Improving Survival Rate by Estimating the Progression of Pulmonary Fibrosis

Abstract

Pulmonary Fibrosis (PF) is a chronic and progressive lung disease, it tightens the lungs and make a person unable to breath. A person suffering with PF can experience at different rates based on their age, health conditions, and lifestyle and so on. Basically this PF occurs without any cause or else when they are exposed to environmental hazards and autoimmune diseases. The outcomes can range from long-term stability to rapid deterioration. There is no cure for PF. The life expectancy of patients with PF is 3–5 years in average after diagnosis. We used CNN model and multivariate regression analysis for the prediction and progression of the PF. Early detection of the disease is the key for slowing progression and happens only when the patient is known of their severity. PF patients will lose 150–200 mL of lung capacity in average, which can be monitored by spirometry. Disease stage of the patient is determined by their lung capacity and the severity of their symptoms. Current procedures make fibrotic related lung diseases became problematic to treat by considering chest CT scan yet, does not cure. By using data science, CT scan of their lungs, machine learning techniques, image, metadata and baseline FVC as input the project predicts the stage of severity and progress of the patient.

Keywords