Breakthrough AI Tool Accurately Detects Pancreatic Cancer Using Non-Contrast CT Scans

Pancreatic ductal adenocarcinoma (PDAC), the most common type of pancreatic cancer, accounts for nearly half a million deaths worldwide annually. Diagnosing this cancer at an early stage is notoriously difficult, as symptoms typically don't appear until the cancer has grown significantly and begun to spread. However, early detection can lead to potentially curative treatments, with up to 10% of early-diagnosed patients recovering fully after treatment. Now, a new AI-powered tool offers a breakthrough in detecting early signs of pancreatic cancer by enhancing and identifying subtle pathological features in standard CT scans.

This innovative AI platform, developed by DAMO Academy (Hangzhou, China), has been named PANDA (Pancreatic Cancer Detection with Artificial Intelligence). It demonstrates a remarkable capacity for identifying and classifying pancreatic lesions. PANDA employs deep learning algorithms to detect cancerous growths within the pancreas through non-contrast CT scans. This type of scan is globally preferred due to its lower radiation dose compared to contrast-enhanced scans. PANDA's initial training involved a dataset from 3,208 patients from a single medical center, making it a promising tool for early PDAC detection, potentially extending the median overall survival rate from approximately 1.5 years for late-stage cases to close to ten years for early detections.

In a comparative study, PANDA proved to be 34.1% more sensitive than human radiologists in identifying abnormalities from screening scans. The tool underwent validation in a real-world clinical environment, screening over 20,000 patients. During these screenings, PANDA identified pancreatic cancer-related pathological changes in 31 patients that were initially overlooked by doctors. Besides its heightened sensitivity, PANDA has been used clinically over 500,000 times, showing an impressively low false-positive rate of only one in every thousand tests, outperforming radiologists by 6.3%. This advanced AI tool holds the potential to revolutionize large-scale pancreatic cancer screening, possibly being incorporated into regular health check-ups or emergency department visits. Furthermore, researchers believe that PANDA might eventually be capable of detecting other cancer types as well.

“AI plus non-contrast CT technology hold the promise to be an effective and cost-efficient tool to achieve detection of pancreatic cancer in the early stages and make large-scale pancreatic cancer screening possible to prevent loss of lives,” said Le Lu, Head of Damo Academy’s medical AI team.

Related Links:
DAMO Academy