Lung Imaging Research Articles

A patient with unexplained hypoxemia after a fall diagnosed with platypnea orthodeoxia syndrome: approaches to resolving discrepancies between level of hypoxemia and clinical presentation.

Platypnea orthodeoxia syndrome (POS) is a rare cause of hypoxemia. Diagnosis of POS is challenging, requiring a high index of clinical suspicion, special investigations, and collaboration with multiple specialists. We describe an 86-yr-old male who presented to the emergency department with hip pain after a witnessed fall. He was noted to be hypoxemic at presentation with a peripheral oxygen saturation (SpO2) of 84% on room air, with an inadequate increase in oxygenation after administration of a fractional concentration of inspired oxygen (FIO2) of 1.00. A chest radiograph, computed tomography pulmonary angiogram, and Doppler ultrasound of the liver were unremarkable. In the supine position with an FIO2 of 0.65, his SpO2 and arterial partial pressure of oxygen (PaO2) (96% and 74mm Hg, respectively) increased significantly relative to the seated position (84% and 50mm Hg, respectively). Contrast transthoracic echocardiography (TTE) showed a large patent foramen ovale (PFO) with right-to-left shunt. Transthoracic echocardiography showed rotation of the patient's heart, enabling direct alignment of the inferior vena cava with the PFO, creating a large anatomical right-to-left shunt in the seated position. Right heart catheterization confirmed a large PFO with normal right-sided heart pressures. He was treated with a septal occlusion and his SpO2 in the seated position improved immediately. The patient was discharged home without requiring supplemental oxygen. Platypnea orthodeoxia syndrome is a rare presentation of hypoxemia. Positional changes in oxygenation are the cardinal feature of POS. Discordance between lung imaging and the severity of hypoxemia should prompt investigation for an intracardiac shunt, which can occur in POS even in the absence of increased right-sided heart pressures. Either contrast TTE or transesophageal echocardiography is necessary to make this diagnosis.

Quantitative texture analysis using machine learning for predicting interpretable pulmonary perfusion from non-contrast computed tomography in pulmonary embolism patients

BackgroundPulmonary embolism (PE) is life-threatening and requires timely and accurate diagnosis, yet current imaging methods, like computed tomography pulmonary angiography, present limitations, particularly for patients with contraindications to iodinated contrast agents. We aimed to develop a quantitative texture analysis pipeline using machine learning (ML) based on non-contrast thoracic computed tomography (CT) scans to discover intensity and textural features correlated with regional lung perfusion (Q) physiology and pathology and synthesize voxel-wise Q surrogates to assist in PE diagnosis.MethodsWe retrospectively collected 99mTc-labeled macroaggregated albumin Q-SPECT/CT scans from patients suspected of PE, including an internal dataset of 76 patients (64 for training, 12 for testing) and an external testing dataset of 49 patients. Quantitative CT features were extracted from segmented lung subregions and underwent a two-stage feature selection pipeline. The prior-knowledge-driven preselection stage screened for robust and non-redundant perfusion-correlated features, while the data-driven selection stage further filtered features by fitting ML models for classification. The final classification model, trained with the highest-performing PE-associated feature combination, was evaluated in the testing cohorts based on the Area Under the Curve (AUC) for subregion-level predictability. The voxel-wise Q surrogate was then synthesized using the final selected feature maps (FMs) and model score maps (MSMs) to investigate spatial distributions. The Spearman correlation coefficient (SCC) and Dice similarity coefficient (DSC) were used to assess the spatial consistency between FMs or MSMs and Q-SPECT scans.ResultsThe optimal model performance achieved an AUC of 0.863 during internal testing and 0.828 on the external testing cohort. The model identified a combination containing 14 intensity and textural features that were non-redundant, robust, and capable of distinguishing between high- and low-functional lung regions. Spatial consistency assessment in the internal testing cohort showed moderate-to-high agreement between MSMs and reference Q-SPECT scans, with median SCC of 0.66, median DSCs of 0.86 and 0.64 for high- and low-functional regions, respectively.ConclusionsThis study validated the feasibility of using quantitative texture analysis and a data-driven ML pipeline to generate voxel-wise lung perfusion surrogates, providing a radiation-free, widely accessible alternative to functional lung imaging in managing pulmonary vascular diseases.Clinical trial numberNot applicable.

