Journal of Thoracic Imaging最新文献

{"title":"Comparative Performance of 3 Artificial Intelligence Systems for Lung Nodule Characterization in Low-Dose Computed Tomography Screening.","authors":"Khulan Khurelsukh, Yen-Po Lin, Hsuan-Ming Chang, Wen-Chi Hsu, Pei-Ching Huang, Chen-Te Wu, Yung-Liang Wan","doi":"10.1097/RTI.0000000000000877","DOIUrl":"https://doi.org/10.1097/RTI.0000000000000877","url":null,"abstract":"<p><strong>Purpose: </strong>This study evaluates 3 artificial intelligence (AI) systems in detecting, characterizing, and classifying lung nodules on low-dose computed tomography (LDCT) scans of 100 subjects, assessing agreement with a reference standard and inter-vendor consistency.</p><p><strong>Materials and methods: </strong>Performance of 3 commercially available AI platforms-AI 1, AI 2, and AI 3-was retrospectively analyzed against evaluations by 2 thoracic radiologists, with discordances resolved by consensus as reference standard. Agreements were assessed for nodule presence, type (solid, part-solid, ground-glass), and Lung-RADS category using Cohen Kappa. Agreement for continuous measurements (nodule diameter and volume) across AI systems was evaluated using intraclass correlation coefficients (ICC). Group comparisons for continuous variables were performed using the Kruskal-Wallis test, with Mann-Whitney U tests for post hoc pairwise comparisons. Categorical variables were compared using χ2 tests. Bland-Altman analysis evaluated variability in diameter and volume measurements.</p><p><strong>Results: </strong>The 3 AI systems detected 435, 152, and 70 nodules, respectively, whereas radiologists identified 126 nodules (P<0.001). Sensitivity, specificity, and accuracy were 77.0%, 8.2%, and 25.7% for AI 1; 72.2%, 83.4%, and 80.6% for AI 2; and 42.9%, 95.7%, and 82.2% for AI 3. Agreement with the reference standard was perfect for AI 2 and almost perfect for AI 3, but absent for AI 1. Inter-AI agreement was substantial (κ=0.66 to 0.78), and diameter/volume measurements showed moderate to good reliability (ICC=0.57 to 0.87).</p><p><strong>Conclusion: </strong>Commercial AI systems show variable performance in nodule detection and classification, underscoring the need for users to understand each system's characteristics and interpret results within clinical context.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147436951","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Variability in Mediastinal Lymph Node Measurements in Chest Contrast-enhanced CT: Time to Change the Paradigm?","authors":"Alon Olesinksi, Richard Lederman, Yusef Azraq, Leo Joskowicz, Jacob Sosna","doi":"10.1097/RTI.0000000000000859","DOIUrl":"10.1097/RTI.0000000000000859","url":null,"abstract":"<p><strong>Purpose: </strong>Measurement of mediastinal lymph nodes (LNs) is an integral part of patient assessment, and is performed by manually measuring the short axis length (SAL) of the LNs on axial slices. LNs with SAL ≥10 mm are considered pathologically enlarged. We aimed to quantify the interobserver agreement and variability of SAL measurements, compare them to automatically computed SALs from manual LN delineations, and establish the mean SAL measurement error.</p><p><strong>Materials and methods: </strong>Two radiologists independently measured the SALs of 451 LNs in 40 contrast-enhanced chest CT (CECT) scans. One of them also manually delineated the LN contours in each CECT slice, and this served to automatically classify LN as normal/enlarged based on their SALs. Differences between SAL measurements and Bland-Altman statistics were computed.</p><p><strong>Results: </strong>The normal/enlarged LN overall agreement (371 normal, 52 enlarged) between both radiologists was 93.8% (423/451). For agreement/disagreement, the SAL differences were 1.1 (1.0) mm (17%) and 3.5 (3.2) mm (40%). The disagreement differences were nearly twice as large as the agreement differences. The agreement between the manual and the computed SALs for both radiologists was 92.7% (418/451), similar to the interobserver variability.</p><p><strong>Conclusion: </strong>Classification of mediastinal lymph nodes based on SAL measurements demonstrates high agreement. It indicates that SAL measurements automatically computed from manual LN delineations could be a reliable and time-saving tool. In cases of disagreement, the ±2 mm error supports the use of 3 size categories: normal (<8 mm), possibly enlarged (8 to 12 mm), and definitely enlarged (>12 mm).</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145309767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Chest Computed Tomography-Based Radiomics and Machine Learning for Classifying Mediastinal Lymphadenopathy Caused By Hematologic Malignancies and Metastatic Abdominopelvic Solid Cancers.","authors":"Haoru Wang, Qian Hu, Yingxue Tong, Huiru Zhu, Ling He, Jinhua Cai","doi":"10.1097/RTI.0000000000000860","DOIUrl":"10.1097/RTI.0000000000000860","url":null,"abstract":"<p><strong>Purpose: </strong>To evaluate the role of chest CT radiomics in classifying mediastinal lymphadenopathy caused by hematologic malignancies and abdominopelvic solid cancers.