Major Breakthroughs in Dementia Made Possible by AI

As society ages, the issue of dementia is becoming increasingly prominent, emerging as a major social concern. According to statistics from the Ministry of Health and Welfare, Taiwan's dementia population is continuously rising, exceeding 300,000 by 2022, with 96% of patients being seniors aged 65 or older. It is predicted that by 2030, the dementia population will increase to over 500,000.

Dementia is a progressively chronic and debilitating disease, primarily characterized by intellectual decline, cognitive impairment, and behavioral and emotional changes. These symptoms severely impact patients' daily lives, making self-care difficult and requiring long-term care, and imposing immense psychological and life pressures on both patients and their families.

Current challenges in dementia treatment involve difficulties in early detection and pharmaceutical intervention. As dementia predominantly affects individuals over 65, when symptoms like forgetfulness and memory impairment appear, it is difficult to determine if they are due to normal brain aging or pathological degeneration. This complicates early detection, and by the time dementia is diagnosed and medication begins, the effectiveness of drugs is already very limited, only able to slightly slow down the rate of brain degeneration without being able to cure the symptoms of dementia.

With the support of the National Science and Technology Council (NSTC), Taipei Medical University has developed the world's first AI model capable of predicting brain age and dementia risk. This model can help high-risk seniors plan their care proactively, while also assisting pharmaceutical companies in identifying more precise candidates for clinical trials, thereby enhancing the effectiveness of dementia treatment.

It is hoped that by using AI to predict dementia risk, patients, caregivers, and doctors can prepare in advance, develop effective treatment plans, and research better medications, thereby improving the effectiveness of dementia treatment while reducing societal costs.

The research team integrated dementia databases from the United States, UK Biobank, and Japan, along with data from four major medical centers: National Taiwan University Hospital, National Cheng Kung University Hospital, Taipei Veterans General Hospital, and Taipei Medical University Hospital System, and launched an AI dementia prediction module. This module can determine brain age and estimate the likelihood of developing dementia within two years based on factors such as CT scans, age, gender, education level, socioeconomic status, and genetics. It also analyzes the degree of brain degeneration and automatically generates a text report to assist doctors in understanding the patient's condition.

The purpose of this module is to assess whether the rate of brain aging matches the actual chronological age by predicting brain age. If the brain age is greater than the actual age, it suggests that the brain may be aging too rapidly, potentially leading to future dementia development. Therefore, this function can help doctors and families recognize potential risks earlier and take appropriate preventive measures.

Based on known research findings, predictions are made for individuals already exhibiting mild cognitive impairment, estimating their likelihood of developing dementia within the next two years. Studies show that individuals with such cognitive deficits have a 50% chance of developing dementia within one year, hence, further prediction of their dementia risk is necessary.

Finally, by analyzing the areas of brain atrophy, the module can determine the type of dementia. Existing clinical types of dementia include four major categories: Alzheimer's disease, frontotemporal dementia, Lewy body dementia, and vascular dementia. Through AI analysis, the areas and patterns of brain atrophy can be identified (Patten), assisting doctors in early determination of the patient's specific dementia type.