Chest Chest Computed Tomography Research Articles

Leveraging deep transfer learning and explainable AI for accurate COVID-19 diagnosis: Insights from a multi-national chest CT scan study

The COVID-19 pandemic has emerged as a global health crisis, impacting millions worldwide. Although chest computed tomography (CT) scan images are pivotal in diagnosing COVID-19, their manual interpretation by radiologists is time-consuming and potentially subjective. Automated computer-aided diagnostic (CAD) frameworks offer efficient and objective solutions. However, machine or deep learning methods often face challenges in their reproducibility due to underlying biases and methodological flaws. To address these issues, we propose XCT-COVID, an explainable, transferable, and reproducible CAD framework based on deep transfer learning to predict COVID-19 infection from CT scan images accurately. This is the first study to develop three distinct models within a unified framework by leveraging a previously unexplored large dataset and two widely used smaller datasets. We employed five known convolutional neural network architectures, both with and without pretrained weights, on the larger dataset. We optimized hyperparameters through extensive grid search and 5-fold cross-validation (CV), significantly enhancing the model performance. Experimental results from the larger dataset showed that the VGG16 architecture (XCT-COVID-L) with pretrained weights consistently outperformed other architectures, achieving the best performance, on both 5-fold CV and independent test. When evaluated with the external datasets, XCT-COVID-L performed well with data with similar distributions, demonstrating its transferability. However, its performance significantly decreased on smaller datasets with lower-quality images. To address this, we developed other models, XCT-COVID-S1 and XCT-COVID-S2, specifically for the smaller datasets, outperforming existing methods. Moreover, eXplainable Artificial Intelligence (XAI) analyses were employed to interpret the models’ functionalities. For prediction and reproducibility purposes, the implementation of XCT-COVID is publicly accessible at https://github.com/cbbl-skku-org/XCT-COVID/.

A clinical-radiomics nomogram based on automated segmentation of chest CT to discriminate PRISm and COPD patients

PurposeIt is vital to develop noninvasive approaches with high accuracy to discriminate the preserved ratio impaired spirometry (PRISm) group from the chronic obstructive pulmonary disease (COPD) groups. Radiomics has emerged as an image analysis technique. This study aims to develop and confirm the new radiomics-based noninvasive approach to discriminate these two groups. MethodsTotally 1066 subjects from 4 centers were included in this retrospective research, and classified into training, internal validation or external validation sets. The chest computed tomography (CT) images were segmented by the fully automated deep learning segmentation algorithm (Unet231) for radiomics feature extraction. We established the radiomics signature (Rad-score) using the least absolute shrinkage and selection operator algorithm, then conducted ten-fold cross-validation using the training set. Last, we constructed a radiomics signature by incorporating independent risk factors using the multivariate logistic regression model. Model performance was evaluated by receiver operating characteristic (ROC) curve, calibration curve, and decision curve analyses (DCA). ResultsThe Rad-score, including 15 radiomic features in whole-lung region, which was suitable for diffuse lung diseases, was demonstrated to be effective for discriminating between PRISm and COPD. Its diagnostic accuracy was improved through integrating Rad-score with a clinical model, and the area under the ROC (AUC) were 0.82(95 %CI 0.79–0.86), 0.77(95 %CI 0.72–0.83) and 0.841(95 %CI 0.78–0.91) for training, internal validation and external validation sets, respectively. As revealed by analysis, radiomics nomogram showed good fit and superior clinical utility. ConclusionsThe present work constructed the new radiomics-based nomogram and verified its reliability for discriminating between PRISm and COPD.

Prognostic importance of extensive coronary calcium on lung cancer screening chest computed tomography.