Pathophysiological Markers of Acute Respiratory Distress Syndrome Severity Are Correlated With Ventilation-Perfusion Mismatch Measured by Electrical Impedance Tomography.

Pulmonary ventilation/perfusion (V/Q) mismatch measured by electrical impedance tomography (EIT) is associated with the outcome of patients with the acute respiratory distress syndrome (ARDS), but the underlying pathophysiological mechanisms have not been fully elucidated. The present study aimed to verify the correlation between relevant pathophysiological markers of ARDS severity and V/Q mismatch. Prospective observational study. General ICU of a university-affiliated hospital. Deeply sedated intubated adult patients with ARDS under controlled mechanical ventilation. Measures of V/Q mismatch by EIT, respiratory mechanics, gas exchange, lung imaging, and plasma biomarkers. Unmatched V/Q units were assessed by EIT as the fraction of ventilated nonperfused plus perfused nonventilated lung units. At the same time, plasma biomarkers with proven prognostic and mechanistic significance for ARDS (carbonic anhydrase 9 [CA9], hypoxia-inducible factor 1 [HIF1], receptor for advanced glycation endproducts [RAGE], angiopoietin 2 [ANG2], gas exchange, respiratory mechanics, and quantitative chest CT scans were measured. Twenty-five intubated ARDS patients were included with median unmatched V/Q units of 37.1% (29.2-49.2%). Unmatched V/Q units were correlated with plasma levels of CA9 (rho = 0.47; p = 0.01), HIF1 (rho = 0.40; p = 0.05), RAGE (rho = 0.46; p = 0.02), and ANG2 (rho = 0.42; p = 0.03). Additionally, unmatched V/Q units correlated with plateau pressure (r = 0.38; p = 0.05) and with the number of quadrants involved on chest radiograph (r = 0.73; p < 0.01). Regional unmatched V/Q units were correlated with the corresponding fraction of poorly aerated lung tissue (r = 0.62; p = 0.01) and of lung tissue weight (rho: 0.51; p = 0.04) measured by CT scan. In ARDS patients, unmatched V/Q units are correlated with pathophysiological markers of lung epithelial and endothelial dysfunction, increased lung stress, and lung edema. Unmatched V/Q units could represent a comprehensive marker of ARDS severity, reflecting the complex organ pathophysiology and reinforcing their prognostic significance.

Malignancy Detection in Lung and Colon Histopathology Images by Transfer Learning with Class Selective Image Processing

Malignancy Detection in Lung and Colon Histopathology Images by Transfer Learning with Class Selective Image Processing

Beneficial effect of sequential treatment with high-dose steroids and short-course oral glucocorticoids in patients with severe influenza virus-associated pneumonia

Accumulating evidence supports that glucocorticoid treatment for viral pneumonia (VPA) can shorten the disease course and improve survival. However, currently, the use of glucocorticoids in treating VPA remains controversial. Moreover, a unified standard for the dosage and duration of glucocorticoid therapy has not been presented in published articles. A retrospective analysis was conducted in patients who were hospitalized for severe influenza virus-associated pneumonia, and they received sequential treatment with high-dose glucocorticoids and short-course oral glucocorticoids. Patients were followed up for 3 months. A total of 11 patients were included in the study (average age 56 years). There was no gender difference, but age and underlying diseases could be risk factors for severe influenza virus-associated pneumonia. The types of viruses causing pneumonia included influenza A/B. The main clinical symptoms of patients were fever, cough, sputum production, and dyspnea. Chest computed tomography showed multiple ground-glass shadows in the lobes, and the presence of bacterial and fungal infections was accompanied by consolidation shadows. After glucocorticoid therapy, the symptoms improved. None of the patients underwent tracheal intubation, and all survived. After a 3-month follow-up, lung CT absorption in all patients had reached more than 80%, and lung imaging absorption in 20% patients was complete. No serious complications occurred in any of the patients. Sequential treatment with high-dose steroids and short-course oral glucocorticoids may be helpful for reducing the tracheal intubation rate and mortality rate in patients with severe influenza virus-associated pneumonia. Additionally, short-course oral glucocorticoids may reduce pulmonary fibrosis in patients with severe influenza virus-associated pneumonia without any serious complications.

Lung MRI: The future is now.

Tomographic imaging via MRI can capture structural abnormalities at high resolution and reveal functional abnormalities if acquired during respiration. In infants and children with chronic lung diseases, MRI can serve as a safe and effective tool to characterize cardiorespiratory disease where PFTs are not feasible. Functional lung imaging via hyperpolarized Xe MRI has been shown to have very high sensitivity to regional airway obstruction, changes in alveolar-airspace size/structure, and gas-exchange due to abnormalities in the interstitium or blood-flow; the technique is routinely used in ages 5 and up. This summary outlines each of the functional measures currently available using Xe MRI.

Correlation between elderly patients with COPD and the impact on immunity in tuberculosis patients: A retrospective study.

The prevalence of chronic obstructive pulmonary disease (COPD) and tuberculosis (TB) is increasing globally, yet their comprehensive impact on the immune system remains underexplored. This study aimed to provide a thorough assessment of the immune status of patients with COPD and tuberculosis (TB-COPD), including their pulmonary conditions, immune cell responses, and changes in lymphocyte subpopulations. A total of 151 patients with TB-COPD patients were included, and clinical data were compared between the TB-COPD group and a group of TB patients without COPD (TB-NCOPD). Lung imaging findings and peripheral blood immune cell levels were compared between the 2 groups. Flow cytometry was used to analyze the absolute counts of lymphocyte subpopulations. The incidence of pulmonary lobe lesions and cavitation in the TB-COPD group aged 70 years or older was significantly higher than that in the control group. At the immune cell level, patients with TB-COPD showed a significant reduction in total lymphocytes, CD4+ T lymphocytes and CD4+/CD8+ ratio. Regardless of COPD status, the CD4+ T cell count in the CMV-infected group was significantly lower than that in the uninfected group (P < .05). Additionally, the CD4+/CD8+ ratio in the COPD + TB CMV + group was significantly lower than that in the uninfected group. Analysis of lymphocyte subpopulations revealed a decrease in the counts of CD4+ T lymphocytes in patients with TB-COPD, potentially associated with the chronic inflammatory state induced by COPD. The one-month treatment outcomes showed that the improvement rate in the control group was 70.58%, which was significantly higher than the 38.92% in the COPD + TB group (P < .001). We observed a significant increase in the number of pulmonary cavity patients in the TB-COPD group, suggesting that COPD may be a potential risk factor for the formation of pulmonary cavities in patients with TB. At the immune cell level, TB-COPD patients showed a notable decrease in lymphocytes and CD4+ T lymphocytes, implying that COPD combined with pulmonary TB may significantly affect the immune system, leading to a reduction in the counts of key immune cells.

Pulmonary Emphysema Disease Classification Model Using Deep Structured Techniques

