Rationale Interstitial lung diseases (ILD) are a diverse group of conditions, often diagnosed using high-resolution chest computed tomography (HRCT), which is susceptible to subjective biases in interpretation. Objectives This study aims to develop and validate SPAIDNet (Spatial Pattern Analysis for ILD Diagnosis using a residual neural Network), a deep learning (DL) model for the automated classification of ILD, to reduce subjective biases and improve diagnostic consistency. Methods The study included 2901 ILD patients who underwent 5213 HRCT scans across multiple centers between July 2017 and June 2023. SPAIDNet, built upon the pre-trained residual neural network with 18 layers, utilizes multi-instance learning in three centers in China. Measurements and main results The model demonstrated exceptional performance, achieving macro-average area under the receiver operating characteristic curve (AUC) of over 0.999 in internal validation, 0.905 in external cohort I, and 0.870 in external cohort II for multiclass classification. SPAIDNet outperformed both a junior radiologist (AUC: 0.737) and a senior radiologist (AUC: 0.763). Furthermore, DL-assisted the two radiologists saw significant improvements in diagnostic accuracy, with AUCs rising to 0.817 and 0.787, respectively. Conclusions These results underscore SPAIDNet's potential to offer high accuracy, robustness, and generalizability in ILD diagnosis, providing a valuable tool to mitigate the subjectivity inherent in HRCT image interpretation.

Background Non-metastatic lymphadenopathy is challenging to diagnose. The comparative diagnostic performance of endobronchial ultrasound (EBUS)-guided transbronchial mediastinal cryobiopsy (TBMC) vs. EBUS-transbronchial needle aspiration (TBNA) remains debated. Methods This multicenter randomized trial was conducted in three hospitals. Patients with at least one mediastinal and/or hilar lesion of ≥1 cm in the short axis who required diagnostic bronchoscopy were included. The patients were randomized in a 1:1 ratio to receive either EBUS-TBNA followed by EBUS-TBMC (EBUS-TBNA-first group) or EBUS-TBMC followed by EBUS-TBNA (EBUS-TBMC-first group). The primary outcome was the diagnostic yields of EBUS-TBMC and EBUS-TBNA. Findings The overall diagnostic yield of EBUS-TBMC for non-metastatic lymphadenopathy was significantly higher than that of EBUS-TBNA for specific benign etiologies and lymphomas (97.1% vs. 79.9%, p < 0.001). In the subgroup analysis, EBUS-TBMC showed a higher sensitivity for sarcoidosis than EBUS-TBNA (98.0% vs. 82.7%, p < 0.001). All patients experienced grade 1 airway bleeding. Conclusions EBUS-TBMC demonstrated a higher diagnostic yield than EBUS-TBNA for non-metastatic lymphadenopathy in a cohort almost exclusively composed of sarcoidosis cases, with a good safety profile. EBUS-TBMC is a potential first-line diagnostic tool for non-metastatic lymphadenopathy. Funding This work was financially supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0528900 and 2024ZD0528902 to G.H.).

Currently, no targeted therapy exists for idiopathic pulmonary fibrosis (IPF). The hallmark pathological feature of excessive extracellular matrix (ECM) deposition severely undermines the efficacy of mesenchymal stem cell (MSC)-based treatments. While existing MSC therapeutic strategies primarily focus on modulating inflammation in early stages, they have not yet established precise interventions addressing the core pathological mechanism-ECM dysregulation. Previous studies demonstrated the therapeutic potential of human embryonic stem cell (hESCs)-derived immunity-and-matrix-regulatory cells (IMRCs) in lung injury and fibrosis models. However, the critical biomarkers and underlying mechanisms mediating IMRCs' efficacy in IPF remain poorly understood. In this study, we generated MMP1 knockout IMRCs (IMRCs-MMP1 KO) using CRISPR-based gene editing. We then characterized whether MMP1 ablation affected key properties of IMRCs, including cell morphology, proliferation, migration, marker protein expression, transcriptomic profile, and cytokine secretion. Subsequently, the ability of IMRCs-MMP1 KO to degrade collagen was tested using in vivo and in vitro pulmonary fibrosis models. MMP1 knockout was successfully achieved and did not compromise typical IMRC characteristics or impair their immunomodulatory capacity. However, MMP1 deficiency significantly attenuated the ability of IMRCs to degrade TGF-β1-induced collagen I deposition in A549 cells. Importantly, wild-type IMRCs demonstrated superior therapeutic efficacy in ameliorating bleomycin-induced lung injury and fibrosis in mice compared with IMRCs-MMP1 KO. Furthermore, IMRCs exhibited significantly greater capability to directly degrade the pericellular collagen I and modulate fibroblasts' activation progression within fibrotic lung tissues in a MMP1-dependent manner. In summary, our data establish that MMP1 plays an essential functional role in IMRC-mediated attenuation of PF. MMP1 thus represents a key therapeutic biomarker for IMRC-based treatment. This work provides a foundation for developing stem cell therapies tailored to the pathological features of IPF, potentially enabling adaptive treatment strategies.