Low-dose chest computed tomography (CT) is used for lung cancer screening, but can also detect coronary artery disease as coronary artery calcium. We sought to determine the prevalence and prognostic utility of coronary artery calcium in a population at high risk of cancer. We reviewed CT scans from consecutive participants screened for lung cancer between March 2017 and November 2018 as part of the Ontario Health Lung Cancer Screening Pilot for People at High Risk. We quantified coronary artery calcium using an estimated Agatston score. We identified the composite primary outcome of all-cause death and cardiovascular events using linked electronic medical record data from The Ottawa Hospital to December 2023. Among 1486 people who underwent screening CT, coronary artery calcium was detected in 1232 (82.9%) and was extensive in 439 (29.5%). On multivariable analysis, extensive coronary artery calcium was associated with the composite primary outcome (hazard ratio [HR] 2.13, 95% confidence interval [CI] 1.35-3.38), all-cause death (HR 2.39, 95% CI 1.34-4.27), and cardiovascular events (HR 2.06, 95% CI 1.13-3.77). Extensive coronary artery calcium remained predictive of cardiovascular events after we adjusted for noncardiovascular death as a competing risk (HR 2.05, 95% CI 1.09-3.85). Among people undergoing low-dose chest CT for lung cancer screening, extensive coronary artery calcium was an independent predictor of all-cause death and cardiovascular events, even after adjustment for noncardiovascular death. The opportunity to identify and reduce risks from coronary artery disease may represent an additional benefit of lung cancer screening.

Radiomics analysis of thoracic vertebral bone marrow microenvironment changes before bone metastasis of breast cancer based on chest CT

BackgroundBone metastasis from breast cancer significantly elevates patient morbidity and mortality, making early detection crucial for improving outcomes. This study utilizes radiomics to analyze changes in the thoracic vertebral bone marrow microenvironment from chest computerized tomography (CT) images prior to bone metastasis in breast cancer, and constructs a model to predict metastasis. Methods: This study retrospectively gathered data from breast cancer patients who were diagnosed and continuously monitored for five years from January 2013 to September 2023. Radiomic features were extracted from the bone marrow of thoracic vertebrae on non-contrast chest CT scans. Multiple machine learning algorithms were utilized to construct various radiomics models for predicting the risk of bone metastasis, and the model with optimal performance was integrated with clinical features to develop a nomogram. The effectiveness of this combined model was assessed through receiver operating characteristic (ROC) analysis as well as decision curve analysis (DCA). Results: The study included a total of 106 patients diagnosed with breast cancer, among whom 37 developed bone metastases within five years. The radiomics model’s area under the curve (AUC) for the test set, calculated using logistic regression, is 0.929, demonstrating superior predictive performance compared to alternative machine learning models. Furthermore, DCA demonstrated the potential of radiomics models in clinical application, with a greater clinical benefit in predicting bone metastasis than clinical model and nomogram. Conclusion: CT-based radiomics can capture subtle changes in the thoracic vertebral bone marrow before breast cancer bone metastasis, offering a predictive tool for early detection of bone metastasis in breast cancer.

Features of Lung Cysts in Birt-Hogg-Dubé Syndrome from Patients with Multiple Lung Cysts.

High-resolution chest computed tomography (CT) is a crucial assessment tool for diagnosis of Birt-Hogg-Dubé syndrome (BHD). This study aims to analyze the differences of lung cyst between BHD and other cystic lung diseases. From January 2020 to December 2022, we retrospectively screened all patients who underwent chest CT at Gangnam Severance Hospital. We included the patients with multiple lung cysts for the analysis of chest CT images. Over a three-year period, out of 52,823 patients who underwent a chest CT scan, 301 patients (0.6%) with multiple lung cysts were enrolled, of which 24 (8.0%) were diagnosed with BHD. Notably, 95.8% and 83.3% of BHD patients exhibited bilateral cysts and basal predominance, and had larger cysts with a maximal diameter (averaging 32.1mm [interquartile range 26.5mm to 43.5mm]) than lymphangioleiomyomatosis (17.0mm [13.2;19.1], p<0.001) and others' group (11.3mm [7.9;17.0], p<0.001). Additionally, 95.8% of BHD patients has a diverse range in cyst sizes and morphologies. Multivariate logistic regression analysis identified bilateral cysts (OR 12.393, 95% CI: 1.613-274.682, p=0.038), basal predominance (OR 8.511, 95% CI: 2.252-39.392, p=0.002), maximum diameter (OR 1.053, 95% CI: 1.009-1.108, p=0.032), and diversity of morphology (OR 19.513, 95% CI: 2.833-398.119, p=0.010) as factors associated with BHD diagnosis. By stepwise selection, a multivariate prediction model for BHD diagnosis was established, demonstrating a sensitivity of 95.83%, a specificity of 81.22%, and an AUC of 0.951 (95% CI: 0.914-0.987). Distinguishing features of lung cyst from other cystic lung diseases include bilateral cysts, basal dominance, large size, and irregular shape. The predictive model can assist in identifying undiagnosed patients with BHD.

