Abstract Bronchiolitis obliterans syndrome (BOS) is an obstructive lung disease caused by a combination of inflammation and immune response that is irreversible in its late stage. Children with BOS are typically diagnosed late because they are unable to perform spirometry, and morbidity and mortality are high. The long-term goal of this work is to improve survival and reduce morbidity from BOS by identifying strategies for accurate screening and prediction of BOS in children and young adults after HSCT and using these tools to identify novel drug targets for early intervention or prevention of BOS. The objective of this application is to validate novel predictive plasma protein biomarkers and establish a dynamic prediction model for BOS for early diagnosis, risk stratification and disease trajectory prediction for BOS after HSCT. We will achieve these goals through the following specific aims: 1) Validate longitudinal predictive performance of newly discovered plasma biomarkers of BOS risk in samples from banked and prospective studies by mass spectrometry and ELISA in both pediatric and adult cohorts. 2) Optimize and validate our dynamic prediction algorithm using pulmonary function and clinical data as well as biomarker levels (needed when no spirometry can be obtained) as covariates to project risks of BOS and rapid BOS lung-function decline to inform treatment decisions. There are currently no biomarkers or predictive tools for BOS so this work is entirely novel. The use of a dynamic prediction algorithm in this clinical setting is innovative allowing for the first time the ability to predict and diagnose early lung disease in HSCT subjects prior to the clinical diagnosis of BOS. Identification of BOS risk and stratification of screening and treatment procedures according to risk and predicted disease course would allow us to modify post-transplant care and reduce morbidity and mortality. We will use our data to inform prospective clinical trials of both early active treatment prior to development of early fibrosis and to test novel prophylactic therapies to reduce incidence in high risk individuals. Our studies will provide blood biomarkers that can be used as frequently as necessary without requiring active participation from small and often very sick children. Our preliminary biomarker and HSCT specific algorithm data demonstrate detection of BOS as soon as 2 weeks to 6 months prior to clinical diagnosis of BOS. This work is both significant and vital because improvements in HSCT techniques and supportive care have led to improved survival. Improved survival increases the number of children at risk for late complications of HSCT that are associated with life-changing morbidity and late mortality and there is urgent need to address these issues. This work will advance prediction and early diagnosis of BOS, as well as providing the framework for future prevention and treatment trials.