Sarcoidosis is a systemic granulomatous disease that has limited treatment options. Emerging evidence suggests that macrophages are essential for sarcoid granuloma initiation. Legumain (LGMN), a cysteine protease, regulates macrophage polarization in various cancers. However, its involvement in sarcoid granuloma formation remains elusive. Herein, LGMN is upregulated in macrophages within sarcoid-like granulomas. Genetic deletion of Lgmn exacerbates granulomatous inflammation in a Propionibacterium acnes (P. acnes)-induced mouse model, accompanied by increased M1 macrophage polarization. Mechanistically, LGMN binds to integrin αvβ3 on the macrophage surface and restrains M1 polarization by inhibiting the mechanistic target of rapamycin complex 1 (mTORC1)/signal transducer and activator of transcription 1 (STAT1) pathway. Furthermore, intratracheal administration of lipid nanoparticles carrying Lgmn plasmid DNA effectively alleviates granuloma formation induced by P. acnes or trehalose 6,6'-dimycolate, concomitant with decreased mTORC1/STAT1 activation and M1 polarization. These findings reveal the pivotal role of LGMN in restraining sarcoid granulomatous inflammation through suppression of mTORC1/STAT1-driven M1 macrophage polarization. Therefore, LGMN supplementation may be a promising therapeutic strategy for sarcoidosis.

The prognostic value of deep-tissue microstructure profiling derived from transbronchial lung cryobiopsy (TBLC) specimens in idiopathic pulmonary fibrosis (IPF) remains poorly explored. This study aims to bridge this gap by quantitatively extracting collagen features in heterogeneous pathological regions using multiphoton microscopy (MPM), and providing novel insights into the potential of collagen features as sensitive biomarkers for predicting severity and disease progression in IPF. In this prospective and observational study, a total of 82 patients with IPF undergoing TBLC and 22 healthy controls were enrolled. The differences in collagen features between the normal/mild and moderate/severe groups were compared. Multivariate logistic regression was used to identify collagen features associated with disease progression. Histopathologically, 62.2

Background:Fibrotic interstitial lung disease (fILD) consists of a heterogeneous group of chronic, progressive interstitial lung diseases characterized by reduced lung elasticity and restrictive ventilatory impairment. Strain analysis based on dynamic ventilation computed tomography (DVCT) has emerged as a novel method for quantifying lung deformation during ventilation and thus monitoring the pathophysiological changes in lungs. This study aimed to quantitatively identify abnormal lung motion in patients with fILD through use of strain analysis and to determine the correlation of these values with spirometric indices. Methods:A total of 27 patients with fILD and 20 healthy controls were prospectively recruited. All participants underwent DVCT scanning on a 320-row computed tomography (CT) scanner. Strain metrics across the full respiratory cycle were computed with computational fluid dynamics software at nine axial levels spanning the upper, middle, and lower lungs. In patients, the percentage of fibrotic lung at the corresponding levels was also quantified. Group differences were assessed, and Pearson correlations were used to determine the associations between strain metrics, percentage fibrosis, and pulmonary function parameters. Results:During the respiratory cycle, lung strain exhibited heterogeneous temporal and spatial distributions. There were significant differences in maximum principal strain, mean principal strain, maximum displacement speed, and mean displacement speed between the upper, middle, and lower parts of both lungs, both in patients and controls (P<0.05). In both groups, peaks in strain were observed during the early expiration and mid-inspiration phases. Patients with fILD exhibited a distinct pattern, with consistently lower strain values across all metrics, while those of healthy controls were all significantly higher. The strain-related parameters were significantly correlated with forced expiratory volume at 1 second (r: 0.646-0.769; P≤0.001), forced vital capacity (r: 0.670-0.827; P≤0.001), and total lung capacity (r: 0.625-0.817; P≤0.001), whereas the percentage of lung fibrosis was not associated with any other parameters. Conclusions:DVCT-derived strain represents a quantitative measure of abnormal regional lung motion in patients with fILD and may complement spirometry by capturing local ventilatory mechanics. This technique shows promise for the evaluation of regional mechanical impairment in patients with fibrotic lung disease.

