Evidence Hub

Histopathological Differences Between Idiopathic and Secondary Interstitial Pneumonia: An Elastic Fiber-Oriented Deep Learning Analysis

Hiroyuki Katsuragawa, Hiroaki Ito, T. Handa, Ryo Sakamoto, Daisuke Nakajima, Hiroshi Date, Hironori Haga, Akihiko Yoshizawa

Human Pathology, 2026

93 participants

Abstract

Fibrotic interstitial lung diseases (ILDs) are classified into histopathological patterns by the idiopathic interstitial pneumonia (IIP) classification, and not only IIP but also secondary interstitial pneumonia (SIP) are evaluated based on the classification system, and the final integrated diagnosis is established through multidisciplinary discussion. However, distinguishing IIP from SIP is often challenging. This study aimed to investigate differences in collagen and elastic fiber composition between idiopathic and secondary ILDs using machine-learning-based quantitative analysis. We analyzed 780 Elastica van Gieson (EVG)-stained whole-slide images obtained from 93 patients with fibrotic ILDs who underwent lung transplantation using the HALO-AI. Fibrotic areas were identified, and collagen and elastic fiber areas were quantitatively measured. Among the elastic fibers, the area of elastic fiber aggregation was identified based on the increased density and dark staining intensity on EVG staining. Although no significant differences were observed between IIP and SIP in the collagen fiber area or total elastic fiber area, the elastic fiber aggregation area/fibrosis area ratio was significantly higher in SIP, particularly in the lower lobes (P = 0.0010). Similar findings were observed in the subgroup analyses of usual interstitial pneumonia pattern and nonspecific interstitial pneumonia pattern, in which secondary cases demonstrated significantly greater elastic fiber…

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free