- The AI model was developed from an analysis of 500 CT scans of people with lung nodules, indicating the spread of pancreatic cancer to the lungs.
- Os researchers identified the people most likely to be diagnosed with pancreatic cancer based on medical records and analysis.
- The model was tested on 6 million people in Denmark and 3 million in the US, yielding an AUC score of 0,88 for estimating risk over the next three years and 0,9 for detecting risk over the next 12 months.
- The AI model can also be used to predict the need for additional intervention in patients with lung nodules who would be considered an average risk of developing cancer.
- The use of AI can help maximize the use of diagnostic tests and improve clinical decision-making, crucial for early disease screening.
- According to the American Society of Clinical Oncology, 56% of people with pancreatic cancer die within five years of diagnosis.
- In Brazil, according to the National Cancer Institute, 1% of all types of cancer diagnosed are pancreatic. Still according to the INCA, 5% of the total deaths caused by the disease are from pancreatic cancer.
See also: