The differential diagnosis of invasive pulmonary aspergillosis (IPA), pulmonary mucormycosis (PM), bacterial pneumonia (BP) and pulmonary tuberculosis (PTB) are challenging due to overlapping clinical and imaging features. Manual CT lesion segmentation is time-consuming, deep-learning (DL)-based segmentation models offer a promising solution, yet disease-specific models for these infections remain underexplored. We aimed to develop and validate dedicated CT segmentation models for IPA, PM, BP and PTB to enhance diagnostic accuracy. Methods：Retrospective multi-centre data (115 IPA, 53 PM, 130 BP, 125 PTB) were used for training/internal validation, with 21 IPA, 8PM, 30 BP and 31 PTB cases for external validation. Expert-annotated lesions served as ground truth. An improved 3D U-Net architecture was employed for segmentation, with preprocessing steps including normalisations, cropping and data augmentation. Performance was evaluated using Dice coefficients. Results：Internal validation achieved Dice scores of 78.83% (IPA), 93.38% (PM), 80.12% (BP) and 90.47% (PTB). External validation showed slightly reduced but robust performance: 75.09% (IPA), 77.53% (PM), 67.40% (BP) and 80.07% (PTB). The PM model demonstrated exceptional generalisability, scoring 83.41% on IPA data. Cross-validation revealed mutual applicability, with IPA/PTB models achieving > 75% Dice for each other's lesions. BP segmentation showed lower but clinically acceptable performance ( ＞72%), likely due to complex radiological patterns. Disease-specific DL segmentation models exhibited high accuracy, particularly for PM and PTB. While IPA and BP models require refinement, all demonstrated cross-disease utility, suggesting immediate clinical value for preliminary lesion annotation. Future efforts should enhance datasets and optimise models for intricate cases.

<b>Rationale:</b> Quantifying functional small airway disease (fSAD) requires additional expiratory computed tomography (CT) scans, limiting clinical applicability. Artificial intelligence (AI) could enable fSAD quantification from chest CT scans at total lung capacity (TLC) alone (fSAD^TLC). <b>Objectives:</b> To evaluate an AI model for estimating fSAD^TLC, compare it with dual-volume parametric response mapping fSAD (fSAD^PRM), and assess its clinical associations and repeatability in chronic obstructive pulmonary disease (COPD). <b>Methods:</b> We analyzed 2,513 participants from SPIROMICS (the Subpopulations and Intermediate Outcome Measures in COPD Study). Using a randomly sampled subset (<i>n</i> = 1,055), we developed a generative model to produce virtual expiratory CT scans for estimating fSAD^TLC in the remaining 1,458 SPIROMICS participants. We compared fSAD^TLC with dual-volume fSAD^PRM. We investigated univariate and multivariable associations of fSAD^TLC with FEV₁, FEV₁/FVC ratio, 6-minute-walk distance, St. George's Respiratory Questionnaire score, and FEV₁ decline. The results were validated in a subset of patients from the COPDGene (Genetic Epidemiology of COPD) study (<i>n</i> = 458). Multivariable models were adjusted for age, race, sex, body mass index, baseline FEV₁, smoking pack-years, smoking status, and percent emphysema. <b>Measurements and Main Results:</b> Inspiratory fSAD^TLC showed a strong correlation with fSAD^PRM in SPIROMICS (Pearson's <i>R</i> = 0.895) and COPDGene (<i>R</i> = 0.897) cohorts. Higher fSAD^TLC levels were significantly associated with lower lung function, including lower postbronchodilator FEV₁ (in liters) and FEV₁/FVC ratio, and poorer quality of life reflected by higher total St. George's Respiratory Questionnaire scores independent of percent CT emphysema. In SPIROMICS, individuals with higher fSAD^TLC experienced an annual decline in FEV₁ of 1.156 ml (relative decrease; 95% confidence interval [CI], 0.613-1.699; <i>P</i> < 0.001) per year for every 1% increase in fSAD^TLC. The rate of decline in the COPDGene cohort was slightly lower at 0.866 ml/yr (relative decrease; 95% CI, 0.345-1.386; <i>P</i> < 0.001) per 1% increase in fSAD^TLC. Inspiratory fSAD^TLC demonstrated greater consistency between repeated measurements, with a higher intraclass correlation coefficient of 0.99 (95% CI, 0.98-0.99) compared with fSAD^PRM (0.83; 95% CI, 0.76-0.88). <b>Conclusions:</b> Small airway disease can be reliably assessed from a single inspiratory CT scan using generative AI, eliminating the need for an additional expiratory CT scan. fSAD estimation from inspiratory CT correlates strongly with fSAD^PRM, demonstrates a significant association with FEV₁ decline, and offers greater repeatability.

