Early diagnosis is one of the strongest weapons you can use in the fight against cancer, especially pancreatic cancer nestled deep in the abdomen is difficult to detect.
I MIT researchers Computer Science and Artificial Intelligence Laboratory (CSAIL), together with Limor Appelbaum, a scientist in the Department of Radiation Oncology at Beth Israel Deaconess Medical Center (BIDMC) have trained artificial intelligence to diagnose pancreatic cancer early.
Early diagnosis of pancreatic cancer thanks to AI
Pancreatic cancer is one of the deadliest forms of cancer, with a five-year survival rate of less than 10%. Often, the disease is diagnosed in advanced stages, when treatment options are limited. However, researchers at MIT have developed artificial intelligence (AI) models that could help detect pancreatic cancer early and accurately.
The models are based on deep learning techniques, which allow you to analyze large amounts of biomedical data and extract relevant information. In particular, the researchers used data from blood tests and pancreas scans of healthy and sick patients. The models were trained to recognize hallmark signs of pancreatic cancer, such as changes in the levels of certain biomarkers in the blood or abnormalities in the structure of the pancreas.
The researchers fed a huge amount of data to two machine learning models, as many as 5 million patient records in the United States, which allowed them to build a neural network called PRISM for the early diagnosis of pancreatic ductal adenocarcinoma (PDAC).
“This large pool of data helped ensure the reliability and generalizability of the models, making them applicable to a wide range of populations, geographic locations, and demographic groups“, the researchers write.
The results were surprising: the models demonstrated a sensitivity significantly higher than current diagnosis methods which stop at 10% of cases, while the PRISM neural network was able to detect pancreatic cancer in 35% of the cases analyzed. Furthermore, the models were able to identify pancreatic cancer at early stages, when the disease is still operable and treatable.
“This report outlines a powerful approach to using big data and artificial intelligence algorithms to refine our approach to identifying cancer risk profiles“says David Avigan, a Harvard Medical School professor and director of the cancer center and chief of the department of hematology and hematologic malignancies at BIDMC, who was not involved in the study. “This approach may lead to new strategies to identify patients at high risk of malignancy who may benefit from targeted screening with the potential for early intervention”.
The use of artificial intelligence is nothing new, it is already used to detect cancer risk. The algorithms can analyze mammograms, CT scans for lung cancer, and help analyze Pap smears and HPV tests.
The development of the PRISM neural network began more than six years ago in search of methods of early diagnosis of pancreatic cancer which appears to be one of the main reasons for the high mortality of this type of cancer diagnosed, unfortunately, too late, even in ‘80% of cases.
The researchers hope that their machine learning models could be used as screening tools for pancreatic cancer in the future, in combination with other diagnostic techniques. This could increase the likelihood of detecting the disease early and offering patients the best possible treatment options.