Scientist develop software to predict tumor markers

Scientists from the University of North Carolina Lineberger Comprehensive Cancer Center have developed a software program capable of predicting which tumor-specific markers will appear on the surface of leukemia cells in patients who received stem cells. Findings were presented at the 59th Annual American Society for Hematology Annual Meeting.

By predicting which markers appear on leukemia cells, researchers would gain a step in developing immune-based therapy for leukemia patients. In the study presented, researchers validated their method in patients with blood cancers using software to predict antigenic targets.

"If you could identify and activate the immune cells that only target leukemia cells, and not normal, healthy cells, that would be a big win," saidBenjamin Vincent, MD, assistant professor in the University of North Carolina School of Medicine Division of Hematology/Oncology. "We've developed a software package that predicts leukemia-specific immune targets in any patient based on DNA and RNA sequencing. The next step of our work is to use that information for patient-specific therapies to try to improve cure rates without making graft-versus-host disease worse."

The study enrolled 101 patients with blood cancer who had an allogenic stem cell transplant. Results showed the software was able to identify all 16 known minor histocompatibility antigens while also predicting more than 100 new antigen targets. Researchers hope the software can be used to develop different treatments, targeting predicted antigens, to improve stem cells transplants and engineer donor immune cells to target cancer cells.

"We have developed a novel computational approach to predicting leukemia-specific antigens," Vincent said. "The biggest takeaway is that we can now develop personalized cell therapies that target these antigens that we hypothesize will increase leukemia cures without causing graft versus host disease."