High-throughput proteomics and machine learning helped uncover new protein signatures that could one day help with the diagnosis and treatment of systemic mastocytosis (SM), according to recently published research to be presented at the upcoming American Society of Hematology Annual Meeting and Exposition.
SM is characterized by mast cells building up in tissues throughout the body, causing a wide range of symptoms including rashes, bone pain and organ infiltration and damage. Because how the disease manifests differs greatly from one person to another, there is a need for reliable ways to predict how each patient’s illness will progress or how severe it will get. Therefore, finding new prognostic tools is a priority for researchers.
In this context the authors combined high-throughput proteomics and machine learning to look for new ways to assess and predict SM. Proteomics is a technique that allows researchers to measure hundreds of different proteins in a person’s blood at the same time. Proteins are essential for nearly every function in the body, and changes in their levels can signal underlying health problems. By analyzing blood samples from 82 adults with SM, the researchers aimed to spot protein patterns that might be linked to certain symptoms or to more aggressive forms of the disease.
The study included patients with various types of SM, ranging from the less severe indolent form to advanced forms like aggressive SM.
The authors used a panel that measured 250 inflammatory proteins and then sought links between these protein levels and specific symptoms. The results showed that specific proteins were more common in people with related complications. For example, proteins related to bone breakdown were linked to bone disease. Proteins related to fluid buildup, liver problems and other conditions were also identified.
Additionally the study used machine learning, a form of artificial intelligence that can spot complex patterns in large datasets, to determine the predictive value for the severity of specific protein patterns. The model was able to identify severe cases with over 80% accuracy, based on just a handful of key proteins.
“These proteins may represent previously unrecognized contributors to SM biology and offer new opportunities for diagnostic and prognostic refinement. All findings are being actively validated in a prospective cohort,” the authors concluded.
Sign up here to get the latest news, perspectives, and information about SM sent directly to your inbox. Registration is free and only takes a minute.