Emphysema is normally employed in the sense of morphology, and thus, imaging the modalities has a significant part in detecting the disorder. Most importantly, the High-Resolution Computed Tomography (HRCT) is a trustworthy component for illustrating the emphysema pathology. Hence, the pulmonary emphysema diagnosis at the starting stage is very important and can secure the life of a human. When there is a failure in the technique, there is a possibility of the death that displays the importance of effective diagnosis. The endotracheal intubation’s delay is prevented by implementing distinct mechanisms that result in various requirements. The presented work develops a deep learning-aided pulmonary emphysema disease classification model that is implemented to defeat the above difficulties. At first, the images are accumulated from a standard dataset. Then, the collected images are provided to the phase of segmentation. Here, the newly developed Adaptive Trans-Residual Dense UNet (ATRDUNet) is employed to perform the segmentation process. Also, the recommended ATRDUNet model helps to locate the disease in the appropriate location. Moreover, the performance of the segmentation process is enriched to fine-tune the parameter using the Fitness Revised Position of Wild Horse Optimizer (FRP-WHO). While optimizing the parameters effectively, it tends to handle the complex issues to strengthen the model accurately. The segmented results from the ATRDUNet are passed to Vision Transformer-based Residual DenseNet with LSTM layer (ViTrans-RDLSTM) for getting the final classified outcome. To this end, the analysis of the suggested model is correlated to find the effectiveness of the recommended model. However, these results are examined with the help of standard performance measures. Also, the experimentation is carried out using the standard “Automatic emphysema detection using weakly labeled HRCT lung images” dataset to attain better outcomes in the developed model. The findings of the recommended approach show 95% in terms of accuracy. Also, the statistical analysis of the developed model shows 23.9%, 12.3%, 15.2%, and 13.0% better performance than BWO, CO, HBA, and WHO in terms of mean analysis to show enriched performance.

Mapping the amplitude and phase of dissolved 129Xe red blood cell signal oscillations with keyhole spectroscopic lung imaging.

To assess the regional amplitude and phase of dissolved 129Xe red blood cell (RBC) signal oscillations in the lung vasculature with keyhole spectroscopic imaging and to compare with previous methodology, which does not account for oscillation phase. 129Xe gas transfer was measured with a four-echo 3D radial spectroscopic imaging sequence. Keyhole reconstruction-based RBC signal oscillation amplitude mapping was applied retrospectively to data acquired from 28 healthy volunteers, 4 chronic thromboembolic pulmonary hypertension (CTEPH) patients, and 5 patients who were hospitalized due to COVID-19 pneumonia and had residual lung abnormalities. Using a sliding window keyhole reconstruction, maps of RBC oscillation amplitude were corrected for regional phase difference. Repeatability of the phase-adjusted oscillation amplitude was assessed in 8 healthy volunteers across three scans. With sliding window keyhole reconstruction, regional phase differences were observed in the RBC signal oscillations: mean phase = (0.27 ± 0.19) rad in healthy volunteers, (0.24 ± 0.13) rad in CTEPH patients, and (0.33 ± 0.19) rad in patients with post-COVID-19 residual lung abnormality. The oscillation amplitude and phase maps were more heterogeneous (i.e., they showed increased coefficient of variation) for the CTEPH patients. The RBC oscillation amplitude was repeatable, and the mean three-scan coefficient of variation was smaller when the phase adjustment was made (0.07 ± 0.04 compared with 0.16 ± 0.05). Sliding window keyhole reconstruction of radial dissolved 129Xe imaging reveals regional phase differences in the RBC oscillations, which are not captured when performing two phase keyhole reconstruction. This regional phase information may reflect the hemodynamic effect of the cardiac pulse wave in the pulmonary microvasculature.

Capacitively Coupled Electrical Impedance Tomography in Lung Imaging

Capacitively Coupled Electrical Impedance Tomography in Lung Imaging

Multi-View Soft Attention-Based Model for the Classification of Lung Cancer-Associated Disabilities.

Background: The detection of lung nodules at their early stages may significantly enhance the survival rate and prevent progression to severe disability caused by advanced lung cancer, but it often requires manual and laborious efforts for radiologists, with limited success. To alleviate it, we propose a Multi-View Soft Attention-Based Convolutional Neural Network (MVSA-CNN) model for multi-class lung nodular classifications in three stages (benign, primary, and metastatic). Methods: Initially, patches from each nodule are extracted into three different views, each fed to our model to classify the malignancy. A dataset, namely the Lung Image Database Consortium Image Database Resource Initiative (LIDC-IDRI), is used for training and testing. The 10-fold cross-validation approach was used on the database to assess the model's performance. Results: The experimental results suggest that MVSA-CNN outperforms other competing methods with 97.10% accuracy, 96.31% sensitivity, and 97.45% specificity. Conclusions: We hope the highly predictive performance of MVSA-CNN in lung nodule classification from lung Computed Tomography (CT) scans may facilitate more reliable diagnosis, thereby improving outcomes for individuals with disabilities who may experience disparities in healthcare access and quality.

