Predicting Incident Stroke Using Blood Biomarkers of Brain Injury
This project will derive a dynamic prognostic model using a RNN (recurrent neural network)-based framework and data collected during the first two weeks post-injury from BOOST-2 (Brain Oxygen Optimization in Severe Traumatic Brain Injury
Inaccurate stroke risk prediction constitutes a critical barrier to effective stroke prevention. Prevention of stroke depends on accurate identification of at-risk individuals. However, according to the American Heart Association, an ideal stroke risk prediction tool does not yet exist. Consequently, inaccurate stroke risk prediction results in unnecessary exposure of low risk individuals to the risks of stroke prevention pharmacotherapy and under-treatment of certain high risk groups. Therefore there is a critical need to identify novel stroke risk predictors that can improve the accuracy of current stroke risk prediction tools. Our long-term goal is to decrease the incidence of stroke by developing more accurate stroke risk prediction tools that allow the right stroke prevention strategies to be targeted to the right at-risk individuals. Our overall objective in this application is to identify blood biomarkers of brain injury that are independent predictors of stroke. Our central hypothesis is that serum levels of brain derived neurotrophic factor (BDNF) and neurofilament light chain (NF- L), two biomarkers of brain injury, are both independent predictors of incident stroke. Our hypothesis has been formulated on the basis of previous work identifying silent brain infarction as an important predictor of future stroke. Additionally, conditions associated with brain cellular death (such as brain infarction) are often accompanied by the release of brain-specific proteins such as BDNF and NF-L into the extracellular matrix and subsequently into blood. Among participants in the Framingham Heart Study, low levels of BDNF were associated with increased risk of future stroke/transient ischemic attack. BDNF is a neurotrophic factor that mediates homeostatic interactions between neurons, glial cells and the cerebral endothelium. It is also stored in platelets. NF-L is a structural axonal protein that is exclusively found in the nervous system. Thus, serum NF-L levels may be even more reflective of brain injury than serum BDNF levels. We propose confirming the role serum BDNF in stroke risk prediction and determining the predictive value of NF-L for stroke. We will also determine whether 12-month changes in NF-L and BDNF are more predictive of stroke risk than baseline levels, as well as determine whether NF-L and BDNF provide predictive information for stroke risk beyond traditional CVD risk factors. The proposed study is aligned with the NHLBI mission of preventing cardiovascular and blood diseases. Our expected outcomes are: a) generation of new knowledge regarding the role of 2 biomarkers of brain injury in predicting stroke, while maximizing the scientific value of NHLBI?s biospecimen collections; and b) provision of critical evidence needed for further studies that will refine current stroke risk prediction tools by incorporating blood biomarkers brain injury. Our results will have a positive impact because they will lead to improved stroke risk prediction and ultimately to more successful stroke prevention.