The goal of this proposal is to develop and evaluate an integrative prediction model for diagnosis of pediatric heart transplant rejection that uses both genomic and histopathological image data. Rejection of the donor heart by the recipient is the most common cause of mortality in the pediatric heart transplant population. Unfortunately, diagnosis of rejection using current technology (i.e., histological analysis of endomyocardial biopsy samples [EMB]), is subjective, inaccurate, and imprecise. Thus, it is important to develop new diagnostic methods that leverage emerging high-throughput genomic technologies, quantitative histopathological wholeslide imaging (WSI) algorithms, and precision medicine to improve the detection of transplant rejection, guide immunosuppressive therapy, and improve patient survival. Bioinformatics and personalized medicine has shifted towards integrative methods that can effectively use the rich and complex repositories that contain multi-modal biomedical data.
Specific Aim #1: To extract quantitative image features from histopathological whole-slide images (WSls) of endomyocardial biopsies (EMBs) and develop an objective prediction model for pediatric heart transplant rejection.
Specific Aim #2: To identify differentially expressed genes [DEGs] from microRNA data extracted from EMBs and develop a gene expression-based prediction model for pediatric heart transplant rejection.
Specific Aim #3: To develop and evaluate an integrated model that combines microRNA and histopathological image data from EMBs to predict pediatric heart transplant rejection.