Sarcoidosis is a systemic granulomatous disease that often affects the lungs, lymph nodes, and skin. The pathological hallmark of sarcoidosis is the formation of epithelioid, non-necrotizing granulomas, with CD4+ T-helper (Th) type 1 (Th1), Th17, and Th17.1 cells scattering throughout the granuloma. The CXCL16-CXCR6 axis plays a crucial role in the development of inflammatory diseases. However, the precise role of CXCR6+ CD4+ T cells in sarcoidosis pathogenesis and their therapeutic potential remain unclear. In this study, we confirmed the enrichment of CXCR6+ CD4+ T cells in the lesional tissue of sarcoidosis patients and Propionibacterium acnes-induced sarcoidosis-like model mice. Flow cytometric analysis and single-cell RNA sequencing revealed that CXCR6+ CD4+ T cells exhibited a Th17/Th17.1 phenotype and displayed pro-inflammatory characteristics. By performing cell-cell communication analysis and validating the findings with flow cytometry and CXCL16-blockade experiments, we demonstrated an active crosstalk between CXCR6+ CD4+ T cells and CXCL16+ macrophages in mice. Furthermore, treating mice with an anti-CXCR6 monoclonal antibody effectively reduced the mRNA levels of pro-inflammatory genes, including Nos2, Cxcl9, Cxcl10, Il17a, Tnf, and Ifng. In addition, anti-CXCR6 treatment reduced the abundance of Th17 and Th17.1 cells, which was associated with inhibition of downstream mTORC1 signaling. Finally, anti-CXCR6 treatment suppressed granuloma formation and attenuated collagen deposition in the lungs. Taken together, our findings highlight CXCR6 as a promising therapeutic target for inhibiting granuloma formation and pulmonary fibrosis in sarcoidosis.

Idiopathic pulmonary fibrosis is a progressive interstitial lung disease with limited treatment options and poor prognosis. Increasing evidence suggests that airway epithelial remodeling contributes to disease pathogenesis, yet the role of basal cells remains incompletely understood. This study aimed to determine whether basal cells participate in pulmonary fibrosis through a senescence-associated secretory phenotype regulated by the transcription factor SP1. METHODS:Lung tissue from patients with idiopathic pulmonary fibrosis who underwent transbronchial cryobiopsy was compared with control samples from patients undergoing lung resection for benign nodules. Histopathological analysis and immunohistochemistry were used to evaluate basal cell distribution and quantify cell-specific markers. Transcriptomic data from the GEO database were analyzed to identify differentially expressed genes and enriched signaling pathways. Findings were validated in a bleomycin-induced pulmonary fibrosis mouse model, and SP1 expression was assessed by protein analysis. RESULTS:Basal cells were markedly increased in fibrotic airways and extended from bronchioles into fibroblast foci, with their abundance correlating positively with fibrosis severity. Differential gene expression analysis identified enrichment of senescence-associated secretory phenotype-related pathways, including upregulation of metalloproteinases and chemokines. In the mouse model, senescence-associated mediators were significantly elevated, and SP1 expression was increased in fibrotic lungs. CONCLUSION:Basal cells may actively contribute to pulmonary fibrosis, and SP1-mediated senescence-associated secretory phenotype may represent a potential novel pathogenic mechanism. Targeting basal cell dysfunction or SP1-related pathways may offer new therapeutic opportunities for idiopathic pulmonary fibrosis.