SYRMEP beamline: state of the art, upgrades and future prospects

SYRMEP is the hard X-ray imaging beamline of Elettra synchrotron offering X-ray full-field techniques, micro-computed tomography (microCT) and phase-contrast modality in the energy range 10–40 keV. The beamline operates in a multidisciplinary research context spanning from biomedical applications to botany, from zoology to food technology and cultural heritage, from materials engineering to geology and earth science. Thanks to the flexibility of SYRMEP setup, in situ experiments can be performed as well, novel imaging methods can be developed and implemented in a synergical manner with interested users and collaborators. SYRMEP peculiar wide beam together with the long sample-to-detector distance enables multiscale phase-contrast studies with optimized contrast and spatial resolution on rather large specimens, such as human lung phantoms. This is particularly relevant in view of future clinical lung imaging foreseen in the framework of Elettra 2.0 program. Here, the current beamline features and recent upgrades are illustrated, an overview of the imaging methods routinely offered to SYRMEP users’ community is presented, and the outlook for the new beamline SYRMEP-Life Science (SYRMEP-LS) is reported.

Multifractal Analysis in Age-Based Classification for COVID-19 Patients’ CT-Scan Images with Different Noise Levels

Recently, many number of people various countries contracted the COVID-19 (C-19) disease. The identification of this harmful disease and other pneumonia diseases is an important process in the medical world. CT-Scan is mainly prescribed to predict the severity of lung diseases affected by C-19 and other pneumonia diseases. Generally, a CT-Scan image is represented as a grayscale image (GI). Biomedical-oriented GIs are more complex and they are too difficult to recognize the status of lung diseases directly from the experimental images. To address this type of medical problem, a multifractal approach can be applied to analyze and illustrate the GIs in detail. Therefore, the multifractal dimensional analysis is used to diagnose and explore the vehemence of the contamination levels in the lungs. In this study, the complexity of CT-Scan images of C-19 patients is analyzed in order to perform a comparison in terms of age and noise levels. It was also discovered that the intricate of CT-Scan images is considerably varied for patients aged 50 and above when compared to younger subjects. This comparative study indicates that the deadly virus affects elderly persons compared to youngsters. The proposed age-based classification is supported by image processing techniques, qualitative measures, and statistical tools. The obtained results are demonstrated graphically by Generalized Fractal Dimensions (GFD) spectrum, 3D visualization, ANOVA table, and box plots to expose the rate of discrimination of complexity in CT-Scan lung images.

Rapid precision targeting of nanoparticles to lung via caveolae pumping system in endothelium.

Modern medicine seeks precision targeting, imaging and therapy to maximize efficacy and avoid toxicities. Nanoparticles (NPs) have tremendous yet unmet clinical potential to carry and deliver imaging and therapeutic agents systemically with tissue precision. But their size contributes to rapid scavenging by the reticuloendothelial system and poor penetration of key endothelial cell (EC) barriers, limiting target tissue uptake, safety and efficacy. Here we discover the ability of the EC caveolae pumping system to outpace scavenging and deliver NPs rapidly and specifically into the lungs. Gold and dendritic NPs are conjugated to antibodies targeting caveolae of the lung microvascular endothelium. SPECT-CT imaging and biodistribution analyses reveal that rat lungs extract most of the intravenous dose within minutes to achieve precision lung imaging and targeting with high lung concentrations exceeding peak blood levels. These results reveal how much ECs can both limit and promote tissue penetration of NPs and the power and size-dependent limitations of the caveolae pumping system. This study provides a new retargeting paradigm for NPs to avoid reticuloendothelial system uptake and achieve rapid precision nanodelivery for future diagnostic and therapeutic applications.

