Pulmonary fibrosis progression prognosis using machine learning

Abstract

Lung fibrosis means scarring of tissue in a patient's lungs and is a common condition that can complicate the course of COVID-19 disease. Pulmonary fibrosis destroys the patient's lungs, preventing oxygenation of the blood. Modern methods of treatment are not highly effective even with access to a patient's CT scan. The problem of predicting the prognosis of pulmonary fibrosis is extremely important, since its solution will make it possible to organize clinical trials to study methods of treating patients with fibrosis more effectively. This article proposes a method for predicting the prognosis of pulmonary fibrosis progression as the volume of inhaled and exhaled air to the individual patient based on tabular patient data using an ensemble of four machine learning algorithms. This solution also provides a forecast accuracy because it is useful in medical applications to assess the 'confidence' of the model in its predictions. Modeling the proposed method shows a better result than other forecasting methods that are compared in the article. Keywords-Pulmonary fibrosis progression prognosis, Machine learning, Computer-aided diagnostics