Impact of individually tailored contrast medium on vascular attenuation in chest CT: a randomized controlled trial.

Individually tailored contrast medium (CM) may improve vascular image quality in chest computed tomography (CT). To evaluate vascular attenuation in chest CT by comparing CM dose calculations using lean body mass (LBM) and body surface area (BSA) with a fixed-dose protocol. Patients referred for contrast-enhanced chest CT were categorized as either normal, muscular, or overweight. Patients were accordingly randomized into three CM dosing protocols: fixed-dose group (n = 87), LBM group (n = 92), and BSA group (n = 93). Of the patients, 94% in the fixed-dose group, 99% in the LBM group, and 98% in the BSA group achieved optimal vascular attenuation. In the overweight category, lower attenuation was demonstrated in the fixed-dose group compared to the LBM group (P = 0.032) and the BSA group (P = 0.010). In the fixed-dose group, vascular attenuation showed a negative correlation with total body weight for all body composition categories. In the LBM group, a positive correlation was observed between attenuation and total body weight in the muscular category (P = 0.041), while a negative correlation was noted for the overweight category in the BSA group (P = 0.049). Fixed-dose CM protocol exhibited larger variations in vascular attenuation between patients of various body weights and body compositions compared to tailored CM doses based on LBM and BSA.

Chest computed tomography for patients with sepsis in the emergency intensive care unit

Sepsis is a systemic inflammatory response syndrome (SIRS) caused by infection, which may lead to multiple organ dysfunction in susceptible patient. The most frequently involved organs/systems include the lung, kidney and circulation system. It is well established that sepsis is a risk factor for acute lung injury. While overt pulmonary infiltrates can be well captured by human operators, subtle structural changes of the lung might be ignored. Since the advantage of chest computed tomography (CT) is its capability of providing fine structural changes in high spatial resolution, the study of chest CT by means of computer science may provide further insights into the underlying pathophysiology. The integration of chest CT into the study of sepsis is limited partly due to the lack of well-curated database. The study aims to establish a database comprising detailed clinical tabular data, as well as the raw chest CT images. The database is intended to support a wide array of research studies involving radiomics in sepsis patients, helping to reduce barriers to the reproducibility of clinical research.

Thymoma-associated paraneoplastic encephalitis resolved by robotic-assisted radical thymectomy: a case report

BackgroundWe present a rare case of autoimmune encephalitis associated with an incidentally discovered thymoma. Limited case reports exist of thymoma-associated paraneoplastic encephalitis. The unique clinical presentation and subsequent surgical intervention offer valuable insights into the management of similar cases.Case presentationA 41-year-old female presented with irritability, personality changes, fatigue, and seizures. Physical examination and bloodwork on first encounter were unremarkable. Neurological evaluation was done, and a brain magnetic resonance imaging (MRI) with and without contrast demonstrated numerous hyperintensity foci within bilateral cerebral cortices on T2-/FLAIR-weighted images affecting only the supratentorial brain, findings highly suggestive of encephalitis. A chest computed tomography (CT) revealed an anterior mediastinal mass compatible with a thymoma. Cerebrospinal fluid from a lumbar puncture was unremarkable. Paraneoplastic autoantibody evaluation was positive, and the patient was diagnosed with autoimmune encephalitis believed to be associated with an incidentally discovered thymoma. Given her rapid neurological deterioration, the thoracic surgery team was urgently consulted.Patient underwent robotic-assisted radical thymectomy with prompt resolution of symptoms. During her 2-week postoperative visit, she was neurologically intact with no residual deficits. Follow-up MRI at 6 months demonstrated resolution of abnormal signal intensities, with continued improvement noted 1-year post-operation. Chest CT scans at 6 and 12 months revealed no residual or recurrent thymic tissue.ConclusionsThis case illustrates the connection between thymoma and autoimmune neurological disorders like encephalitis. It emphasizes the significance of comprehensive neurological and oncological assessments for patients with unexplained neurological symptoms. Prompt diagnosis and treatment led to positive outcomes in this case, underscoring the potential for full recovery in similar cases.

