A Precision Medicine Approach Based on Discrete Time Windows for Predicting Outcomes of Polytrauma Patients

Background: Trauma is a leading cause of morbidity and mortality in both military and civilian populations. It is unclear why some trauma victims follow a complicated clinical course and/or die, while others, with apparently similar injury characteristics, do not. Objectives: We propose to leverage Precision Medicine approaches in a three-phase study of military and civilian trauma, incorporating: (1) Phase I Base (Narrow-Window Diagnostic): A novel, time window-based trauma patient stratification scheme will be refined with genomic and admission clinical/inflammation biomarkers using both retrospective and prospective data on patients with polytrauma ± traumatic brain injury (TBI). Admission variables will be correlated with a range of outcome variables based on multiple organ dysfunction syndrome (MODS) and nosocomial infection (NI) rates. We will define the admission variables that most accurately prognosticate for these adverse outcome categories. (2) Phase 2 Option I (Wide-Window Diagnostic): The stratification algorithm from Phase 1, which is based on single time point data, will be compared against a Wide-Window algorithm involving multiple initial readings in the first 24 hours post-injury, using the dataset obtained in Phase 1. We will test the hypothesis that widening the time window for data acquisition will increase the precision of the prognostication.(3) Phase 3 Option II (Optimized Patient Stratification): A prospective study testing the optimal stratification algorithm in patients with polytrauma ± TBI. Hypotheses and Specific Aims: Phase 1 Aim 1: Optimization of a Narrow Window-based patient stratification platform formulated using patient-specific genomic, clinical, and inflammation biomarkers obtained upon hospital admission to stratify expected clinical trajectories based on meaningful in-hospital outcomes. Phase 1 Aim 2: Enriched, outcome-stratified analysis of prior interventional and observational studies in both military and civilian blunt trauma based on the Narrow-Window platform. Phase 2 Aim 1: Optimization of a Wide-Window-based patient stratification platform formulated using genomic, clinical, and inflammation biomarkers obtained at multiple time points post-admission to stratify expected clinical trajectories based on meaningful in-hospital outcomes. Phase 2 Aim 2: Enriched, outcome-stratified analysis interventional and observational studies in both military and civilian blunt trauma based on the Wide-Window platform. Phase 3 Aim 1: Optimization of a window-based patient stratification platform based on data obtained in Phases 1 and 2. Phase 3 Aim 2: A prospective, observational clinical study of polytrauma ± TBI aimed at validating a streamlined set of variables, obtained at key time points post-injury using the window-based paradigm, with respect to prognosticating adverse outcomes of patients with polytrauma ± TBI. Phase 3 Aim 3: Extend the observational period to 6 months to understand how early stratification and selection of in-hospital outcome variable correlates with long-term outcomes. These aims will be carried out by collaborative team at the University of Pittsburgh, Uniformed Services University of the Health Sciences/Surgical Critical Care Initiative (SC2i), and Indiana University, which bring expertise in all aspects of trauma care, inflammation biomarker analysis, statistical and computational modeling, and enriched, outcome-stratified analysis. Study Design: Phase 1: Refine the Narrow Window-based patient stratification algorithm using existing data from interventional and observational studies of military (SC2i) and civilian (Pittsburgh, Indiana, SC2i) trauma (>1000 patients) as well as additional genomic data. Recruit an additional 200 patients and assay both genomic (defined single-nucleotide polymorphisms [SNPs]) and circulating inflammatory mediators. Stratify patients by outcomes, including MODS characteristics, adverse in-hospital outcomes, and inflammatory networks. Utilize enriched, outcome-stratified analysis methods to determine retrospectively the potential of Narrow-Window algorithm for improved outcomes of interventional studies. Phase 2: Refine the Wide Window-based patient stratification algorithm using the methods utilized in Phase 1. Phase 3: Multiple statistical and machine learning methods will be leveraged to optimize the prognostic value of the most streamlined number of genomic, clinical, and inflammation variables for both Narrow-Window and Wide-Window platforms. Validate with data from an additional 100 patients using the two platforms in sequence in the same patients. Impact: Deliver datasets and algorithms involving genomic, clinical, and inflammation biomarkers for optimal stratification of polytrauma patients into outcome-based cohorts, thus bringing Precision Medicine approaches to combat casualty care.