Advances in Artificial Intelligence could detect pancreatic cancer

The progress of Artificial Intelligence (AI) has brought new benefits for health research and early detection of diseases in recent years.

The European Society of Medical Oncology (ESMO) organized the World Congress on Gastrointestinal Cancer, which would take place in early July. However, because of the current health situation in the world it was postponed and will offer virtual days between July 16 and August 27.

However, this conference will present the progress of a study promising the early detection of pancreatic cancer. Pancreatic cancer is a disease that affects 12 in 100,000 people, meaning that a method of screening through testing would not be efficient. Other cancers may have tests or tests that determine a possible disease, because in some cancers they exist in populations with more risk to develop the disease. By age group or by gender, for example, breast cancer and prostate cancer.

For this preliminary study, electronic health records (EHR), used by general practitioners of patients in the UK, were supported. The analysis included 1,378 patients aged 15 to 99 diagnosed with pancreatic cancer between 2005 and 2010. In addition, each patient was paired by age and sex with four other people without a cancer diagnosis.

At that stage is where AI was the most important part, since information on symptoms, diseases, medications in the two years prior to diagnosis and thus develop a model predicting who might develop cancer. The pilot study found that the model can predict which people under the age of 60 were most at risk of getting sick and make a diagnosis more than 20 months before developing the disease.

“Our model has estimated that around 1,500 tests need to be performed to save one life from pancreatic cancer” explained Dr. Ananya Malhotra researcher in statistics, London School of Hygiene &Tropical Medicine, London, UK. In addition, he explained that they used AI to study an extensive volume of data and find exact combinations to predict who might develop the disease. He further stated that: “It’s not possible for the human eye to recognize these trends in such large amounts of data”.

In this way, the combination of traditional non-invasive techniques and modern Digital Health techniques such as Artificial Intelligence or EHR can achieve a more than positive result for early diagnosis of diseases where time is a key factor.

Related Content