Coal-dust, a persistent airborne pollutant, induces dose-related pulmonary fibrosis; however, plasma biomarkers for pre-clinical toxicity remain lacking. We enrolled 158 participants, including 28 healthy controls (HCs), 30 dust-exposed workers (DEWs), and 100 patients with coal workers’ pneumoconiosis (CWP) at different stages (n CWP−I=40, n CWP−II=30, n CWP−III=30). Plasma proteomic profiling was performed via data-independent acquisition (DIA) mass spectrometry. Differentially expressed proteins were identified and functionally annotated. Key proteins were selected and multiple machine learning algorithms were employed to construct and validate predictive models. We identified 1,239 plasma proteins, including 645 high-confidence candidates. Functional enrichment revealed significant associations between disease progression and pathways such as PPAR signaling, cholesterol metabolism, Epstein-Barr virus infection, and the pentose phosphate pathway. These alterations converge on dysregulated lipid metabolism, chronic inflammatory signaling and virus-induced immune evasion, suggesting a metabolic-immune axis that orchestrates early fibrotic progression. We successfully constructed the first plasma proteomics-based machine learning models for pneumoconiosis grading and early screening. Notably, a single biomarker, PRSS3, demonstrated exceptional performance in distinguishing DEW patients from early-stage pneumoconiosis patients (CWP-I), achieving an area under the curve (AUC) of 1.00 and an accuracy of 1.00 in the training set and an AUC of 1.00 with an accuracy between 0.93 and 1.00 in the validation set. This study establishes innovative machine learning-based models for the grading and early screening of pneumoconiosis via plasma proteomics. The identification of PRSS3 as a potential biomarker highlights the clinical utility of our approach. These findings provide a foundation for noninvasive diagnostic strategies and future translational research in occupational lung diseases.

Background: Individuals with chronic obstructive pulmonary disease (COPD) have a significantly increased risk of major adverse cardiovascular events (MACE). Active case-finding of COPD in population with high cardiovascular disease (CVD) risk may have substantial clinical implications. This study aimed to develop and validate an interpretable machine learning (ML) model to guide targeted spirometry use in this specific population. Methods: The national COPD screening data and CVD screening (China Health Evaluation And risk Reduction through nationwide Teamwork, ChinaHEART) data were matched to form the derivation cohort (2319 subjects for training and 993 for internal validation). The China Pulmonary Health (CPH) and ChinaHEART data were matched as external validation cohort 1 (n=373), while data from the National Health and Nutrition Examination Survey (NHANES) 2007-2012 were included as external validation cohort 2 (n=1013). Five feature selection methods and ten ML algorithms were employed, and the final models were interpreted using the SHapley Additive exPlanations (SHAP) framework. Findings: The derivation cohort included 3312 individuals with high CVD risk, among whom 591 (17.8%) had COPD. Ten variables were selected, and the gradient boosting machine (GBM) performed best, achieving area under the receiver operating characteristic curves (AUCs) of 0.821 (95% CI: 0.784-0.857) in internal validation, 0.816 (95% CI: 0.760-0.872) in external validation cohort 1, and 0.781 (95% CI: 0.748-0.814) in external validation cohort 2. Additionally, an 8-variable GBM model (excluding laboratory-derived variables) demonstrated acceptable performance, with AUCs of 0.803 (95% CI: 0.766-0.840), 0.787 (95% CI: 0.726-0.847), and 0.768 (95% CI: 0.734-0.802) across the respective cohorts. Online calculators based on these models were developed and deployed for clinical application. Interpretation: Leveraging nationwide screening data, we developed and validated an interpretable ML model for identifying COPD in population with high CVD risk. Its online calculators may assist clinicians in early identification of individuals requiring spirometry.

Chronic inhalation of coal dust causes coal workers' pneumoconiosis (CWP), which is one of the leading occupational diseases. Coal workers' pneumoconiosis is currently incurable, posing a serious public health threat. Identifying the underlying molecular mechanisms of CWP is critical to overcome this challenge. Nowadays, metabolomics has bridged underlying molecular alterations with disease progression, providing a useful tool for researching the pathogenesis and finding biomarkers. In this study, a comprehensive view of metabolic characterization of serum from CWP patients at all stages was provided using untargeted metabolomic analysis. As a result, when compared to healthy controls, the specific alteration patterns of each stage were observed. The results showed arginine and cortisol could be core metabolites in CWP progression. Moreover, five metabolites that significantly changed when going from "solely chronic coal-dust exposure" to an early stage were screened out as potential biomarkers. The receiver operating characteristic results were 0.691-0.862 (individual) and 0.884-0.907 (combined). These findings will benefit the application of metabolomics to understand the pathological mechanism and identify diagnostic biomarkers for CWP.

