Cedars-Sinai investigators developed an artificial intelligence tool that accurately predicted who would develop pancreatic cancer based on what their CT scan results looked like years prior to being diagnosed with the disease. These findings, which may help prevent death through early detection of one of the most harmful cancers to treat, are published in the journal Cancer Biomarkers.
In CT scans, Debiao Li, PhD, the director of the Biomedical Imaging Research Institute, and the senior and related author of the study, was able to identify and quantify very subtle, early signs of pancreatic ductal adenocarcinoma. These are signs that the human eye would never be able to discern, according to Li, who is also the Chair of the Minimally Invasive Surgery in Honor of George Berci.
The pancreatic ductal adenocarcinoma is not only the most common type of pancreatic cancer, but it is also the most deadly. Nearly 10% of people diagnosed with the disease live five years after being diagnosed or starting treatment. However, recent research has demonstrated that finding the cancer early can increase survival rates by up to 50%. There is currently no easy way to locate pancreatic cancer early.
People with this type of cancer may experience symptoms such as general abdominal pain or unexplained weight loss, but these symptoms are often ignored or overlooked as signs of the cancer, because they are common in many health situations.
According to Stephen J. Pandol, MD, the director of the Basic and Translational Pancreas Research and program at Cedars-Sinai, and another author of the study, there are no particular symptoms that might assist with a diagnosis for pancreatic ductal adenocarcinoma. This AI tool may eventually be used to detect early illness in people who undergo CT scans for abdominal pain or other problems.
The investigators reviewed electronic medical records to identify persons who had been diagnosed with the cancer in the last 15 years, and who received CT scans six months to three years prior to their diagnosis. These CT images were considered normal at the time they were taken. The team identified 36 patients who had compliance with these criteria, most of whom had abdominal pain done in the ER.
The AI tool was developed to identify pre-diagnostic CT images from people with pancreatic cancer and compare them with CT images from 36 people who didn''t develop the cancer. The researchers found that the model was 86% accurate in identifying individuals who would eventually be discovered to have pancreatic cancer and those who would not develop the cancer.
The AI model was based on interactions between people with cancer and healthy controls on the surface of the pancreas. These textural differences may be the result of molecular changes that occur during the development of pancreatic cancer.
We believe that this tool might detect the cancer early enough to make it possible for more people to have their tumor completely removed through surgery, according to Touseef Ahmad Qureshi, the PhD, a scientist at Cedars-Sinai and the first author.
Investigators are currently collecting data from tens of thousands of patients at healthcare locations across the United States to continue to investigate the intelligence tools available.