An artificial intelligence system is helping researchers inch towards being able to detect pancreatic cancer in its earliest stages.
That’s according to a new study published in Cancer Biomarkers that used the system to retrospectively analyze 108 computed tomography scans for subtle imaging alterations that could eventually develop into pancreatic ductal adenocarcinoma (PDAC). Using those image features, the AI system was able to achieve an average classification accuracy of 86% when predicting which patients would go on to develop pancreatic cancer. Experts involved in the study suggested that their findings could eventually be used to detect pancreatic cancer in its earliest stages when patients are most likely to respond to interventions favorably.
“This AI tool was able to capture and quantify very subtle, early signs of pancreatic ductal adenocarcinoma in CT scans years before occurrence of the disease. These are signs that the human eye would never be able to discern,” corresponding author Debiao Li, PhD, director of the Biomedical Imaging Research Institute and professor of Biomedical Sciences and Imaging at Cedars-Sinai, and co-authors explained.
PDAC is not only the most common form of pancreatic cancer, but it’s also the deadliest. With a 5-year survival rate of just 10%, it is currently the 4th leading cause of cancer deaths in both men and women. Research has shown that detecting PDAC in its earliest stages could significantly increase survival rates, but since its symptoms are typically very generalized (abdominal pain, weight loss, etc.) they are frequently overlooked or dismissed.
“There are no unique symptoms that can provide an early diagnosis for pancreatic ductal adenocarcinoma,” said Stephen J. Pandol, MD, program director of the Gastroenterology Fellowship Program at Cedars-Sinai, and another author of the study. “This AI tool may eventually be used to detect early disease in people undergoing CT scans for abdominal pain or other issues.”
Using medical records that dated back 15 years, researchers at Cedars-Sinai identified individuals who had been diagnosed with PDAC and had also undergone CT scans six months to three years prior to their cancer diagnosis. This resulted in 36 patients, each of whom had scans that were considered normal at the time of completion. Their scans, along with the imaging from a control group of individuals who did not go on to develop cancer, were used to train the model to differentiate the subtle changes that would later be used to indicate whether a person would be diagnosed with pancreatic cancer.
The AI system was able to identify textural differences on the surface of the pancreas in those who were eventually diagnosed with cancer. This led to an accuracy of 86% in distinguishing between those who would and would not go on to develop PDAC.
“Our hope is this tool could catch the cancer early enough to make it possible for more people to have their tumor completely removed through surgery,” said first author Touseef Ahmad Qureshi, PhD, a scientist at Cedars-Sinai.
Related artificial intelligence content:
AI-based mammo screening protocol reduces radiologist workload by 62%
Explainable AI model accurately auto-labels chest X-rays from open access datasets
Algorithm performs at expert level when distinguishing between benign and malignant ovarian tumors
Lack of diverse datasets in AI research puts patients at risk, experts suggest
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She joined Innovate Healthcare in 2021 and has since put her unique expertise to use in her editorial role with Health Imaging.