Hypersensitivity pneumonitis (HP) manifests as fibrotic (FHP) and non-fibrotic (NFHP) phenotypes. Clinically distinguishing FHP from idiopathic pulmonary fibrosis (IPF) remains challenging owing to phenotypic overlap, despite divergent management protocols. This investigation sought to develop a plasma proteomics-based framework for differential diagnosis between these entities. A total of 119 subjects were enrolled from the Chinese Interstitial Lung Disease (ILD) National Cohort and the PORTRAY IPF Cohort between July 2018 and June 2022, comprising 32 healthy controls (HCs), 31 NFHPs, 28 FHPs, and 28 IPF patients. The plasma samples were subject to quantitative proteomic profiling, weighted gene co-expression network analysis (WGCNA), and bioinformatics analysis to identify differentially expressed proteins, core pathways, and co-expression modules. Key proteins were selected to construct and validate diagnostic models via seven machine learning algorithms. This study delineated the plasma proteomic landscape of FHP and IPF, identifying 813 proteins. WGCNA revealed significant enrichment of the glycolysis/gluconeogenesis and pyruvate metabolism pathways, implicating metabolic reprogramming in FHP pathogenesis. Differential analysis identified nine differentially expressed proteins, from which a six-protein signature (H2BC12, SHBG, APCS, PTPRG, IGHV1-58, and GAPDH) was derived through LASSO regression and recursive feature elimination. Among seven machine learning algorithms, support vector machine (SVM) achieved the optimal performance on the independent test set with an accuracy of 71.4

BACKGROUND:Cryptogenic organizing pneumonia (COP) is classified as a subtype of idiopathic interstitial pneumonia (IIP), which has a good prognosis. Relapse remains one of the most challenging and intriguing aspects of COP. This study aimed to characterize the clinical features, prognosis, and relapse patterns of COP. METHODS:In this prospective cohort study, patients diagnosed with COP between March 1, 2004 and June 30, 2022 were enrolled after a comprehensive multidisciplinary review. Patients were followed up through regular clinical visits every 3-6 months until October 30, 2023. Data collected included demographic characteristics, clinical presentation, laboratory findings, radiologic features, and prognostic outcomes. The primary outcome measure was relapse. Cox regression analyses were performed to explore factors related to the presence of relapse. RESULTS:This study included 268 patients with a final diagnosis of COP based on multidisciplinary discussion. After diagnosis, the one-year survival rate was 99.6%, decreasing to 97.7% at 3 years, 94.5% at 5 years, and 81.7% at 10 years. Relapses occurred in 20.9% (56/268) of patients, with a median time to relapse of 12 months. Of these relapses, 28.6% (16/56) occurred when prednisone was tapered to below 10 mg/day, and 62.5% (35/56) occurred after discontinuation of prednisone therapy; fewer than 10% of relapses occurred at prednisone doses above 15 mg/day. Ground-glass opacities (GGOs) (225/268, 84.0%) and consolidations (212/268, 79.1%), predominantly with bilateral lung distribution, were the most common computed tomography (CT) features at diagnosis. No patient developed residual lung fibrosis on high-resolution CT, regardless of relapse status. Furthermore, no deaths were attributable to COP progression or relapse, and all-cause mortality was unrelated to COP itself. CONCLUSION:This study indicated that patients with COP had a good prognosis with favorable survival outcomes. However, relapse was relatively common, occurring in more than 20% of patients. Therefore, close follow-up is essential for early detection of relapses, particularly during corticosteroid tapering or after treatment discontinuation.

Sarcoidosis is a multisystem disease pathologically characterized by non-caseating epithelioid granulomas. Its neurological form, neurosarcoidosis (NS), can affect the central nervous system (CNS) and/or the peripheral nervous system (PNS), posing a significant diagnostic challenge. The diagnostic dilemma is further complicated by the significant clinicopathological overlap between sarcoidosis and tuberculosis (TB). Although rare, sarcoidosis and TB can coexist. Our case series included three NS patients: two had CNS involvement and one had PNS involvement. Notably, two patients had confirmed active tuberculosis infection and one patient had a latent TB infection. This highlights the diagnostic and therapeutic challenges of these coexisting conditions. Fluorodeoxyglucose positron emission computed tomography (FDG PET) helps detect extraneural sarcoidosis locations and find biopsy sites. Regarding the similarity and potential coexistence of TB and sarcoidosis, thorough screening for TB infection is essential before initiating any therapy for sarcoidosis.

