A new artificial intelligence (AI) model was able to detect 13 types of cancer with 98.2% accuracy using only DNA data from tissue samples, according to a new study. The AI model, called EMethylNET, was developed by researchers at the University of Cambridge in the UK and has the potential to accelerate early cancer detection, diagnosis and treatment.
The findings, released last week, Biological methods and protocolsfocused on DNA methylation, a chemical reaction that occurs early when cells, including cancer cells, begin to grow. The researchers trained a machine learning model to find the structures and pathways of cancer that form early on.
“Cancer, a collective term for more than 200 diseases, remains a leading cause of morbidity and mortality worldwide,” the study noted. “Metastatic cancer is usually detected at an advanced stage of the disease and accounts for 90 percent of cancer-related deaths.”
“Early cancer detection, combined with current treatments, will therefore have a significant impact on survival and treatment across a range of cancer types,” the report continues.
The researchers trained EMethylNET using data from more than 6,000 tissue samples from The Cancer Genome Atlas, representing 13 types of cancer, including breast, lung and colorectal cancer, and then tested it on more than 900 samples from an independent dataset.
The most notable finding was that the method showed over 98% accuracy in classifying 13 types of cancer and non-cancer samples.The study looked at a cross-section of diverse datasets from different countries. 3,388 methylation sites linked to cancer-related genes and pathways.
According to the study, the AI model combines two AI approaches – XGBoost to select relevant features and deep neural networks for classification – which not only enables accurate cancer detection but also provides insight into the body’s regulation of non-genetic factors that mutate healthy cells into cancer cells.
“these Epigenetic The modification is one of the earliest Neoplastic Related Events Carcinogenesis“,” the study noted, highlighting the approach’s potential for early cancer detection.
Although this early study is promising, the authors caution that further research and testing is needed before the technology can be used clinically. The team said they are currently working on adapting the model to liquid tissue samples, potentially enabling non-invasive early cancer testing.
“Depending on the availability of training data, the method can be scaled to detect hundreds of cancer types,” the report claims.
As AI continues to make inroads in healthcare, EMethylNET marks a major step toward leveraging machine learning for earlier and more accurate cancer diagnoses — innovations like this could have far-reaching implications for public health.
More than 19 million new cases of cancer are diagnosed each year, and 10 million people die from cancer. According to the latest estimates From the International Agency for Research on Cancer.
The lead researcher did not respond to a request for comment. Decrypt.
Editor: Ryan Ozawa.