Ai model detects pancreatic cancer over a year before diagnosis
An artificial intelligence system can identify early signs of pancreatic cancer on routine scans more than a year before clinical diagnosis, according to new research. The model, known as REDMOD, analyzes subtle tissue patterns that standard imaging and human observation fail to detect, enabling earlier identification of one of the deadliest cancers.
Developed by researchers at Mayo Clinic and collaborators, REDMOD detected what scientists describe as a preclinical signature of pancreatic ductal adenocarcinoma an average of 475 days before diagnosis. The system relies on radiomics, extracting quantitative data from medical images to reveal microscopic structural changes. These changes remain invisible to radiologists using conventional methods.
The model outperformed experienced radiologists in multiple tests. It showed a sensitivity rate of 73 percent compared with 39 percent for human experts when identifying early-stage disease. In cases detected more than two years before diagnosis, the gap widened further, with REDMOD achieving 68 percent accuracy versus 23 percent. The system also demonstrated reliability across independent datasets, correctly identifying more than 81 percent of cancer-free scans in a group of 539 patients and 87.5 percent in a public dataset from the National Institutes of Health.
Pancreatic cancer remains the most lethal cancer in the United States, with a five-year survival rate of 13 percent. It is the third leading cause of cancer-related deaths, with an estimated 67,530 new cases and 52,740 deaths expected in 2026. Survival improves significantly when the disease is detected early, rising to 44 percent for localized cases. This gap underscores the potential impact of earlier detection enabled by AI systems such as REDMOD.
The framework includes automated pancreas segmentation, removing the need for manual outlining by radiologists, a process that is time-consuming and inconsistent. Researchers caution that further validation is required, particularly in high-risk patients such as those with unexplained weight loss or newly diagnosed diabetes. They describe the model as a step toward shifting pancreatic cancer care from late-stage diagnosis to proactive early detection.
-
14:10
-
14:06
-
13:47
-
13:30
-
13:15
-
13:00
-
12:45
-
12:40
-
12:30
-
12:15
-
12:00
-
11:42
-
11:40
-
11:25
-
11:11
-
10:47
-
10:32
-
10:21
-
10:15
-
10:00
-
09:45
-
09:30
-
09:15
-
09:02
-
09:00
-
08:45
-
08:30
-
08:15
-
23:00
-
22:30
-
22:00
-
21:00
-
20:30
-
20:00
-
19:00
-
18:55
-
18:40
-
18:25
-
18:10
-
17:55
-
17:40
-
17:25
-
17:10
-
16:55
-
16:40
-
16:25
-
16:10
-
15:55
-
15:40
-
15:25
-
15:10
-
14:55
-
14:40
-
14:25
