Photo: snapshot from CCTV News.

Early disease screening powered by artificial intelligence (AI) technology has become possible. Leveraging groundbreaking research achievements in human health and disease proteome atlases,MKsport a research team from the Huashan Hospital affiliated to Fudan University has developed a method of using AI to help doctors detect risks for hundreds of diseases, including Alzheimer's disease.

"For example, under the support of AI, for the first time globally we can predict a patient's risk of having Alzheimer's disease up to 15 years in advance through a blood test with the diagnostic accuracy exceeding 90 percent," Deng Yueting, lead author of the study that was published in the journal Cell in January, told the Global Times on Tuesday.

Deng said that the biggest advantage of the disease screening method is that it can detect diseases at an early stage and at low cost. "We are currently developing a rapid detection kit that can allow routine health checkups in the future to include a protein screening test costing approximately dozens of yuan (less than $13) for a major disease risk assessment. This would be as convenient as current blood glucose or blood pressure measurements. Middle-aged and elderly people who need regular health monitoring will find it easy to use," Deng said.

According to Deng, the study comprehensively mapped human health and disease proteomes and combined AI big data analysis methods to construct disease diagnosis prediction models, leading to the discovery of 26 new targets for drug therapy. 

Deng said that AI played a pivotal role in addressing the colossal data challenges arising from cross-correlations between thousands of diseases and proteins. AI-powered algorithmic models enabled targeted screening through vast protein datasets to identify potential biomarkers for specific diseases, while also facilitating the establishment of predictive and diagnostic frameworks.

After conducting an in-depth analysis, Deng and her teammates demonstrated that over 650 proteins are linked to at least 50 diseases, while more than 1,000 proteins exhibit sex- and age-related heterogeneity. The landmark achievement uncovered group-specific variations in disease susceptibility, providing a scientific foundation for precision diagnosis and treatment, according to Deng.

"Taking Alzheimer's disease as an example, less than 1 percent of clinical cases are caused by specific genetic mutations that almost invariably lead to onset. However, approximately 99 percent of patients cannot be fully explained by genetic mutation analysis," Deng said.

In such cases, proteins can provide critical clues about disease pathways, as proteins directly reflect the body's biological processes and pathological changes, making them crucial for understanding disease mechanisms and developing new therapies, she noted.

Deng said that the research team will enhance their database by integrating multi-omics data including metabolomic and genomic information, aiming to optimize algorithmic models to process higher-dimensional data for more accurate disease prediction.

Additionally, researchers are encouraged to validate their discoveries, facilitate clinical translation of findings, and explore collaborations with pharmaceutical companies to assess therapeutic potential, particularly for proteins that show promise as drug targets.