Fractal Analysis of Chest Radiographs Using Image-J-FIJI Software- A Pilot Study

Background: Lung vasculature has nutritive and functional roles. Unlike the bronchial tree which branches dichotomously into twenty-one generations, the pulmonary arteries give supernumerary branches to perfuse the neighboring parenchyma. The pulmonary arteries additionally branch for five more generations than airways before forming capillaries. Further pulmonary veins are interlobular in position. Hence the characterization quantifying the pulmonary vascular networks is challenging. Objective: In this study, we assessed the pulmonary vasculature in chest radiographs using the fractal analysis method on Image-J-Fiji software. Design: Cross-sectional study Settings: Patients referred to the Department of Radiodiagnosis, Akash Institute of Medical Sciences and Research Centre and Dr. B. R. Ambedkar Medical College, Bangalore, Karnataka, India Participants: One hundred and thirty-two chest radiographs of normal healthy individuals (aged 2 months to 80 years) were photographed. Each of these images was processed in the Image-J-Fiji software. A box-counting algorithm quantified the images. Data results of the fractal dimensions were validated at the probability of significance [0.05]. Results: The mean fractal dimension of the pulmonary vasculature was 1.39. For males and females, the Pearson's correlation coefficient between the fractal dimension and age in years was 0.102 and 0.16, respectively. In males, a chi-square value of 0.58, in females, a chi-square value of 1.03, degree of freedom 2 and critical value of p-value 0.05, showed the relation was statistically not significant. Comparison between fractal dimension and Gender using Cramer’s V test in males, 0.066, and in females, 0.088, indicates a weak association between fractal dimension and gender. Conclusion: The applicability of the fractal dimensions is to screen the high risks of severe chronic obstructive pulmonary diseases. The determination of fractal values helps evaluate the complexity of lung tumors. The baseline data concerning fractal properties of pulmonary vasculature obtained from this study helps to evaluate lung diseases like emphysema and vascular abnormalities during the progression of chronic obstructive pulmonary disease. KEYWORDS: Chest Radiograph, Image Analysis, Fractal dimension, Pulmonary vasculature.