Multifractal analysis of chest CT images of patients with the 2019 novel coronavirus disease (COVID-19)

Highlights

• A multifractal analysis was performed on CT images of patients with COVID-19.

• We found a way to early detection and quantitatively assessment of lung infection.

• The capacity dimension (D₀) as a new identification factors is introduced.

• We found patients with the D0 value lower than 1.4 can be healed by treatment.

• Multifractality of the CT images increased with increasing lung infection.

Abstract

The present study was done to evaluate chest computed tomography (CT) images of patients with 2019 novel coronavirus disease (COVID-19) by multifractal technique as a new method to find a way for comparing lung infection quantitatively and identifying progression pattern of the disease. The multifractal spectra extracted from analysis of CT images showed that these spectra were correlated with lung infection amount and disease progression so that, multifractal parameters (α_min, α_max, ∆α, f(α_min), f(α_max), ∆f(α), α(q = 0), and f(α) max) were strongly dependent on amount of lung infection. The results demonstrated that multifractality of chest CT images was increased with the increase in lung infection in patients. The interesting and promising result was that capacity dimension (D₀) as a new diagnostic parameter varied linearly with progression and reduction of lung infection. A critical value was found for D₀, according to which patients with D₀ lower than 1.4 can be healed by treatment. Therefore, herein, a way was found for quantitative assessment of lung infection of patients with COVID-19 by analyzing chest CT images using the multifractal method. This method can be very effective for physicians in diagnosis and treatment of pneumonia caused by COVID-19 and timely identification of therapeutic effects.