Heritable thoracic disease (HTAD) represents a heterogeneous group of genetic conditions predisposing to thoracic aortic aneurysm and dissection, with important implications for patients and families. Accurate diagnosis requires integration of clinical assessment-including subtle syndromic features-and molecular genetic testing, supported by family screening. While aortic root pathology is a hallmark, extra-aortic manifestations such as myocardial dysfunction, arrhythmias, premature atherosclerosis and aneurysms in distal or branch vessels are increasingly recognised, with gene-specific associations informing risk stratification.Imaging plays a central role in diagnosis and longitudinal monitoring. Transthoracic echocardiography remains the first-line tool, but cross-sectional imaging (cardiovascular magnetic resonance or cardiovascular CT) is essential for complete aortic assessment and detection of extra-aortic involvement. Surveillance intervals and imaging techniques must be standardised and tailored to genotype and clinical features.Medical therapy aims to control blood pressure and reduce aortic growth. Beta-blockers and angiotensin receptor blockers are first-line in Marfan syndrome; evidence for other HTAD subtypes is emerging. Surgical thresholds differ by genotype, emphasising the importance of personalised care. Paediatric management follows similar principles but requires adapted imaging techniques, growth-adjusted interpretation and careful timing of intervention.Pregnancy in women with HTAD demands multidisciplinary planning, individualised risk assessment and close follow-up to minimise maternal and fetal complications. Finally, recent data support moderate aerobic activity while avoiding isometric and contact sports; exercise prescriptions should be individualised.Overall, HTAD care requires lifelong, multidisciplinary, gene-informed management, integrating imaging, genetics and lifestyle considerations to optimise outcomes across the lifespan.

Background: Chest X-ray (CXR) is widely used for the assessment of thoracic diseases, yet automated multi-label interpretation remains challenging due to subtle visual patterns, overlapping anatomical structures, and frequent co-occurrence of abnormalities. While recent deep learning models have shown strong performance, limitations in interpretability, anatomical awareness, and robustness continue to hinder their clinical adoption. Methods: The proposed framework employs a hybrid ConvNeXtV2–Vision Transformer (ViT) architecture that combines convolutional feature extraction for capturing fine-grained local patterns with transformer-based global reasoning to model long-range contextual dependencies. The model is trained exclusively using image-level annotations. In addition to classification, three complementary post hoc components are integrated to enhance model trust and interpretability. A segmentation-aware Gradient-weighted class activation mapping (Grad-CAM) module leverages CheXmask lung and heart segmentations to highlight anatomically relevant regions and quantify predictive evidence inside and outside the lungs. An ontology-driven neuro-symbolic reasoning layer translates Grad-CAM activations into structured, rule-based explanations aligned with clinical concepts such as “basal effusion” and “enlarged cardiac silhouette”. Furthermore, a lightweight out-of-distribution (OOD) detection module based on confidence scores, energy scores, and Mahalanobis distance scores is employed to identify inputs that deviate from the training distribution. Results: On the VinBigData test set, the model achieved a macro-AUROC of 0.9525 and a Micro AUROC of 0.9777 when trained solely with image-level annotations. External evaluation further demonstrated strong generalisation, yielding macro-AUROC scores of 0.9106 on NIH ChestXray14 and 0.8487 on CheXpert (frontal views). Both Grad-CAM visualisations and ontology-based reasoning remained coherent on unseen data, while the OOD module successfully flagged non-thoracic images. Conclusions: Overall, the proposed approach demonstrates that hybrid convolutional neural network (CNN)–vision transformer (ViT) architectures, combined with anatomy-aware explainability and symbolic reasoning, can support automated chest X-ray diagnosis in a manner that is accurate, transparent, and safety-aware.