Primary Sjögren’s syndrome-associated interstitial lung disease (pSS-ILD) shows heterogeneous patterns and variable prognosis. We previously observed elevated serum CA125 in pSS-ILD. This study investigated whether CA125 predicts adverse outcomes in pSS-ILD. Data were derived from the ILD-China cohort (ClinicalTrials.gov: NCT04370158). Among enrolled pSS-ILD patients, a total of 395 with follow-up data were included. The adverse outcomes were death or lung transplantation (LTx). Associations were evaluated using Cox proportional hazards models, with restricted cubic splines to assess nonlinearity. Among 395 patients, 87 experienced adverse outcomes (83 deaths and 4 LTx). CA125 levels were higher in patients with adverse outcomes and were associated with indices of disease severity. In ROC analysis, CA125 showed the highest discrimination among the evaluated biomarkers (AUC 0.764), with a higher AUC than KL-6 (0.580). In multivariable Cox models, higher CA125 modeled as a continuous biomarker was independently associated with adverse outcomes (adjusted HR 1.005 per 1 U/mL increase, 95

Sarcoidosis is a multisystem granulomatous disease that primarily affects the lungs. Although macrophages and T cells are implicated in the pathogenesis of sarcoidosis, their heterogeneity from granuloma formation to pulmonary fibrosis remains unclear. Using single-cell RNA sequencing (scRNA-seq), we characterized macrophages and T cells at different disease stages in lung tissues from Propionibacterium acnes (PA)-induced sarcoidosis-like model mice. The findings revealed that Lgals3hi macrophages showed notable enrichment during both granuloma formation and fibrosis. Temporal transcriptional profiling of Lgals3hi macrophages revealed a functional shift from pro-inflammatory signaling during the granulomatous stage to MHC-II antigen processing during fibrosis. Investigation of CD4+ T cell heterogeneity showed that CD4⁺ Th1 cells exhibited pro-inflammatory NF-κB profiles during granuloma formation, while adopting a migratory phenotype during fibrosis. Moreover, reduced Treg cell proliferative capacity was observed during fibrosis. Cellular communication analysis further revealed an upregulated MHC-II-mediated interaction between macrophages and CD4⁺ T cells both prior to PA re-exposure and during fibrosis. Together, our findings provide deeper insight into the cellular mechanisms underlying sarcoidosis progression, facilitating the identification of new therapeutic targets for the disease.

BackgroundAnti-synthetase syndrome (ASS) associated interstitial lung disease (ILD) usually responds to immunosuppressive therapy, but recurrence is common. We report a 58-year-old man with ASS-ILD who developed recurrent ILD within one year after bilateral lung transplantation (LTx). Despite triple immunosuppression (glucocorticoids, tacrolimus, mycophenolate mofetil), systemic inflammation persisted. This case represents an in vivo model of ASS-ILD recurrence, warranting further investigation into underlying mechanisms and novel therapeutic strategies.MethodsPeripheral blood mononuclear cells (PBMCs) were collected at 56 and 84 weeks post-transplant for single-cell RNA sequencing (scRNA-seq), while lung tissue was analyzed via spatial transcriptomics. Control data came from five naïve ASS-ILD patients and two clinically stable connective tissue disease associated ILD (CTD-ILD) patients post-LTx. Differential gene expression and pathway enrichment analyses were performed to identify therapeutic targets.ResultsExploratory PBMC scRNA-seq analysis suggested enrichment of interferon-, interleukin- and JAK-STAT-related signaling programs in circulating monocytes and neutrophils. Based on this immune activation profile, a multidrug treatment adjustment was implemented, including short-term glucocorticoid augmentation, replacement of mycophenolate mofetil with Janus kinase inhibitor tofacitinib, and replacement of tacrolimus with cyclosporine A. Following treatment adjustment, systemic inflammatory markers declined and interstitial lesions in both lungs were markedly alleviated on imaging. Subsequent spatial transcriptomic analysis of lung tissue revealed persistent interferon-related signaling and identified pro-fibrotic transcriptional programs in alveolar macrophages and transitional type II alveolar cells. Functional enrichment suggested a potential association between systemic inflammatory activation and localized fibrotic remodeling within the lung microenvironment.ConclusionsThis case illustrates the potential utility of scRNA-seq and spatial transcriptomics for characterizing immune-related transcriptional programs in recurrent ASS-ILD after LTx. Transcriptomic profiling informed therapeutic decision-making in this complex clinical setting, and clinical improvement was observed following treatment adjustment. These findings generate exploratory insights into potential therapeutic targets in recurrent ASS-ILD.