Abstract 4127513: Cardiopulmonary long-term effects 6, 18 and 30 months after severe covid-19 infection

Background: SARS-CoV-2 infection affects the cardiopulmonary system in both the acute and long-term phase. This study aimed to comprehensively assess symptoms and potential long-term impairments 6, 18 and 30 months in patients previously hospitalized for severe Covid-19 infection. Methods: This prospective registry included patients hospitalized for PCR-confirmed Covid-19 infection. Approximately 6 months post-discharge, follow-up examination included patient history, clinical examination, echocardiography, electrocardiogram, cardiac magnetic resonance imaging (cMRI), chest computed tomography (CT) scan, pulmonary function test (PFT), six-minute walk test (6MWT) and a comprehensive laboratory panel. Patients with pathologic findings during the first visit underwent a second (at 18 months) and third (at 30 months) follow-up examination. Those without pathologic findings or who refused further medical examinations were contacted via phone to inquire about symptoms. Results: Between July 2020 and April 2022, 200 patients (91% general ward, 9% intensive care unit) were recruited. Due to dropouts, the second visit was conducted in 170 patients, and the third visit in 139 (74 in person, 65 via telephone). Long Covid criteria were fulfilled by 73% at 6 months, 52% at 18 months and 49% at 30 months post-discharge, with fatigue being the most common symptom (Figure 1). Echocardiography at 6 months showed impaired left ventricular function in 15 patients, with normalization in 80% at 18 months and further 66% at 30 months (Figure 2). cMRI revealed pericardial effusions in 28 patients at 6 months, which resolved in 47% at 18 months and in further 60% at 30 months. Signs of peri- or myocarditis were present in 7 patients at 6 months and were resolved in all 4 patients who attended control studies at 18 months. Chest CT scans at 6 months identified post-infectious residues in 41 patients, with full recovery in 20% at 18 months without further normalization after 30 months. The length of in-hospital stay was identified as a significant predictor for persisting Long Covid 6 months after discharge (95% CI: 1.005 - 1.12, p=0.03). Conclusion: While the prevalence of Long Covid decreased over time, a significant symptom burden persisted at 6, 18 and even 30 months after severe Covid-19 infection. Structural and functional abnormalities were less frequent compared to reported symptoms, posing a challenge in substantiating the causes of these symptoms.

Coronary artery calcium on lung cancer radiation planning CT aids cardiovascular risk assessment

BackgroundPatients with non-small cell lung cancer (NSCLC) undergoing thoracic radiation are at high cardiovascular risk. Semiquantitative assessment of coronary artery calcification (CAC) on baseline planning non-gated chest computed tomography (CT) scans may help further risk stratify patients.ObjectivesThis study aimed to characterize the association between CAC and major adverse cardiovascular events (MACE; myocardial infarction or stroke) and assess the utility of semiquantitative assessment of CAC.MethodsPatients with NSCLC with non-contrast planning chest CT scans were evaluated for CAC. Planning scans were visually graded using the CAC-DRS method, stratifying patients into no, mild, moderate, and severe CAC groups. Demographics, comorbidities, and radiation treatment characteristics were gathered, and CAC groups were assessed for the incidence of MACE after initiation of radiation therapy.ResultsOut of 137 patients, 39 patients had no CAC, and 98 patients had any CAC (38 with mild CAC, 34 with moderate CAC, and 26 with severe CAC). There was 1 MACE event in the no CAC group and 11 in patients with any CAC. The presence of CAC was associated with increased MACE compared to no CAC (p = 0.034). Semiquantitative CAC analysis correlated with formal CAC scoring.ConclusionThere is a significantly lower incidence of MACE in patients with no CAC on planning CT compared to patients with higher burdens of CAC. CAC burden is an important risk factor for adverse cardiovascular events in patients with NSCLC undergoing thoracic radiation. Semiquantitative CAC scoring may be a useful proxy when formal CAC scoring is unavailable.

