The identification of molecules that reflect and/or predict disease progression and can be easily measured in chronic lung disease is highly desirable. Such biomarkers that can be rapidly and reliably measured to establish change in the course of the disease over time or in response to therapy would significantly improve care. For example, measurement of short term markers that track with or predict long term changes in FEV1 (a clinical measure of lung function) may shorten the length, significantly reduce the cost of clinical trials, and accelerate the examination of potential therapies allowing faster delivery to the patients. Furthermore, the behavior of molecular biomarkers can inform therapy and increase success of alleviating disease progression. In preliminary studies, we analyzed differences in pathways in serum samples from 44 children with mild lung disease (FEV1>80th percentile) and 44 with more severe lung disease (FEV1<45th percentile). We found significant differences in a number of proteins and pathways associated with disease. The top markers in our analysis correlated with disease progression and significantly improved novel disease progression prediction models. Here we propose to integrate the longitudinal behavior of novel disease markers with novel Functional Data (FD) analysis of FEV1 and other clinical data to develop an algorithm that models lung function decline. To achieve this we propose to; 1) identify and validate serum proteome changes in banked samples collected from patients with stable and declining FEV1, 2) develop a dynamic prediction model that integrates validated proteomic biomarkers with Functional Data (FD) analysis of longitudinal FEV1 values to produce a novel diagnostic algorithm that identifies individuals at risk of lung function decline, and 3) test the capacity of dynamic prediction modeling to identify modulator treated CF patients who demonstrate rapid pulmonary decline compared to highly responsive individuals in banked longitudinal samples. If we are successful, we will clinically translate a novel lung function prediction model that will significantly improve therapeutic intervention and accelerate clinical studies.
Although CF disease progression is well characterized, forecasting disease progression is a pressing need and has not been successful. We have discovered novel molecular markers and created an algorithm that predicts lung function, and propose here to develop it for clinical translation in CF patients.