Background/Objectives: The clinical adoption of AI in radiology requires models that balance high accuracy with interpretable, anatomically plausible reasoning. This study presents an integrated diagnostic framework that addresses this need by unifying a hybrid deep-learning architecture with explicit anatomical guidance and neuro-symbolic inference. Methods: The proposed system employs a dual-path model: an enhanced EfficientNetV2 backbone extracts hierarchical local features, whereas a refined Vision Transformer captures global contextual dependencies across the thoracic cavity. These representations are fused and critically disciplined through auxiliary segmentation supervision using CheXmask. This anchors the learned features to lung and cardiac anatomy, reducing reliance on spurious artifacts. This anatomical basis is fundamental to the interpretability pipeline. It confines Gradient-weighted Class Activation Mapping (Grad-CAM) visual explanations to clinically valid regions. Then, a novel neuro-symbolic reasoning layer is introduced. Using a fuzzy logic engine and radiological ontology, this module translates anatomically aligned neural activations into structured, human-readable diagnostic statements that explicitly articulate the model’s clinical rationale. Results: Evaluated on the NIH ChestX-ray14 dataset, the framework achieved a macro-AUROC of 0.9056 and a macro-accuracy of 93.9% across 14 pathologies, with outstanding performance on emphysema (0.9694), hernia (0.9711), and cardiomegaly (0.9589). The model’s generalizability was confirmed through external validation on the CheXpert dataset, yielding a macro-AUROC of 0.85. Conclusions: This study demonstrates a cohesive path toward clinically transparent and trustworthy AI by seamlessly integrating data-driven learning with anatomical knowledge and symbolic reasoning.

Adapting international evidence-based guidelines to local challenges: A Lebanese perspective on the latest American Thoracic Society and Infectious Diseases Society of America (ATS/IDSA) community-acquired pneumonia antibacterial therapy recommendations.

Chest X-ray is the most popular examination type for thoracic diseases, but its interpretation exhibits error rates which are still subject to inter-observer variability and economic workload constraints. This work can be found in this paper: "A reproducible deep learning pipeline for pneumonia identification vs normal based on DenseNet-121 from chest X-ray". The dataset originated from the NIH ChestX-ray14 corpus and was downsampled to 8,500 frontal radiographs (1,050 pneumonia-positive, 7,450 normal) and split at the patient level into training, validation and testing sets. Preprocessing: Grayscale normalization, Resize, Targeted Augmentation and Training (with) Early Stopping, Learning Rate Scheduling, Class Weighting and Post-Hoc Probability Calibration. In the held-out test set, the model achieved ROC-AUC: 0.87, PR-AUC: 0.72, as well as a general accuracy of 93.2%, sensitivity: 82.8% and specificity: 94.6%. Calibration analysis contributed to improving the Brier score from 0.042 to 0.019 and led to good-fitting reliability curves. Interpretability was built into the inference using Grad-CAM and Integrated Gradients, with explanation faithfulness quantitatively checked (deletion AUC = 0.84, insertion AUC = 0.87, sanity check pass rate = 98%, pointing-game hit rate = 76%). Based on the above results, it can be seen that CNN-based diagnosis is promising to achieve a good accuracy as well as interpretability and reproducibility simultaneously. Hence, the proposed framework provides a white-box baseline for clinical examination and future multi-label thoracic disease detection extensions

Accurate classification of chest X-ray (CXR) images is vital for early detection of thoracic diseases such as COVID-19, Tuberculosis, and Pneumonia, particularly in regions with limited radiological expertise. While deep learning has shown promise in CXR interpretation, many existing models rely solely on internal datasets, risking overfitting and poor generalizability. Furthermore, inadequate tuning of network architectures may limit robustness across varied imaging conditions. This study presents an externally validated deep learning framework based on Convolutional Neural Networks (CNNs) for multi-disease CXR classification. This study compared a baseline CNN with two convolutional layers against a tuned architecture with three layers across multiple image resolutions (64×64, 112×112, 224×224). The proposed model employs transfer learning with a pre-trained CNN, fine-tuned for four-class classification using a softmax output layer. Training was performed with the Adam optimizer (learning rate: 0.0001, batch size: 32) and categorical cross-entropy loss, for up to 50 epochs with early stopping. Internal validation showed the tuned model outperformed the baseline, achieving 0.97 accuracy and an F1-score of 0.89. External validation confirmed superior generalizability, with the tuned model attaining an F1-score of 0.83 and an AUC of 0.97 at 112×112 resolution, compared to the baseline’s F1-score of 0.79 and AUC of 0.94. These results highlight the potential of optimized CNN architectures as reliable, scalable tools for radiological decision support in resource-limited healthcare systems. Future work will incorporate explainable AI methods and real-world clinical validation to ensure safe, interpretable deployment.

