AI Detects Pancreatic Cancer Early and Saves Lives

Researchers from Alibaba have developed a artificial intelligence (AI) tool that identifies pancreatic cancer early in routine exams, saving lives. In use since 2024, the technology has shown effectiveness in detecting signs of the disease that are often overlooked. The AI system, known as PANDA, has been utilized by the Ningbo University Affiliated Popular Hospital in China.

According to The New York Times, so far, the AI has analyzed over 180,000 CT scans of the abdomen and chest. The use of this tool has resulted in faster and more accurate diagnoses.

Faster and More Accurate Diagnoses

Since its implementation, the technology has detected over 20 cases of pancreatic cancer, including 14 in the early stage. The system is able to identify both the common and aggressive form of the disease, ductal adenocarcinoma, as well as the rare and less deadly neuroendocrine tumor.

• All diagnosed patients presented symptoms such as swelling or nausea and had not consulted a specialist before being alerted by the AI.
• The success of PANDA is due to its training with contrast-enhanced CT scans from over 2,000 patients who tested positive, detailing the location of lesions in the pancreas.
• Furthermore, when tested on non-contrast CT scans, the AI successfully identified 93% of cases of pancreatic lesions.
• At the hospital, doctors use the tool to review exams classified as high risk and, as needed, call patients for more detailed tests.

Despite the initial success, experts emphasize that more data from real cases is necessary to evaluate whether the AI can accurately identify early-stage cases to minimize false positive risks. Other countries are also developing similar resources.

Classified as an “innovative device” by the U.S. Food and Drug Administration (FDA), PANDA may have its market availability expedited, according to information from Alibaba.

Pancreatic cancer, which has high mortality rates, is difficult to diagnose in early stages. Symptoms include abdominal pain, loss of appetite and weight, weakness, and jaundice. Treatment may involve surgery, chemotherapy, and radiation therapy.

This innovation could redefine the diagnostic approach to pancreatic cancer, allowing for improvements in survival rates and changes in future disease treatment.