</p><p><strong>Materials and methods: </strong>A total of 231 patients with mediastinal lymphadenopathy were selected from the Mediastinal-Lymph-Node-SEG collection in The Cancer Imaging Archive, including 145 patients with hematologic malignancies (74 with chronic lymphocytic leukemia and 71 with lymphoma) and 86 with abdominopelvic solid cancers. Patients were randomly stratified into train and test sets in a 7:3 ratio. Radiomics features were extracted from enhanced CT images of mediastinal lymph nodes, followed by feature selection using univariate analysis and least absolute shrinkage and selection operator regression. A support vector machine algorithm was used to develop classification models, with performance evaluated using the area under the receiver operating characteristic curve (AUC-ROC), accuracy, and 95% CI.</p><p><strong>Results: </strong>For differentiating mediastinal lymphadenopathy between hematologic malignancies and abdominopelvic solid cancers, the model incorporated 23 features and achieved an AUC-ROC of 0.931 (95% CI: 0.891-0.971) and an accuracy of 0.866 in the train set, and an AUC-ROC of 0.830 (95% CI: 0.730-0.929) and an accuracy of 0.759 in the test set. For distinguishing chronic lymphocytic leukemia from lymphoma, the model utilized 4 features, achieving an AUC-ROC of 0.880 (95% CI: 0.813-0.947) and an accuracy of 0.752 in the train set, and an AUC-ROC of 0.872 (95% CI: 0.763-0.982) and an accuracy of 0.836 in the test set.</p><p><strong>Conclusions: </strong>Chest CT radiomics shows promise for classifying mediastinal lymphadenopathy in patients with hematologic malignancies and abdominopelvic solid cancers.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145240495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging Artificial Intelligence to Transform Thoracic Radiology for Lung Nodules and Lung Cancer: Applications, Challenges, and Future Directions.","authors":"Geewon Lee, Hwan-Ho Cho, Dong Young Jeong, Jong Hoon Kim, You Jin Oh, Sung Goo Park, Ho Yun Lee","doi":"10.1097/RTI.0000000000000866","DOIUrl":"10.1097/RTI.0000000000000866","url":null,"abstract":"<p><p>This review traces the historical path of artificial intelligence (AI) methods that have been applied to medical image interpretation. Early AI approaches, which were based on clinical expertise and domain-specific medical knowledge, established the basis for data-driven methods, initiating the radiomics era and leading to the widespread use of deep learning in medical imaging. More recently, transformer architectures-originally developed for natural language processing-have been adapted for medical image analysis. In the first section, we explore the literature on the use of AI, specifically addressing lung nodules and lung cancer. AI has been effective in detecting lung nodules, evaluating their characteristics, and predicting cancer risk, while also addressing technical issues like kernel conversion. In lung cancer, AI has been applied to various clinical needs, including prognosis evaluation, mutation identification, treatment response analysis, operability prediction, treatment-related pneumonitis, and clinical information extraction. In the following section, we explore foundation models, multimodal AI, and a multiomic approach in the field of lung nodules and lung cancer. Finally, as AI models continue to evolve, so too must the approaches for evaluating their real-world utility; thus, we outline relevant methods for evaluating the performance and application of AI in thoracic radiology.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145543660","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial Intelligence in Cardiovascular MRI: From Imaging to Biomechanics and Diagnosis.","authors":"Amin Mahmoodi, Akhilesh Yeluru, Jerjes Aguirre-Chavez, Kathryn Lamar-Bruno, Karan Punjabi, Shant Malkasian, Albert Song, Evan Masutani, Albert Hsiao","doi":"10.1097/RTI.0000000000000864","DOIUrl":"10.1097/RTI.0000000000000864","url":null,"abstract":"<p><p>In this review, we highlight how artificial intelligence, specifically deep learning, is reshaping every aspect of cardiovascular magnetic resonance imaging: from planning and acquisition to reconstruction, analysis, and clinical report generation. We first introduce core machine learning paradigms and concepts, then survey recent deep learning advances to automate and enhance multiple aspects of MRI. We highlight the range of recent advances to provide a conceptual understanding of how the field has rapidly evolved in the last 10 years, enabling improvements in acquisition speed, spatial resolution, suppression of artifacts, and correction for motion. Automation of postprocessing is providing us a deeper look into detailed analysis of regional cardiac function and measurement of hemodynamics, and a greater ability to automatically integrate interpretation with nonimaging clinical data to support prognostication and management. Advances in artificial intelligence will continue to shape our practice of clinical cardiovascular MRI to provide greater efficiency and enrich our ability to guide the management of patients with cardiovascular disease.