BackgroundThe 2019 American Thoracic Society and Infectious Diseases Society of America community acquired pneumonia guidelines recommend empiric coverage of methicillin-resistant Staphylococcus aureus and Pseudomonas aeruginosa based on previous respiratory isolation, recent IV antibiotic use, and locally validated risk factors. This study aims to describe how local risk factors may be determined efficiently using data retrieved electronically.MethodsThis retrospective cohort study focused on the time period May 13, 2020, through June 30, 2024. Consecutive adults admitted to one of five acute care facilities with confirmed community-acquired pneumonia were included. Community-acquired pneumonia was defined as the presence of one or more pneumonia diagnosis codes and an order for a respiratory culture or an antimicrobial with the indication of pneumonia or sepsis, 24 h before or within 48 h after the date and time of admission. Patients were excluded if they had a diagnosis code for hospital-acquired or ventilator-associated pneumonia, any subsequent admission in the study period, or if they had a previous respiratory culture positive for methicillin-resistant Staphylococcus aureus or Pseudomonas aeruginosa within a year of admission. The causative pathogen and the presence or absence of evaluated risk factors were electronically abstracted from billing data and health records. Serial quality assessments of electronic data were performed to improve accuracy until a well validated population was determined.ResultsThere were 4,558 unique patients included. Methicillin-resistant Staphylococcus aureus and Pseudomonas aeruginosa rates were 0.6% and 0.7%, respectively. Only age was inversely associated with risk of methicillin-resistant Staphylococcus aureus (OR = 0.86, 95% CI: 0.76–0.98). No significant risk factors for Pseudomonas aeruginosa were found.ConclusionsIn rural or otherwise resource limited healthcare settings, risk factors for methicillin-resistant Staphylococcus aureus and Pseudomonas aeruginosa community-acquired pneumonia may be determined using only electronic data capture and the methodology described in this article.Supplementary InformationThe online version contains supplementary material available at 10.1186/s41479-025-00188-6.

Transthoracic needle biopsy (TNB) is one of the routine procedures for thoracic diseases, especially nodules or consolidation. The procedure can be guided by imaging tools, such as computed tomography (CT) scan and ultrasonography (US). This study reports the results of a five-year experience of transthoracic imaging-guided needle biopsy in a respiratory referral hospital. We searched for a monthly sampling database in the procedure room from 2020 to 2024 and identified all transthoracic imaging-guided needle biopsies, either by CT or US-guided. We excluded a few data samples if there was a repetition of the data register. Data regarding pathology and procedure-related complications were analyzed, with the primary outcomes being disease proportion and positivity rate of the procedure. A total of 1,591 procedures were included in our final analysis. Almost all procedures (99.6%) used a 16-gauge needle core biopsy size. Computed tomography was used predominantly (89.9%) to guide the procedure rather than ultrasound. Adenocarcinoma was the most frequent pathology result of TNB (37.7%). The complications were rare (1.6%) and there was zero mortality reported within 24 hours after TNB procedures. Lung cancer was the most reported case, followed by lymphoma and tuberculosis (TB). The overall accuracy of the TNB procedure in lung and mediastinal consolidation was 96.3%. Transthoracic needle biopsy has high accuracy and is considered a safe procedure with minor complications.

Diagnosing thoracic diseases from chest radiographs remains challenging, especially in resource-limited environments. This study presents YOLOv8n-cls, a lightweight deep learning model for real-time classification of five pathologies: 1) COVID-19, 2) fibrosis, 3) normal, 4) pneumonia, and 5) tuberculosis. The model was trained on a dataset of 11,019 chest X-ray images, combining public data from NIH ChestX-ray14 and a private clinical dataset, and achieving a Top-1 accuracy of 92.23%. Preprocessing included format conversion and text removal, while data augmentation techniques such as flipping, rotation, brightness/contrast adjustment, and affine translation were applied to improve model generalization. Performance evaluation relied on confusion matrices, precision, recall, F1-score, specificity, and ROC-AUC curves. Moreover, Grad-CAM visualizations were employed to enhance interpretability and analyze misclassification patterns. YOLOv8n-cls provides a strong balance between accuracy and computational efficiency, making it suitable for real-time clinical deployment.