A full-scale lung image segmentation algorithm based on hybrid skip connection and attention mechanism

The segmentation accuracy of the lung images is affected by the occlusion of the front background objects. To address this problem, we propose a full-scale lung image segmentation algorithm based on hybrid skip connection and attention mechanism (HAFS). The algorithm uses yolov8 as the underlying network and enhancement of multi-layer feature fusion by incorporating dense and sparse skip connections into the network structure, and increased weighting of important features through attention gates. Finally the proposed algorithm was applied to the lung datasets Montgomery County chest X-ray and Shenzhen chest X-ray. The experimental results show that the proposed algorithm improves the precision, recall, pixel accuracy, Dice, mIoU, mAP and GFLOPs metrics compared to the comparison algorithms, which proves the advancement and effectiveness of the proposed algorithm.

6857 Adrenal Incidentaloma In A Patient With Surgically Treated Lung Adenocarcinoma A Decade After Complete Remission

Abstract Disclosure: L. Kaur: None. R. Kaur: None. F. Haider: None. A. Iqbal: None. N. Patel: None. Adrenal incidentaloma is an adrenal mass &amp;gt;1 cm in size commonly discovered incidentally in up to 5-7% of patients on imaging performed for reasons other than adrenal causes. Metastasis to the adrenal glands is common due to their rich blood supply, with lung cancer being the most common primary site, followed by the stomach, esophagus and the liver. Bilateral adrenal metastasis is common, however isolated metachronous or synchronous adrenal metastasis is rare, with an incidence of 1-6%. We present a case of incidentally discovered isolated right adrenal mass ten years after curative surgical resection of right lung adenocarcinoma. A 74 year old male with history significant for adenocarcinoma of the right upper lung lobe stage 1A , negative for metastasis status post lobectomy and no adjuvant chemotherapy or radiation 10 years ago presented to the hospital for his routine follow up evaluation. His vitals were within normal limits and physical examination was unremarkable. The follow up diagnostic CT imaging was done which showed an incidental 2.1 x 1.6 x 1.2 cm right adrenal mass without any pathological lesion or lymphadenopathy. A review of CT scan done one year ago revealed no prior adrenal mass or other suspicious findings. The patient was referred to endocrinology for evaluation of the adrenal mass. He denied abdominal or right flank pain, headaches, palpitations, elevated blood pressure, sweating, muscle weakness, weight gain/loss and was healthy appearing. Lab work up including biochemical studies for hormonal over production, including aldosterone-renin ratio, adrenocorticotropic hormone, plasma catecholamines with urine metanephrines and urine cortisol were within physiological ranges. CT scan was repeated after 2 months which revealed right adrenal mass of 3.2 x 1.8 x 1.7 cm in size with heterogeneity and high pre contrast density. Given the patient's history of lung cancer, imaging characteristics and rapid interval increase in size, there was high suspicion for metastatic disease. Other differentials considered due to fast growth rate included adrenocortical carcinoma and less likely adrenal adenoma. A decision was made to proceed with right adrenalectomy for definitive resection and histological diagnosis secondary to quickly growing adrenal mass with no lymphadenopathy. The patient underwent laparoscopic adrenalectomy and the final pathological report was consistent with metastatic lung adenocarcinoma. Postoperatively the patient recovered well. At 2 week and 3 month follow up he reported no concerns. Isolated adrenal metastasis is unusual and rare but may occur in lung cancer patients years after curative resection of the primary tumor and hence long term annual follow up imaging and monitoring is recommended for prompt diagnosis and timely treatment. Presentation: 6/2/2024

LungHist700: A dataset of histological images for deep learning in pulmonary pathology

Accurate detection and classification of lung malignancies are crucial for early diagnosis, treatment planning, and patient prognosis. Conventional histopathological analysis is time-consuming, limiting its clinical applicability. To address this, we present a dataset of 691 high-resolution (1200 × 1600 pixels) histopathological lung images, covering adenocarcinomas, squamous cell carcinomas, and normal tissues from 45 patients. These images are subdivided into three differentiation levels for both pathological types: well, moderately, and poorly differentiated, resulting in seven classes for classification. The dataset includes images at 20x and 40x magnification, reflecting real clinical diversity. We evaluated image classification using deep neural network and multiple instance learning approaches. Each method was used to classify images at 20x and 40x magnification into three superclasses. We achieved accuracies between 81% and 92%, depending on the method and resolution, demonstrating the dataset’s utility.