The Study of Lung Ultrasound, Chest CT and Chest X-ray in Children with COVID-19 Hospitalized in Mofid Children's Hospital During a Six-Month Period

Background: Coronavirus disease (COVID-19) is caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). The primary diagnostic tool for pediatric COVID-19 patients is polymerase chain reaction (PCR). Chest imaging findings in pediatric COVID-19 cases are often normal or mild. The correlation of lung ultrasound (LUS) with chest computed tomography (CT) and chest X-ray (CXR) in pediatric COVID-19 patients has not been extensively studied. Objectives: This study aimed to assess the correlation of LUS with chest CT and CXR in detecting COVID-19 pneumonia in children. Methods: This single-center cross-sectional study included patients under 18 years of age diagnosed with COVID-19 by PCR or by abnormal chest CT findings suggestive of COVID-19, admitted to Mofid Pediatric Hospital between December 2021 and August 2022. All patients underwent LUS. Approximately half of the patients also had a chest CT, and CXR was performed on 35 patients. Lung ultrasound and CT scores were calculated, and the correlation between these scores was evaluated. The correlation between LUS and CXR was also analyzed. Results: Sixty patients were included, of whom 21 were female, with a mean age of 4.9 ± 4.0 years. A significant correlation was observed between LUS and CT scores (correlation coefficient = 0.467, P = 0.011). Lesion distribution was similar between LUS and CT. However, no significant correlation was found between LUS scores and CXR findings (P = 0.392). Sixteen out of 19 patients with normal CXRs had LUS scores ≤ 4. Notably, three patients with normal CXRs had LUS scores of 8, 14, and 14. Conclusions: Lung ultrasound was more sensitive than CXR and demonstrated a significant correlation with CT. Lung ultrasound may serve as a safe alternative to CT for detecting COVID-19 pneumonia in pediatric patients.

Self-supervised learning for chest computed tomography: training strategies and effect on downstream applications.

Self-supervised pre-training can reduce the amount of labeled training data needed by pre-learning fundamental visual characteristics of the medical imaging data. We investigate several self-supervised training strategies for chest computed tomography exams and their effects on downstream applications. We benchmark five well-known self-supervision strategies (masked image region prediction, next slice prediction, rotation prediction, flip prediction, and denoising) on 15M chest computed tomography (CT) slices collected from four sites of the Mayo Clinic enterprise, United States. These models were evaluated for two downstream tasks on public datasets: pulmonary embolism (PE) detection (classification) and lung nodule segmentation. Image embeddings generated by these models were also evaluated for prediction of patient age, race, and gender to study inherent biases in models' understanding of chest CT exams. The use of pre-training weights especially masked region prediction-based weights, improved performance, and reduced computational effort needed for downstream tasks compared with task-specific state-of-the-art (SOTA) models. Performance improvement for PE detection was observed for training dataset sizes as large as with a maximum gain of 5% over SOTA. The segmentation model initialized with pre-training weights learned twice as fast as the randomly initialized model. While gender and age predictors built using self-supervised training weights showed no performance improvement over randomly initialized predictors, the race predictor experienced a 10% performance boost when using self-supervised training weights. We released self-supervised models and weights under an open-source academic license. These models can then be fine-tuned with limited task-specific annotated data for a variety of downstream imaging tasks, thus accelerating research in biomedical imaging informatics.

Radiation dose assessment of pediatric computed tomography of the chest: the need to consider patient size.