This article provides apractice-oriented overview of how thoracic diseases can be reliably assessed using small specimens. It demonstrates how cytologic and histologic methods complement each other, why careful preanalytical handling and specimen triage/tissue stewardship are critical, and how ancillary studies can be used judiciously to conserve tissue. The article also addresses current developments such as digital analysis workflows and offers practical recommendations-including common pitfalls and guidance on interpreting results.

European Society for Vascular Surgery (ESVS) 2026 Clinical Practice Guidelines on the Management of Descending Thoracic and Thoraco-Abdominal Aortic Diseases.

Cardiac Rehabilitation After Thoracic Aortic Surgery.

Treatment of thoracic aortic diseases remains a major challenge for cardiovascular surgeons, especially in the ascending aorta and aortic arch. This study aims to evaluate long-term outcomes, exploring risk factors of mortality and complications in thoracic aortic aneurysm treated by hybrid procedure. This is a retrospective and prospective descriptive study from the hospital database at Viet Duc hospital University, which includes patients diagnosed with thoracic aortic aneurysm and treated with hybrid procedure from January 2014 to December 2023. 50 patients, composed of 44 males (88.0%) and 6 females (12%) were included in the analysis The mean age of the patients was 64.6 ± 10.96 years old (30 - 81). 36 patients had left thoracic pain (62.0%), 18 cases (36.0%) underwent aorto-bicarotid bypass with median sternotomy, 27 patients (54.0%) had carotid–carotid bypass with or without left subclavian artery revascularization, 5 patients (10.0%) had left subclavian artery transposition into left carotid artery. The survival estimation was 78.0% at 1 year, 73.8 at 3 years, 69.7% at 5 years and 53.4% at 7 years. Zone 0 intervention had increased risk of procedure-related mortality. At present, hybrid procedure can provide successful treatment suitable younger and older patients in Vietnamese condition.

BackgroundThere is a lack of data regarding the use of hybrid stent graft prostheses in patients with hereditary thoracic aortic disease (HTAD) involving the aortic arch and proximal descending aorta. This retrospective analysis aimed to evaluate the short- and mid-term outcomes of hybrid stent-graft prostheses in Frozen Elephant Trunk (FET) procedures for patients with HTAD, with a particular focus on its safety and feasibility.MethodsA total 280 patients who underwent FET procedures between October 2010 and November 2024 were retrospectively analysed in compliance with the 2024 EACTS/ STS recommendations for shared decision-making within the multidisciplinary aortic team. Among them, 51 patients had genetically confirmed HTAD (Marfan syndrome (FBN1), Loeys-Dietz syndrome (TGFBR1, TGFBR2, SMAD3, TGFB2), vascular Ehlers-Danlos syndrome (COLSA1), and non-syndromic HTAD (ACTA2, MYH11, MYLK)). The Thoraflex™ prosthesis was implanted in 50 of the 51 patients. Short- and mid-term outcomes were assessed descriptively. Survival and subsequent thoracic aortic intervention rates were analysed using the Kaplan-Meier method.ResultsThe overall 30-day mortality was 2.0% (n = 1). Perioperatively, permanent neurological deficit was 3.9% (n = 2), with minor disability on the modified Rankin Scale (mRS 1 and 2). There were no instances of paraplegia. The median follow-up was 4.0 years. The 1-, 3- and 5-year overall survival rate was 93.9%, 90.6%, and 90.6%, respectively. Freedom from subsequent aortic interventions was at 1, 3, 5 years 55.8%, 45.6%, and 33.1%. Early device-related complications occurred in 7 patients (13.7%), including intraluminal FET thrombosis in 4 patients (12.5%) and distal stent graft-induced new entry in 3 patients (9.4%). Mid-term device-related complications occurred 2 patients (4.3%).ConclusionsHybrid stent graft prostheses can be safely implanted with the FET technique in elective and acute HTAD patients with arch and proximal descending aortic disease. Our single-center short- and mid-term outcomes are encouraging, but long-term durability and efficacy are not yet established. This warrants multi-center studies with extended follow-up.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13019-025-03738-7.