Background:Interstitial lung abnormalities (ILAs) are incidental abnormal lung findings on computed tomography that can co-exist with lung cancer, but risk factors for their progression and impact on survival after lung cancer resection remain unclear. This study aimed to identify risk factors for ILA progression, develop a radiomics-based model to predict it, and evaluate the association between ILA progression and mortality. Methods:Patients with ILAs who underwent lung cancer resection in Shanghai Chest Hospital between January 2016 and May 2019 were retrospectively selected and divided into three subcategories at baseline: non-subpleural ILAs, subpleural non-fibrotic ILAs, and subpleural fibrotic ILAs; and three subcategories during follow-up: improved ILAs, unchanged ILAs, and progressive ILAs. Multivariate logistic regression was used to identify clinical and laboratory risk factors associated with ILA progression, and radiomics-based machine learning models were independently constructed to predict ILA progression using baseline CT imaging features. Survival data were analyzed using Kaplan-Meier curves and Cox proportional hazards regression. Results:There were 1363 ILA cases among 10,295 patients who underwent primary lung cancer resection, with a proportion of 13.24%. After 2- and 4-year follow-up, the progression rates of ILAs were 9.86% (65/659) and 10.19% (43/422), respectively. Subpleural fibrotic ILAs (2-year follow-up: OR = 6.078, 95% confidence interval [CI]: 2.633-14.028, P < 0.001; 4-year follow-up: OR = 3.339, 95% CI: 1.085-10.272, P = 0.035), and radiotherapy (OR = 13.595, 95% CI: 4.540-40.710, P < 0.001; OR = 11.496, 95% CI: 2.864-46.141, P = 0.001) were risk factors for ILA progression. Using radiomics features extracted from baseline CT, the AutoGluon machine learning model demonstrated high performance in predicting ILA progression among patients with confirmed ILAs (for all ILAs, mean area under curve value: 0.790, accuracy: 0.801 ± 0.049). The all-cause mortality rates at 2- and 4-year follow-up were 2.11% (27/1281) and 4.26% (53/1245), respectively. Both subpleural fibrotic (log-rank χ² = 23.910, P < 0.001) and progressive ILA groups (log-rank χ² = 26.098, P < 0.001) had higher mortality rates compared with non-subpleural, subpleural non-fibrotic ILA group and improved, unchanged ILA group; and progressive ILAs were a risk factor for death (HR = 2.824, 95% CI: 1.065-7.483, P = 0.037). Conclusions:ILA progression occurs in about 10% of patients after lung cancer resection and is associated with higher mortality. Subpleural fibrotic ILAs and radiotherapy are key risk factors for progression. Radiomics features effectively predict ILA progression, enabling early identification of high-risk patients.

IntroductionMetabolomics analysis shows great promise in identifying non-invasive biomarkers for interstitial lung diseases (ILDs). However, the relevant data are scattered across numerous disparate publications, hindering their full utilization.ObjectivesTo comprehensively leverage the metabolomic data disseminated throughout the literature, we manually curated and integrated them into the ILDMDB database (https://ildmdb.shinyapps.io/ILDMDB/). This database will be regularly updated and maintained.MethodsWe conducted a systematic literature search and extracted key metabolomics data, including changes in metabolites, relevant clinical parameters, and predictive model performance metrics etc. These data were then manually integrated into the ILDMDB database.ResultsThe current version of ILDMDB contains 3,969 entries, representing 20 ILD types and over 1,000 metabolites derived from Homo sapiens, animal models, and cell line experiments. Each entry comprises detailed information, including the metabolite name, disease type, and original reference. In addition, we have incorporated model data on metabolites used for ILD diagnosis, disease severity, and prognosis, along with information on metabolites associated with clinical parameters. Users can search for target metabolites freely, view their expression patterns and detailed information, and manage metabolite collections in the database.ConclusionILDMDB serves as an exploratory platform designed to assist researchers in swiftly and conveniently accessing the metabolic landscape of ILDs, thereby advancing research into the diagnosis, prognosis, and treatment of ILDs from a metabolic perspective.