To evaluate the radiation dose of chest computed tomography (CT) examinations of pediatric patients and the extent to which volume CT dose index (CTDIvol) underestimates radiation dose in comparison to size specific dose estimates (SSDE). Single-center, retrospective study of consecutive unenhanced pediatric (age<18years) chest CTs between October 2015 and October 2016. Radiation dose as well as demographic and clinical data were recorded from 133 chest CTs. Patients were grouped into 4 categories based on mean effective diameter of the chest. SSDE was generated for each patient according to the water equivalent and effective diameter and compared to CTDIvol. Factors associated with higher radiation doses were assessed. CTDIvol underestimated radiation dose by 54.7%, 47.6%, 40.2%, and 31.2% (P<.001) for effective diameter groups 1 to 4, respectively, when compared to SSDE (calculated according to the water equivalent). When calculated according to the effective diameter, CTDIvol underestimated radiation dose by 47.6%, 39.4%, 27%, and 12.3% (P<.001) for effective diameter groups 1 to 4, respectively, when compared to SSDE. CT dose parameters, age, weight, Dw, and mean effective diameter were variables associated with higher radiation doses. CTDIvol systematically underestimated radiation dose in comparison to SSDE in pediatric patients submitted to chest CT and should not be used as the primary parameter to monitor CT protocols in these patients. SSDE calculated according to effective diameter also underestimates the radiation dose compared to SSDE calculated based on water equivalent.

Respiratory manifestations of pediatric granulomatosis with polyangiitis: A 12-year experience from a tertiary care facility.

Pediatric granulomatosis with polyangiitis (GPA) is associated with several pulmonary manifestations. This study aims to describe these manifestations at time of diagnosis and longitudinally at a tertiary-care pediatric hospital. We performed a retrospective chart review of patients with GPA treated at our facility between 1 January 2010 through 31 December 2021. We collected baseline demographics, reported symptoms, imaging findings, pulmonary function tests (PFTs), and laboratory data at time of diagnosis. Data were collected using 6-month observation intervals to follow recurrence of respiratory manifestations, testing during recurrence, and resultant treatment modifications. Of 13 patients treated for GPA during the study period, 12 developed respiratory tract involvement. A total of 87 6-month observation periods were analyzed. At time of diagnosis, 83% (10/12) of subjects reported respiratory symptoms, 92% (11/12) had abnormal chest computed tomography (CT) imaging, and 42% (5/12) had abnormal PFTs. Fewer than half of the patients were seen by pulmonology within 6 months of diagnosis. Eight subjects (75%) had respiratory manifestations during subsequent observation periods. Chest CT or PFTs were obtained in 23/44 (52%) of observations periods with respiratory symptoms, with pulmonary consultation in only 9/44 (20%). This is the first US study to describe respiratory manifestations in pediatric GPA patients longitudinally, finding they are common and frequently recurrent. Our cohort had almost universally abnormal imaging at diagnosis regardless of respiratory symptoms. Early collaboration with pediatric pulmonology in the care of GPA patients may allow rheumatology teams to efficiently evaluate recurrent symptoms and address concomitant lung disease.

Chest computed tomography-based artificial intelligence-aided latent class analysis for diagnosis of severe pneumonia.

There is little literature describing the artificial intelligence (AI)-aided diagnosis of severe pneumonia (SP) subphenotypes and the association of the subphenotypes with the ventilatory treatment efficacy. The aim of our study is to illustrate whether clinical and biological heterogeneity, such as ventilation and gas-exchange, exists among patients with SP using chest computed tomography (CT)-based AI-aided latent class analysis (LCA). This retrospective study included 413 patients hospitalized at Xinhua hospital diagnosed with SP from June 1, 2015 to May 30, 2020. AI quantification results of chest CT and their combination with additional clinical variables were used to develop LCA models in an SP population. The optimal subphenotypes were determined though evaluating statistical indicators of all the LCA models, and clinical implications of them such as guiding ventilation strategies were further explored by statistical methods. The two-class LCA model based on AI quantification results of chest CT can describe the biological characteristics of the SP population well and hence yielded the two clinical subphenotypes. Patients with subphenotype-1 had milder infections (P <0.001) than patients with subphenotype-2 and had lower 30-day (P <0.001) and 90-day (P <0.001) mortality, and lower in-hospital (P=0.001) and 2-year (P <0.001) mortality. Patients with subphenotype-1 showed a better match between the percentage of non-infected lung volume (used to quantify ventilation) and oxygen saturation (used to reflect gas exchange), compared with patients with subphenotype-2. There were significant differences in the matching degree of lung ventilation and gas exchange between the two subphenotypes (P<0.001). Compared with patients with subphenotype-2, those with subphenotype-1 showed a relatively better match between CT-based AI metrics of the non-infected region and oxygenation, and their clinical outcomes were effectively improved after receiving invasive ventilation treatment. A two-class LCA model based on AI quantification results of chest CT in the SP population particularly revealed clinical heterogeneity of lung function. Identifying the degree of match between ventilation and gas-exchange may help guide decisions about assisted ventilation.

