Korean researchers have developed an artificial intelligence model that can determine early liver metastasis in pancreatic cancer patients using a simple blood test.
A research team led by Prof. Lee Hee-seung of the Department of Gastroenterology at Severance Hospital said Sunday that they developed an AI model that predicts the risk of early liver metastasis based on blood test results of pancreatic cancer patients and confirmed its clinical applicability.
Pancreatic cancer remains one of the deadliest cancers, with a five-year survival rate of just 17.0% despite advances in diagnosis and treatment. Because the pancreas is located deep in the abdomen and surrounded by other organs, cancer in the pancreas rarely produces early symptoms. As a result, a significant number of patients are diagnosed only after the cancer has already spread to other organs. Even when symptoms appear, they are difficult to distinguish from other gastrointestinal disorders, making early detection extremely challenging. The critical issue is that survival rates vary dramatically depending on the stage at diagnosis. For patients whose cancer is confined to the pancreas with no metastasis, the five-year survival rate reaches 47.8%. However, when the cancer has spread to distant organs such as the lungs, peritoneum or liver, the rate plummets to 2.4%. Surgery is the only treatment that offers a potential cure for pancreatic cancer, but only about 20% of patients are eligible for curative surgery.
