Doctors working to confirm mesothelioma diagnoses are often challenged by the rarity of the disease and the diverse ways that it can manifest. But a study published in the journal Nature Communications details work done by a group of Scottish researchers, which offers a powerful new diagnostic tool created using artificial intelligence.
Artificial Intelligence Used to Aid in Mesothelioma Diagnoses.
Diagnosing mesothelioma is difficult for many reasons. The disease is highly aggressive, poorly understood from a biological perspective, and presents itself in extremely complex ways that can challenge even the most skilled specialists. Early detection is critical for effective treatment, but diagnosing mesothelioma from stained images is both time-intensive and extremely subjective, especially because of a lack of standards and protocols for dealing with differing and sometimes conflicting characteristics seen in patients’ tissue samples.
To address this problem, a group of researchers from the School of Cancer Sciences, University of Glasgow, the Cancer Research UK Scotland Institute, and the Pathology Department, Queen Elizabeth University Hospital, NHS Greater Glasgow and Clyde, in Glasgow, Scotland, analyzed 3,446 whole-slide images from resected tumor slides to create a comprehensive histomorphological atlas of the disease. The result was an identification of 16 recurrent patterns that matched well with later outcomes and which were compiled into an atlas that will offer an unprecedented ability to advance both clinical diagnosis and scientific understanding of mesothelioma’s biology.
Mesothelioma Atlas Created Through Advanced Machine Learning
Mesothelioma researchers took advantage of Artificial Intelligence learning methods to create the atlas. They comprehensively described the disease’s morphological complexities from thousands of digitized pathology slides to generate a quantitative visual dictionary. The approach allowed the AI to systematically analyze patterns across a massive dataset of whole-slide images, identifying features that might not be apparent through traditional manual examination methods.
This is just one example of how mesothelioma diagnosis and scientific discovery can benefit from artificial intelligence and its ability to achieve high performance in predicting outcomes and accurately subtyping the disease based on morphological patterns. The results demonstrate the atlas’s usefulness for assessing patient prognosis and its accuracy in distinguishing different forms of the disease. Beyond its immediate usefulness for diagnosis, the atlas furthers scientists’ understanding of mesothelioma biology, revealing what lies beneath its diverse patterns and offers researchers new avenues for investigating the biological mechanisms driving this rare and aggressive asbestos-caused cancer. The scientists hope that it will help identify novel therapeutic targets to help achieve better outcomes and eventually a cure.
If you or someone you love has been diagnosed with mesothelioma, the Patient Advocates at Mesothelioma.net are here to help. Contact us today at 1-800-692-8608 to learn more.