Sex Differences in Allergic Bronchopulmonary Aspergillosis and its Impact on Exacerbations.

The impact of sex on allergic bronchopulmonary aspergillosis (ABPA) outcomes remains uncertain. We retrospectively included ABPA subjects per the revised International Society for Human and Animal Mycology ABPA working group criteria over 13years. We compared the clinical features, lung function, immunological tests, imaging, and ABPA exacerbation rates between men and women. Our primary objective was to assess whether women experience higher ABPA exacerbations than men. We included 731 ABPA subjects (mean age, 34.5years; 49.5% women). Women with ABPA were older and had underlying asthma more frequently than men. There was no difference in lung function, immunological investigations, and imaging between men and women. ABPA exacerbations occurred in a slightly higher proportion of women than men (44.5% vs. 38.2%) but did not reach statistical significance (p = 0.09). We did not find a significant sex difference in ABPA exacerbation rates. Prospective studies should confirm our findings.

Active learning based on multi-enhanced views for classification of multiple patterns in lung ultrasound images

There are several main patterns in lung ultrasound (LUS) images, including A-lines, B-lines, consolidation and pleural effusion. LUS images of healthy lungs typically only exhibit A-lines, while other patterns may emerge or coexist in LUS images associated with different lung diseases. The accurate categorization of these primary patterns is pivotal for effective lung disease screening. However, two challenges complicate the classification task: the first is the inherent blurring of feature differences between main patterns due to ultrasound imaging properties; and the second is the potential coexistence of multiple patterns in a single case, with only the most dominant pattern being clinically annotated. To address these challenges, we propose the active learning based on multi-enhanced views (MEVAL) method to achieve more precise pattern classification in LUS. To accentuate feature differences between multiple patterns, we introduce a feature enhancement module by applying vertical linear fitting and k-means clustering. The multi-enhanced views are then employed in parallel with the original images, thus enhancing MEVAL’s awareness of feature differences between multiple patterns. To tackle the patterns coexistence issue, we propose an active learning strategy based on confidence sets and misclassified sets. This strategy enables the network to simultaneously recognize multiple patterns by selectively labeling of a small number of images. Our dataset comprises 5075 LUS images, with approximately 4% exhibiting multiple patterns. Experimental results showcase the effectiveness of the proposed method in the classification task, with accuracy of 98.72%, AUC of 0.9989, sensitivity of 98.76%, and specificity of 98.16%, which outperforms than the state-of-the-art deep learning-based methods. A series of comprehensive ablation studies suggest the effectiveness of each proposed component and show great potential in clinical application.

Preoperative markers for identifying CT ≤2 cm solid nodules of lung adenocarcinoma based on image deep learning.

The solid pattern is a highly malignant subtype of lung adenocarcinoma. In the current era of transitioning from lobectomy to sublobar resection for the surgical treatment of small lung cancers, preoperative identification of this subtype is highly important for patient surgical approach selection and long-term prognosis. A total of 1489 patients with clinical stage IA1-2 primary lung adenocarcinoma were enrolled. Based on patient clinical characteristics and lung imaging features obtained via deep learning, highly correlated diagnostic factors were identified through LASSO regression and decision tree analysis. Subsequently, a logistic model and nomogram were constructed. A restricted cubic spline (RCS) was used to calculate the optimal inflection point of quantitative data and the differences between the groups. The three-dimensional proportion of solid component (PSC), sex, and smoking status was identified as being highly correlated diagnostic factors for solid predominant adenocarcinoma. The logistic model had good prediction efficiency, and the area under the ROC curve was 0.85. Decision curve analysis demonstrated that the application of diagnostic factors can improve patient outcomes. RCS analysis indicated that the proportion of solid adenocarcinomas increased by 4.6 times when the PSC was ≥72%. A PSC of 72% is a good cutoff point. The preoperative diagnosis of solid-pattern adenocarcinoma can be confirmed by typical imaging features and clinical characteristics, assisting the thoracic surgeon in developing a more precise surgical plan.