BackgroundThoracic endovascular aortic repair (TEVAR) has become the reference treatment for thoracic diseases in high-risk patients. Regarding the aging of the population, the indications for TEVAR in octogenarians are constantly increasing with little data on the results of this treatment in this population. The aim of this study was to evaluate the short- and medium-term results of TEVAR for descending thoracic aorta diseases in octogenarians’ patients.MethodsIn a single-center retrospective study, patients aged 80 years and older who underwent TEVAR for descending thoracic aortic aneurysm or dissection between 2000 and 2021 were included. Demographic data, perioperative morbi-mortality, mortality risk factors, and anatomical results were analyzed using the latest available scanner during follow-up.ResultsIn this study from 2000–2021, 35 octogenarians’ patients (57% men, 43% women) underwent treatment for thoracic aortic pathologies, 74% were treated for aneurysm, 26% for aortic dissection and 34% required urgent surgery (10 ruptures, 2 malperfusions). Technical success was 100%, with 14.0% intra-hospital mortality (4 respiratory distress, 1 cardiac arrest). Morbidity rate was 42% (7 respiratory complications, 2 paraplegias, 2 strokes, 3 renal failures). After 14 months of follow-up, death rate was 20% (6/30). Aortic diameter analysis showed stable lesions in 69.2% (n=18), regression in 26.9% (n=7), and increased diameter in 3.8% (n=1). Type I endoleaks were observed in 8.6% (n=3), type 2 in 11.4% (n=4).ConclusionsEndovascular treatment of descending thoracic aortic pathologies in patients aged 80 and older carries higher risks. Careful patient selection is crucial, and therapeutic abstention may be a viable option for this population.

Aim: This study aimed to compare perioperative and long-term outcomes of zone 0 versus zone 1 hybrid aortic arch repair in a Vietnamese cohort. Material and Methods: The present study is a retrospective–prospective cohort study of 117 patients who underwent hybrid arch repair at a single tertiary centre from 2014 to 2023. Hybrid arch repair was performed in patients with thoracic aortic diseases, including aneurysm, dissection, intramural hematoma, penetrating aortic ulcer, and blunt thoracic aortic injury. Patients were stratified by proximal landing zone 0 or 1. Perioperative outcomes, complications, and long-term survival were analyzed using Kaplan–Meier and Cox regression methods. Results: Thirty-four patients underwent zone 0 repair and 83 underwent zone 1 repair. The median follow-up duration was 40.7 ± 29.6 months (range 0.2–122.6 months), with a 97.4% follow-up completion rate. The overall 30-day mortality rate was 9.0%, significantly higher in the zone 0 group (26.5%) compared to zone 1 (6.0%; p = 0.002). Stroke occurred in 6.0% of patients. The overall survival rates at 1, 3, 5, and 10 years were 96.2%, 89.6%, 86.9%, and 75.0%, respectively. At 10 years, survival in zone 0 was 66.8% versus 82.5% in zone 1 (p = 0.019). Retrograde type A dissection was observed in 2.6% of patients, one in zone 0. Late complications, including endoleak type Ia (1.7%) and graft occlusion (0.9%), were infrequent and not statistically different between groups. Conclusion: Zone 1 hybrid arch repair was associated with better early and long-term outcomes than zone 0. These findings support preferential use of zone 1 landing when anatomically feasible and underscore the importance of proximal landing zone selection in optimizing hybrid TEVAR outcomes in high-risk populations.

ObjectivesTo characterise the aetiological spectrum of thoracic granulomatous diseases and identify diagnostic features that facilitate differentiation among causes.DesignRetrospective multicentre observational study.SettingTwo tertiary hospitals in south-central China. Patient data were consecutively enrolled from 1 June 2020 to 30 June 2023.ParticipantsOf 2486 patients with pathologically confirmed thoracic granulomas initially identified, exclusions were applied for specimens outside the lung/pleura/mediastinum (579), incomplete demographic/imaging/pathology data (280) or lack of follow-up (231). A total of 1396 patients met all criteria (853 from hospital 1 and 543 from hospital 2) and were included in the final analysis.Primary and secondary outcome measuresWe quantified the aetiologic distribution of thoracic granulomatous diseases and examined age-stratified/lesion-location differences in aetiologic patterns. We also evaluated associations between histopathological features and specific aetiologies and compared the diagnostic accuracy across sampling modalities.ResultsAmong the 1396 enrolled cases, a confident, probable and uncertain diagnosis was made in 1086 cases, 307 cases and 84 cases. Infectious granulomas predominated (89.4%; 1248/1396), with tuberculosis comprising 87.8% (1096/1248) of infectious cases. Among non-infectious granulomas, sarcoidosis was most common (65.8%; 50/76). Patients aged ≥60 years had a higher proportion of infectious granulomas than younger groups (≥60 years: 389/422; 40–60 years: 633/714; <40 years: 226/260; p=0.044). Fungal granulomas were more frequent in those aged 40–59 years (10.9% of infectious granulomas). Diagnostic accuracy was highest for surgical biopsy (93.5%, 288/305), followed by thoracoscopy (92.3%, 108/117) and transbronchial lung biopsy (74.7%, 510/683). Necrosis was present in 88.6% of granulomas; positive special stains were strongly associated with fungal infection.ConclusionsMost thoracic granulomas arise from mycobacterial or fungal infection, while sarcoidosis is the leading non-infectious cause. Thoracoscopy and surgical biopsy show superior diagnostic yields, and special staining aids differentiation of fungal aetiologies. Findings support a multidisciplinary approach to improve diagnostic accuracy.

