD) and Parkinson disease (PD). Building digital twins that reflect a patient's lifestyle and physiological status makes it possible to observe and predict how specific lifestyle habits (e.g., exercise, diet) and behavioral patterns affect a patient's cognitive function and condition. Furthermore, using digital twins to model the progression of brain atrophy and changes in functional connectivity over time can help predict future cognitive decline and the risk of worsening disease. Digital twins utilizing digital devices and AI will not only enable objective evaluation, but will also enable quantitative evaluation of gait patterns specific to PD (such as shuffling gait) and distinguish PD from other diseases. These advances are based on the technology of real–time, data–driven simulation of individual brain function, which is the core value of digital twins, and are expected to contribute to the elucidation, prediction, and development of personalized treatments for diseases such as AD and PD. This chapter will consider the possibilities and challenges of digital twins through recent project examples from Japan and abroad in the field of neurological disorders.