Artificial Intelligence in Lung Cancer Detection: Clinical and Economic Assessment of Retrospective CT Analysis Two Years Post-COVID-19 Pandemic

BACKGROUND: The use of chest computed tomography (CT) scans in Krasnoyarsk Krai, Russia, has increased since 2020, during the COVID-19 pandemic. This period also saw a 5.2% decrease in lung cancer (LC) incidence. The potential for missed LC cases has led to the investigation of new diagnostic methods, including the use of artificial intelligence (AI) for analyzing retrospective data. AIM: The study aimed to evaluate the effectiveness of an AI algorithm in identifying patients at high risk for LC using chest CT data from the COVID-19 pandemic. METHODS: A retrospective analysis was conducted on chest CTs from patients diagnosed with COVID-19 in the Krasnoyarsk region, using scans from November 1, 2020, to February 28, 2021. The AI algorithm "Chest-IRA" was applied to detect pulmonary nodules larger than 100 mm³. Radiologists classified the nodules detected by AI into three categories based on LC probability. The economic assessment of the AI algorithm included salary costs and potential savings from early LC treatment. RESULTS: Out of 10,500 CTs, the AI algorithm found nodules in 484 cases. Of these, 192 were highly likely to have LC, 103 showed no signs, and 60 were inconclusive. 112 patients with high or intermediate risk did not seek treatment. The AI confirmed lung cancer in 100 cases, 28.2% of those detected. Early-stage LC was found in 35% of cases, while 65% were at later stages. AI could save about 25.04 months of radiologist work, costing 2.43 million RUB. Early detection savings are estimated at 10.6 to 12.5 million RUB per 10,500 CTs, with a five-year economic impact of 259.4 to 305.1 million RUB. CONCLUSION: AI has proven effective in identifying pulmonary nodules amidst COVID-19, highlighting its potential to improve early detection and diagnostic accuracy, leading to earlier and more precise treatments.

Streptococcus salivarius pneumonia-associated pneumomediastinum: a case report and literature review

BackgroundStreptococcus salivarius is an opportunistic pathogen, and there have been no reported cases of Streptococcus salivarius pneumonia to date. Pneumomediastinum is usually secondary to tracheal or esophageal injury and is very rare as a complication of pneumonia. We report a case of Streptococcus salivarius pneumonia complicated by pneumomediastinum, aiming to enhance clinicians’ awareness of rare pathogens and uncommon complications in pneumonia.Case presentationThe patient, a 36-year-old male, presented with a persistent cough and sputum production for one week, accompanied by a sore throat that had developed just one day prior. Chest computed tomography (CT) disclosed pneumomediastinum alongside obstructive atelectasis in the left lower lobe. Streptococcus salivarius infection was conclusively identified through bronchoalveolar lavage metagenomic next-generation sequencing (mNGS), as well as smear and culture analyses. The patient was administered intravenous amoxicillin-clavulanate potassium for a duration of seven days as part of the anti-infection regimen. Given the stability of the patient’s respiratory and circulatory systems, a tube drainage procedure was deemed unnecessary. Post-treatment, the patient’s clinical symptoms notably improved. A subsequent chest CT scan revealed the re-expansion of the left lower lung and near-complete resolution of pneumomediastinum.ConclusionThere are numerous pathogens that can cause pneumonia. While focusing on common pathogens, it is important not to overlook rare ones. When considering infections from rare pathogens, it is recommended to promptly perform a bronchoscopy and submit bronchoalveolar lavage fluid for mNGS to improve pathogen detection rates. During the diagnosis and treatment of pneumonia, it is crucial to be vigilant for rare complications. When a patient presents with symptoms such as dyspnea or subcutaneous emphysema, it is advisable to immediately perform a chest CT scan to rule out pneumomediastinum.