Mendelian Randomization Suggests a Causal Link Between Glycemic Traits and Thoracic Aortic Structures and Diseases

The approval of programmed death-ligand 1 (PD-L1) inhibitors in the first line of treatment has transformed the therapeutic landscape of extensive-stage small cell lung cancer (ES-SCLC); real-world (rw) evidence of clinical benefit is currently limited. In this study, we investigated the rw efficacy of first-line chemoimmunotherapy and the role of potential biomarkers. We retrospectively assessed patients with SCLC receiving first-line chemoimmunotherapy at Sotiria Thoracic Diseases Hospital of Athens, Athens, Greece. Kaplan-Meier curves were used to calculate real-world progression-free survival (rwPFS) and real-world overall survival (rwOS). Cox proportional hazards regression analysis was utilized to identify associations between patient characteristics and outcomes. 188 patients were included. Median rwPFS was 6.5months (95% CI 5.8-7.1months) and median rwOS was 11.2months (95% CI 9.1-12.0months). rwOS was higher in the atezolizumab group compared with the durvalumab group (median, 12.0 vs 9.2months; hazard ratio [HR], 1.51; 95% CI 1.06-2.15; p = 0.02), similar results were observed for rwPFS (median, 6.5 vs. 6.0months, HR, 1.55; 95% CI 1.10-2.16; p = 0.01). In multivariate analysis, the difference between atezolizumab and durvalumab was not statistically significant, while lung, bone and liver metastases, ECOG PS, LDH and NLR were associated with an increased risk of death. Associations were utilized for the generation of a novel prognostic score with good discriminatory power (C-statistic: 0.73). Real-world efficacy of first-line chemoimmunotherapy in patients with ES-SCLC is comparable to randomized trials. The association between prognostic scores and survival outcomes in ES-SCLC should be explored in prospective studies.Query.

Chest X-ray (CXR) represents one of the most widely utilized clinical diagnostic tools for thoracic diseases. Nevertheless, computer-aided diagnosis based on chest radiographs still faces considerable challenges in anomaly detection. Certain lesions in CXRs exhibit subtle radiographic characteristics with ambiguous boundaries, low pixel occupancy, and weak contrast. While existing studies primarily focus on improving multi-scale feature fusion, they frequently overlook complications arising from background noise and varied lesion morphology. This study introduces YOLOv11-MFF, an enhanced YOLOv11 network with three key innovations. Specifically, a novel Frequency-Adaptive Hybrid Gate (FAHG) is developed to improve contrast differentiation between lesions and background. A Multi Scale Parallel Large Convolution (MSPLC) block is designed and integrated with the original C3k2 module to expand receptive fields and enhance long-range modeling capacity. Furthermore, a Feature Fusion module (FF) is introduced to reinforce target-relevant feature representation through channel-wise modulation via weight recalibration mechanisms. Benefiting from these advancements, the network achieves significant improvements in detecting multi-scale and overlapping lesions. Experimental results on the public VinDr-CXR dataset demonstrate that YOLOv11-MFF outperforms state-of-the-art models, achieving a precision of 48.2%, recall of 42.5%, mAP@0.5 of 41.5%, and mAP@0.5:0.95 of 22.6%.