Lung diseases represent one of the most prevalent health challenges globally, necessitating accurate diagnosis to improve patient outcomes. This work presents a novel deep learning-aided lung disease classification framework comprising three key phases: image acquisition, segmentation, and classification. Initially, chest X-ray images are taken from standard datasets. The lung regions are segmented using an Adaptive Recurrent Residual U-Net (AR2-UNet), whose parameters are optimised using Enhanced Pufferfish Optimisation Algorithm (EPOA) to enhance segmentation accuracy. The segmented images are processed using "Attention-based Densenet with Long Short Term Memory(ADNet-LSTM)" for robust categorisation. Investigational results demonstrate that the proposed model achieves the highest classification accuracy of 93.92%, significantly outperforming several baseline models including ResNet with 90.77%, Inception with 89.55%, DenseNet with 89.66%, and "Long Short Term Memory (LSTM)" with 91.79%. Thus, the proposed framework offers a dependable and efficient solution for lung disease detection, supporting clinicians in early and accurate diagnosis.

Interstitial lung disease (ILD) diagnosis is complex, continuously evolving, and increasingly reliant on thin-section chest CT. Multidisciplinary discussion aided by a thorough radiologic review can achieve a high-confidence diagnosis of ILD in the majority of patients and is currently the reference standard for ILD diagnosis. CT also allows the early recognition of interstitial lung abnormalities, possibly reflective of unsuspected ILD and progressive in a substantial proportion of patients. Beyond diagnosis, CT has also become essential for ILD prognostication and follow-up, aiding the identification of fibrotic and progressive forms. The presence of fibrosis is a critical determinant of prognosis, particularly when typical features of usual interstitial pneumonia (UIP) are identified. The UIP-centric imaging approach emphasized in this review is justified by the prognostic significance of UIP, the prevalence of UIP in idiopathic pulmonary fibrosis, and its strong radiologic-pathologic correlation. In nonidiopathic pulmonary fibrosis ILD, progressive pulmonary fibrosis carries clinically significant prognostic and therapeutic implications. With growing evidence and the emergence of novel ILD-related concepts, recent updates of several imaging guidelines aim to optimize the approach to ILD. Artificial intelligence tools are promising adjuncts to the qualitative CT assessment and will likely augment the role of CT in the ILD realm.

Tuberculosis (TB) remains a significant global health challenge, as current diagnostic methods are often resource-intensive, time-consuming, and inaccessible in many high-burden communities, necessitating more efficient and accurate diagnostic methods to improve early detection and treatment outcomes. This study aimed to evaluate the performance of 6 convolutional neural network architectures-Visual Geometry Group-16 (VGG16), VGG19, Residual Network-50 (ResNet50), ResNet101, ResNet152, and Inception-ResNet-V2-in classifying chest x-ray (CXR) images as either normal or TB-positive. The impact of data augmentation on model performance, training times, and parameter counts was also assessed. The dataset of 4200 CXR images, comprising 700 labeled as TB-positive and 3500 as normal cases, was used to train and test the models. Evaluation metrics included accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve. The computational efficiency of each model was analyzed by comparing training times and parameter counts. VGG16 outperformed the other architectures, achieving an accuracy of 99.4%, precision of 97.9%, recall of 98.6%, F1-score of 98.3%, and area under the receiver operating characteristic curve of 98.25%. This superior performance is significant because it demonstrates that a simpler model can deliver exceptional diagnostic accuracy while requiring fewer computational resources. Surprisingly, data augmentation did not improve performance, suggesting that the original dataset's diversity was sufficient. Models with large numbers of parameters, such as ResNet152 and Inception-ResNet-V2, required longer training times without yielding proportionally better performance. Simpler models like VGG16 offer a favorable balance between diagnostic accuracy and computational efficiency for TB detection in CXR images. These findings highlight the need to tailor model selection to task-specific requirements, providing valuable insights for future research and clinical implementations in medical image classification.

