Prediction of severity and subtype of fibrosing disease using model informed by inflammation and extracellular matrix gene index.

Fibrosis is a chronic disease with heterogeneous clinical presentation, rate of progression, and occurrence of comorbidities. Systemic sclerosis (scleroderma, SSc) is a rare rheumatic autoimmune disease that encompasses several aspects of fibrosis, including highly variable fibrotic manifestation and rate of progression. The development of effective treatments is limited by these variabilities. The fibrotic response is characterized by both chronic inflammation and extracellular remodeling. Therefore, there is a need for improved understanding of which inflammation-related genes contribute to the ongoing turnover of extracellular matrix that accompanies disease. We have developed a multi-tiered method using Naïve Bayes modeling that is capable of predicting level of disease and clinical assessment of patients based on expression of a curated 60-gene panel that profiles inflammation and extracellular matrix production in the fibrotic disease state. Our novel modeling design, incorporating global and parametric-based methods, was highly accurate in distinguishing between severity groups, highlighting the importance of these genes in disease. We refined this gene set to a 12-gene index that can accurately identify SSc patient disease state subsets and informs knowledge of the central regulatory pathways in disease progression.