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12810196/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145649848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantitative CT and Artificial Intelligence in Chronic Lung Disease.","authors":"Andrea S Oh, Stephen M Humphries, Augustine Chung, S Samuel Weigt, Matthew Brown, Grace Hyun J Kim, David Lee, John A Belperio, Jonathan G Goldin","doi":"10.1097/RTI.0000000000000867","DOIUrl":"10.1097/RTI.0000000000000867","url":null,"abstract":"<p><p>Computed tomography (CT) is routinely used in diagnosing and managing patients with chronic lung diseases such as chronic obstructive pulmonary disease (COPD) and fibrosing interstitial lung disease (ILD). Visual assessment of disease morphology/phenotype and extent correlates with lung function and patient prognosis, but it is limited by reader subjectivity and interobserver variability. Quantitative CT (QCT) techniques based on density and texture-based features of the lungs have shown stronger correlations with physiologic and survival outcomes in both COPD and ILD cohort studies. Moreover, recent advances in computer processing capabilities have led to the implementation of machine and deep learning-based approaches, allowing for greater robustness and reproducibility beyond visual assessment and density-based methods. This review focuses on QCT and artificial intelligence (AI) techniques for COPD, ILD, and bronchiolitis obliterans syndrome in lung and hematopoietic stem cell transplant recipients. Current challenges and limitations for adoption of these techniques and future directions of QCT and AI in thoracic imaging are also discussed.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145795266","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial Intelligence in Coronary Computed Tomography: Current Applications, Future Potentials, and Real-world Challenges.","authors":"Lorenzo Giarletta, Brian Zhou, Riccardo Marano, Carlo N De Cecco, Marly van Assen","doi":"10.1097/RTI.0000000000000873","DOIUrl":"10.1097/RTI.0000000000000873","url":null,"abstract":"<p><p>Artificial intelligence (AI) is rapidly transforming cardiac computed tomography (CT) imaging by enhancing image acquisition, reconstruction, and analysis to improve diagnostic accuracy and overall clinical workflow. Deep learning reconstruction (DLR) algorithms optimize image quality while reducing radiation and contrast media doses. AI-driven tools for coronary artery segmentation and CAD-RADS classification ensure greater reproducibility and efficiency in coronary artery disease (CAD) assessment. Beyond anatomic evaluation, AI enhances functional imaging with CT-derived fractional flow reserve and myocardial CT perfusion imaging, improving the noninvasive identification of myocardial ischemia associated with flow-limiting coronary lesions. AI also plays a key role in CAD phenotyping through automating quantification and characterization of total plaque burden and identifying rupture-prone plaques and high-risk patients. Radiomics and machine learning models analyzing pericoronary adipose tissue (PCAT) propose new biomarkers of coronary inflammation, refining risk stratification and disease monitoring. Fusion models integrating clinical, imaging, and laboratory data are emerging as powerful tools for comprehensive cardiovascular risk prognostication, surpassing traditional clinical risk scores. Looking ahead, generative AI and large language models (LLMs) could revolutionize radiology workflows by automating report generation and relevant clinical data extraction and integration, while digital twins may enable real-time simulation of patient-specific models that predicts disease progression and treatment response. Despite these advances, challenges like data diversity and standardization, model interpretability, and regulatory approval must be further addressed for AI to reach full integration into clinical practice. As AI-driven technologies continue to evolve, interdisciplinary collaboration will be essential to ensure responsible implementation, ultimately advancing precision medicine in cardiovascular care.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145960668","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantitative Chest Computed Tomography and Machine Learning for Subphenotyping Small Airways Disease in Long COVID.","authors":"Rodrigo Caruso Chate, Carlos Roberto Ribeiro Carvalho, Marcio Valente Yamada Sawamura, João Marcos Salge, Eduardo Kaiser Ururahy Nunes Fonseca, Paula Terra Martins Almeida Amaral, Celina de Almeida Lamas, Luis Augusto Visani de Luna, Fernando Uliana Kay, Antonildes Nascimento Assunção Junior, Cesar Higa Nomura","doi":"10.1097/RTI.0000000000000861","DOIUrl":"10.1097/RTI.0000000000000861","url":null,"abstract":"<p><strong>Purpose: </strong>To investigate imaging phenotypes in posthospitalized COVID-19 patients by integrating quantitative CT (QCT) and machine learning (ML), with a focus on small airway disease (SAD) and its correlation with plethysmography.