Due to the complex anatomical structure and dynamic involution process of the thymus, segmentation and evaluation of the thymus in medical imaging present significant challenges. The aim of this study is to develop a deep-learning tool "Thy-uNET" for automatic segmentation and measurement of the thymus or thymic region on chest CT imaging, and to validate its performance with multicenter data. Utilizing the segmentation and measurement results from two experts, training of Thy-uNET was conducted on training cohort (n = 500). The segmented regions include thymus or thymic region, and 7 features of the thymic region were measured. The automatic segmentation performance was assessed using Dice and Intersection over Union (IOU) on CT data from three test cohorts (n = 286). Spearman correlation analysis and intraclass correlation coefficient (ICC) were used to evaluate the correlation and reliability of the automatic measurement results. Six radiologists with varying levels of experience were invited to participate in a reader study to assess the measurement performance of Thy-uNET and its ability to assist doctors. Thy-uNET demonstrated consistent segmentation performance across different subgroups, with Dice = 0.83 in the internal test set, and Dice = 0.82 in the external test sets. For automatic measurement of thymic features, Thy-uNET achieved high correlation coefficients and ICC for key measurements (R = 0.829 and ICC = 0.841 for CT attenuation measurement). Its performance was comparable to that of radiology residents and junior radiologists, with significantly shorter measurement time. Providing Thy-uNET measurements to readers reduced their measurement time and improved residents' performance in some thymic feature measurements. Thy-uNET can provide reliable automatic segmentation and automatic measurement information of the thymus or thymic region on routine CT, reducing time costs and improving the consistency of evaluations.

This study aimed to compare CT features of COVID-19 and Influenza A pneumonia, develop a diagnostic differential model, and explore a prognostic model for lesion resolution. A total of 446 patients diagnosed with COVID-19 and 80 with Influenza A pneumonitis underwent baseline chest CT evaluation. Logistic regression analysis was conducted after multivariate analysis and the results were presented as nomograms. Machine learning models were also evaluated for their diagnostic performance. Prognostic factors for lesion resolution were analyzed using Cox regression after excluding patients who were lost to follow-up, with a nomogram being created. COVID-19 patients showed more features such as thickening of bronchovascular bundles, crazy paving sign and traction bronchiectasis. Influenza A patients exhibited more features such as consolidation, coarse banding and pleural effusion (P < 0.05). The logistic regression model achieved AUC values of 0.937 (training) and 0.931 (validation). Machine learning models exhibited area under the curve values ranging from 0.8486 to 0.9017. COVID-19 patients showed better lesion resolution. Independent prognostic factors for resolution at baseline included age, sex, lesion distribution, morphology, coarse banding, and widening of the main pulmonary artery. Distinct imaging features can differentiate COVID-19 from Influenza A pneumonia. The logistic discriminative model and each machine - learning network model constructed in this study demonstrated efficacy. The nomogram for the logistic discriminative model exhibited high utility. Patients with COVID-19 may exhibit a better resolution of lesions. Certain baseline characteristics may act as independent prognostic factors for complete resolution of lesions.

