Interstitial lung diseases computer-aided imaging diagnosis, using complex networks

Abstract

Abstract The article aims to explore how a Complex Network (CN) computer-aided technique targeted for interstitial lung disease (ILD) approach can enhance the work of clinicians and if a CN-based computer-aided diagnosis can provide new data to help manage ILDs more successfully. The CN technique is used to evaluate the progression of the disease by analyzing relevant axial HRCT slices and dynamic CN evaluation using the relative speed for each layer. The article presents the results from a study of 65 patients with interstitial lung disease (ILD), comprising 18 females with a mean age of 59.35 years (ranging from 34 to 76). The initial clinical diagnosis was idiopathic pulmonary fibrosis (IPF) in 28 patients (43.07%), Non-Specific Interstitial Pneumonia (NSIP) in 11 patients, and other ILDs in the remaining patients. Each CT scan fulfilled the criteria for high-resolution CT with constant characteristics across the group. All patients underwent imagistic follow-up for at least 11 months, and additional data were provided for each investigation. The cohort was chosen based on concordant lung function decline and imaging evolution decline. The article concludes that the complex network approach provides both a qualitative visual map and quantitative metrics to enhance ILD diagnosis and progression tracking. The results suggest that a CN-based computer-aided diagnosis can provide new required data to manage ILDs more effectively. This approach may enable clinicians to make more precise conclusions regarding the structure of the analyzed lung area, which can help tailor disease management strategies to individual patient profiles.