</p><p><strong>Materials and methods: </strong>In this single-center cross-sectional retrospective study, a subanalysis of a larger prospective cohort, 257 adult survivors from the initial COVID-19 peak (mean age, 56±13 y; 49% male) were evaluated. Patients were admitted to a quaternary hospital between March 30 and August 31, 2020 (median length of stay: 16 [8-26] d) and underwent plethysmography along with volumetric inspiratory and expiratory chest CT 6 to 12 months after hospitalization. QCT parameters were derived using AI-Rad Companion Chest CT (Siemens Healthineers).</p><p><strong>Results: </strong>Hierarchical clustering of QCT parameters identified 4 phenotypes among survivors, named \"SAD,\" \"intermediate,\" \"younger fibrotic,\" and \"older fibrotic,\" based on clinical and imaging characteristics. The SAD cluster (n=37, 14%) showed higher residual volume (RV) and RV/total lung capacity (TLC) ratios as well as lower FEF 25-75 /forced vital capacity (FVC) on plethysmography. The older fibrotic cluster (n=42, 16%) had the lowest TLC and FVC values. The younger fibrotic cluster (n=79, 31%) demonstrated lower RV and RV/TLC ratios and higher FEF 25-75 than the other phenotypes. The intermediate cluster (n=99, 39%) exhibited characteristics that were intermediate between those of SAD and fibrotic phenotypes.</p><p><strong>Conclusion: </strong>The integration of inspiratory and expiratory chest CT with quantitative analysis and ML enables the identification of distinct imaging phenotypes in long COVID patients, including a unique SAD cluster strongly associated with specific pulmonary function abnormalities.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145304350","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Preface for Symposium on Artificial Intelligence in Cardiothoracic Imaging.","authors":"Chi Wan Koo","doi":"10.1097/RTI.0000000000000851","DOIUrl":"10.1097/RTI.0000000000000851","url":null,"abstract":"","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":"41 2","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146259753","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bridging the Gap: A Comprehensive Review of Radiology and Pathology in Acute Lung Injury.","authors":"Ana P S Lima, Desiree A Marshall, Eric Morrell, Sudhakar N J Pipavath","doi":"10.1097/RTI.0000000000000837","DOIUrl":"10.1097/RTI.0000000000000837","url":null,"abstract":"<p><p>Acute respiratory distress syndrome (ARDS) is a life-threatening condition characterized by widespread inflammation in the lungs. It is associated with high mortality and morbidity in critically ill patients. ARDS are conditions that cause acute respiratory failure due to noncardiogenic pulmonary edema, leading to severe hypoxemia and diffuse, bilateral lung injury. These conditions represent a spectrum of lung injury with varying severity and complexity. ARDS is a more severe form of ALI. ALI can also describe a range of clinical and paraclinical findings that include one or both pathologic patterns of organizing pneumonia (OP) or diffuse alveolar damage (DAD). The pathologic correlate of ARDS is DAD. This damage can be triggered by various risk factors, including pneumonia, sepsis, trauma, and the inhalation of harmful substances. The alveolar capillary damage that accompanies DAD leads to a loss in barrier function and is associated with the accumulation of fluid into the alveolar space. This fluid accumulation (pulmonary edema), along with subsequent organization and scarring, impairs gas exchange, which leads to hypoxemia and respiratory failure. Despite advances in understanding the pathophysiology of ARDS and improvements in supportive care, the mortality rates from ARDS still range from 25% to 45%. It is crucial to recognize that radiographic and histologic findings in a patient with ARDS can vary significantly depending on the phase of the disease. This is because the pathophysiological processes underlying these conditions evolve over time, leading to changes in both clinical presentation and imaging findings. Misinterpretation of these findings could lead to incorrect diagnoses and inappropriate treatment strategies. Therefore, understanding the temporal evolution of this condition is essential for accurate diagnosis and effective management. Our paper seeks to examine the existing literature focusing on radiology and pathology at different phases of injury and resolution to enhance management of ARDS.</p>","PeriodicalId":49974,"journal":{"name":"Journal of Thoracic Imaging","volume":" ","pages":""},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144477632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}