This study aims to assess and compare two state-of-the-art deep learning approaches for segmenting four thoracic organs at risk (OAR)-the esophagus, trachea, heart, and aorta-in CT images in the context of radiotherapy planning. We compare a multi-organ segmentation approach and the fusion of multiple single-organ models, each dedicated to one OAR. All were trained using nnU-Net with the default parameters and the full-resolution configuration. We evaluate their robustness with adversarial perturbations, and their generalizability on external datasets, and explore potential biases introduced by expert corrections compared to fully manual delineations. The two approaches show excellent performance with an average Dice score of 0.928 for the multi-class setting and 0.930 when fusing the four single-organ models. The evaluation of external datasets and common procedural adversarial noise demonstrates the good generalizability of these models. In addition, expert corrections of both models show significant bias to the original automated segmentation. The average Dice score between the two corrections is 0.93, ranging from 0.88 for the trachea to 0.98 for the heart. Both approaches demonstrate excellent performance and generalizability in segmenting four thoracic OARs, potentially improving efficiency in radiotherapy planning. However, the multi-organ setting proves advantageous for its efficiency, requiring less training time and fewer resources, making it a preferable choice for this task. Moreover, corrections of AI segmentation by clinicians may lead to biases in the results of AI approaches. A test set, manually annotated, should be used to assess the performance of such methods. Question While manual delineation of thoracic organs at risk is labor-intensive, prone to errors, and time-consuming, evaluation of AI models performing this task lacks robustness. Findings The deep-learning model using the nnU-Net framework showed excellent performance, generalizability, and robustness in segmenting thoracic organs in CT, enhancing radiotherapy planning efficiency. Clinical relevance Automatic segmentation of thoracic organs at risk can save clinicians time without compromising the quality of the delineations, and extensive evaluation across diverse settings demonstrates the potential of integrating such models into clinical practice.

To evaluate the image quality and lung nodule detectability of ultralow-dose CT (ULDCT) with adaptive statistical iterative reconstruction-V (ASiR-V) post-processed using a deep learning image reconstruction (DLIR)-based image domain compared to low-dose CT (LDCT) and ULDCT without DLIR. A total of 210 patients undergoing lung cancer screening underwent LDCT (mean ± SD, 0.81 ± 0.28 mSv) and ULDCT (0.17 ± 0.03 mSv) scans. ULDCT images were reconstructed with ASiR-V (ULDCT-ASiR-V) and post-processed using DLIR (ULDCT-DLIR). The quality of the three CT images was analyzed. Three radiologists detected and measured pulmonary nodules on all CT images, with LDCT results serving as references. Nodule conspicuity was assessed using a five-point Likert scale, followed by further statistical analyses. A total of 463 nodules were detected using LDCT. The image noise of ULDCT-DLIR decreased by 60% compared to that of ULDCT-ASiR-V and was lower than that of LDCT (p < 0.001). The subjective image quality scores for ULDCT-DLIR (4.4 [4.1, 4.6]) were also higher than those for ULDCT-ASiR-V (3.6 [3.1, 3.9]) (p < 0.001). The overall nodule detection rates for ULDCT-ASiR-V and ULDCT-DLIR were 82.1% (380/463) and 87.0% (403/463), respectively (p < 0.001). The percentage difference between diameters > 1 mm was 2.9% (ULDCT-ASiR-V vs. LDCT) and 0.5% (ULDCT-DLIR vs. LDCT) (p = 0.009). Scores of nodule imaging sharpness on ULDCT-DLIR (4.0 ± 0.68) were significantly higher than those on ULDCT-ASiR-V (3.2 ± 0.50) (p < 0.001). DLIR-based image domain improves image quality, nodule detection rate, nodule imaging sharpness, and nodule measurement accuracy of ASiR-V on ULDCT. Question Deep learning post-processing is simple and cheap compared with raw data processing, but its performance is not clear on ultralow-dose CT. Findings Deep learning post-processing enhanced image quality and improved the nodule detection rate and accuracy of nodule measurement of ultralow-dose CT. Clinical relevance Deep learning post-processing improves the practicability of ultralow-dose CT and makes it possible for patients with less radiation exposure during lung cancer screening.