Dynamic forecasting module for chronic graft-versus-host disease progression based on a disease-specific subpopulation of B cells

Abstract

Due to its dynamic nature and the absence of reliable real-time monitoring tools, predicting chronic graft-versus-host disease (cGVHD) progression was challenging. This caused a significant investment of both time and financial resources to ensure good management of cGVHD. In response to this challenge, we identified a distinct B-cell subpopulation characterized by CD27+CD86+CD20-, which could precisely distinguish cGVHD from healthy donors. Leveraging this discovery, we developed cGPS, a user-friendly tool based on marker distribution, which demonstrated exceptional efficacy in tracking cGVHD progression. Its validation, conducted through retrospective and prospective studies involving 91 patients (25 non-GVHD and 66 cGVHD cases), confirmed cGPS's predictive prowess. Remarkably, our retrospective analysis revealed an impressive area under the curve (AUC) of 0.9773 for identifying non-GVHD patients at risk of cGVHD and 0.8846 for predicting disease progression in cGVHD patients. Subsequent validation in an independent prospective study yielded equally promising results, with cGPS accurately predicting all instances of cGVHD development or progression within a three-month observation window. With three independent cohorts, cGPS underscores its robust ability for sensitive and dynamic monitoring of cGVHD progression, provides a solution for early diagnosis and assessment of treatment effectiveness for cGVHD.