Accurate segmentation of interstitial lung disease (ILD) patterns from computed tomography (CT) images is an essential prerequisite to treatment and follow-up. However, it is highly time-consuming for radiologists to pixel-by-pixel segment ILD patterns from CT scans with hundreds of slices. Consequently, it is hard to obtain large amounts of well-annotated data, which poses a huge challenge for data-driven deep learning-based methods. To alleviate this problem, we propose an end-to-end semi-supervised learning framework for the segmentation of ILD patterns (ESSegILD) from CT images via self-training with selective re-training. The proposed ESSegILD model is trained using a large CT dataset with slice-wise sparse annotations, i.e., only labeling a few slices in each CT volume with ILD patterns. Specifically, we adopt a popular semi-supervised framework, i.e., Mean-Teacher, that consists of a teacher model and a student model and uses consistency regularization to encourage consistent outputs from the two models under different perturbations. Furthermore, we propose introducing the latest self-training technique with a selective re-training strategy to select reliable pseudo-labels generated by the teacher model, which are used to expand training samples to promote the student model during iterative training. By leveraging consistency regularization and self-training with selective re-training, our proposed ESSegILD can effectively utilize unlabeled data from a partially annotated dataset to progressively improve the segmentation performance. Experiments are conducted on a dataset of 67 pneumonia patients with incomplete annotations containing over 11,000 CT images with eight different lung patterns of ILDs, with the results indicating that our proposed method is superior to the state-of-the-art methods.
This study aimed to assess the feasibility and image quality of free-breathing 3D isotropic zero echo time (ZTE) whole-lung imaging and explore a clinically appropriate protocol for MR lung imaging. The study was approved by the local ethics committee. A total of thirty healthy volunteers were enrolled in this study from October 2022 to May 2023. Free-breathing pulmonary 3D isotropic ZTE scans were implemented with various acquisition planes and the number of excitations (NEX). ZTE images were evaluated by two radiologists for the overall Image quality and visibility of intrapulmonary structures as well as the signal-to-noise ratio (SNR) of the lung parenchyma. ZTE images with different acquisition parameters were compared. For preliminary clinical visual assessment, three patients with interstitial lung disease underwent both ZTE imaging and computed tomography (CT). The overall image quality of the lung in healthy subjects was good to excellent. The visibilities of pulmonary arteries and bronchus were up to the 7th and 5th generation, respectively. The display of lung fissures was poor. The overall image quality, the visibility of the pulmonary artery, and lung fissures in the axial acquisition were better than in the coronal acquisition (P = 0.011, 0.008, 0.010, respectively) but not statistically different from those in the sagittal acquisition (all P > 0.05). The free-breathing pulmonary ZTE is feasible and may serve as an alternative method in chest imaging. Either axial or sagittal ZTE image acquisition would be preferred in clinical practice.
Background: Vascular Ehlers-Danlos syndrome (vEDS) is a rare inherited connective tissue disease caused by mutations in the COL3A1 gene. The disease can cause fatal complications such as rupture of the arteries, uterus, and intestine, as well as pulmonary complications, including spontaneous pneumothorax and hemoptysis. Since hemoptysis in vEDS is rare and often misdiagnosed, this study aims to summarize its clinical features, CT findings and the diagnosis and treatment process, thereby improving clinical understanding. Methods: Patients with vEDS presenting with hemoptysis treated at the First Affiliated Hospital of Guangzhou Medical University from May 2017 to December 2024 were collected. Inclusion criteria included hemoptysis, chest CT, and COL3A1 gene test results. The clinical manifestations, CT and pathological data of the patients were collected. Results: The cohort included eight males and one female (mean age 22.22 ± 5.72 years). All patients presented with recurrent hemoptysis and cough. The CT findings included patchy ground-glass opacities, fibrous cords, nodules, and cavities, which were predominantly located in the lower lungs. 5 cases showed diffuse erythrocyte exudation and increased hemosiderin-laden macrophages. Eight cases were initially misdiagnosed as infectious diseases, of which 4 cases were treated with diagnostic anti-tuberculosis therapy. Conclusion: This study demonstrates that the clinical and imaging manifestations in vEDS patients presenting with hemoptysis lack specificity and are frequently misdiagnosed as infectious diseases. For young men with repeated hemoptysis and persistent CT abnormalities, especially when anti-infective treatment proves ineffective, vEDS should be considered and early genetic testing should be performed. Early diagnosis is crucial to prevent severe complications. Clinical trial number: Not applicable.