12.17.21

Machine Learning Predicts GVHD and Post-Transplant Cyclophosphamide Signatures

Research published in the journal Blood presents the findings of a study in which authors used machine learning to investigate signatures of Graft Versus Host Disease (GVHD) to explore post bone marrow transplant outcomes. The results of their investigation highlight the utility of post-transplant cyclophosphamide (PTCy) in enhancing alloimmune response in patients receiving HLA-haploidentical (haplo) and HLA-matched bone marrow transplants.

Lymphoma patients receiving bone marrow transplants are at risk of developing GVHD as a result of donor T cells attacking the recipient's own tissue. Prophylactics are commonly used as preventive therapy to reduce inflammation and infection associated with GVHD, with specific signatures playing a critical role in nonrelapse mortality (NRM), relapse, and overall survival (OS) rates in patients receiving transplants. Currently, there are no known immunologic signatures associated with the clinical outcomes of PTCy. Therefore, the researchers used machine learning with immunophenotypic, proteomic, transcriptomics, and clinical data to find possible signatures.

Analysis from the machine learning, which investigated 35 potential candidate signatures, showcased that a predictor of acute GVHD is increased levels of CD4+ conventional T cells (Tconv) combined with high CXCL9 28 days post-transplantation. Additionally, higher numbers of natural killer (NK) cells were linked to enhanced OS because of NK cells' ability to reduce NRM and relapse. Taken together, these findings help guide potential new therapies for GVHD.

Reference:

McCurdy SR, Radojcic V, Tsai HL, et al. Signatures of GVHD and Relapse after Post-Transplant Cyclophosphamide Revealed by Immune Profiling and Machine Learning [published online ahead of print, 2021 Oct 17]. Blood. 2021;blood.2021013054. https://doi.org/10.1182/blood.2021013054

Theme picker

Search