In 2021, over 60,000 individuals in the U.S. alone will be diagnosed with pancreatic cancer. Researchers are turning to machine learning (ML) to help identify those most at risk of developing the disease, such that they can access early interventions.
In the study, published in PLOS ONE, researchers describe how they analyzed health records from over 1,000 patients in the UK, up to two years before their pancreatic cancer diagnosis. They used this data to train an ML algorithm to distinguish patients at risk.
The ML system found 41 percent of the patients to be in the high-risk category, up to 20 months before they receive the diagnosis. The test had a 72 percent sensitivity, meaning the ability to successfully identify those who went on to develop pancreatic cancer. 59 percent of people who did not end up getting pancreatic cancer were categorized as low risk by the algorithm. The specificity and sensitivity of the system can be refined and improved with additional work, say the authors.
"Each year, 460,000 people worldwide are diagnosed with pancreatic cancer, and only around 5% of those diagnosed survive for five years or more,” explained co-lead author Ananya Malhotra. “This low survival is because patients are usually diagnosed very late. Recent progress has been made in identifying biomarkers in the blood and urine, but these tests cannot be used for population screening as they would be very expensive and potentially harmful due to the psychological distress of excess testing.
"Although preliminary, this study offers some hope for a new early diagnosis for pancreatic cancer which until now remains elusive."