Quantification of pulmonary edema using automated lung segmentation on computed tomography in mechanically ventilated patients with acute respiratory distress syndrome

BackgroundQuantification of pulmonary edema in patients with acute respiratory distress syndrome (ARDS) by chest computed tomography (CT) scan has not been validated in routine diagnostics due to its complexity and time-consuming nature. Therefore, the single-indicator transpulmonary thermodilution (TPTD) technique to measure extravascular lung water (EVLW) has been used in the clinical setting. Advances in artificial intelligence (AI) have now enabled CT images of inhomogeneous lungs to be segmented automatically by an intensive care physician with no prior radiology training within a relatively short time. Nevertheless, there is a paucity of data validating the quantification of pulmonary edema using automated lung segmentation on CT compared with TPTD.MethodsA retrospective study (January 2016 to December 2021) analyzed patients with ARDS, admitted to the intensive care unit of the Department of Anesthesiology and Critical Care Medicine, University Hospital Mannheim, who underwent a chest CT scan and hemodynamic monitoring using TPTD at the same time. Pulmonary edema was estimated using manually and automated lung segmentation on CT and then compared to the pulmonary edema calculated from EVLW determined using TPTD.Results145 comparative measurements of pulmonary edema with TPTD and CT were included in the study. Estimating pulmonary edema using either automated lung segmentation on CT or TPTD showed a low bias overall (− 104 ml) but wide levels of agreement (upper: 936 ml, lower: − 1144 ml). In 13% of the analyzed CT scans, the agreement between the segmentation of the AI algorithm and a dedicated investigator was poor. Manual segmentation and automated segmentation adjusted for contrast agent did not improve the agreement levels.ConclusionsAutomated lung segmentation on CT can be considered an unbiased but imprecise measurement of pulmonary edema in mechanically ventilated patients with ARDS.

Automatic Virtual Contrast-Enhanced CT Synthesis Using Dual-Energy CT and Residual U-Net with Attention Module for Detecting Pulmonary Hilar Lymphadenopathy

Rationale and ObjectivesTo propose an automatic virtual contrast-enhanced chest computed tomography (CT) synthesis using dual-energy CT and a Residual U-Net with an attention module to detect clinically significant hilar lymphadenopathy. Materials and MethodsWe conducted a retrospective analysis of 2,082 patients who underwent dual-energy chest CT scans. Our approach utilized a Residual U-Net combined with a Convolutional Block Attention Module (CBAM) to transform non-contrast CT images into virtual contrast-enhanced CT images. We evaluated the effectiveness of our method through quantitative and qualitative analyses and an observer study involving thoracic radiologists, focusing on the detection of significant hilar lymph nodes. ResultsOur method achieved an average peak signal-to-noise ratio (PSNR) of 25.082, a structural similarity index (SSIM) of 0.833, and mutual information (MI) of 1.568. The mean absolute error (MAE), mean squared error (MSE), and root mean squared error (RMSE) were reported as 0.040, 0.023, and 0.102, respectively. Compared to other methods, our proposed approach demonstrated superior performance across all evaluation metrics. In the observer study, our method exhibited a higher diagnostic accuracy for detecting hilar lymphadenopathy (69.2%) compared to the Residual U-Net-based GAN with CBAM (53.7%). ConclusionThe integration of dual-energy computed tomography (DECT) with a Residual U-Net framework augmented by CBAM presents a highly effective technique for generating synthetic contrast-enhanced chest CT images. This novel approach significantly enhances the detection of clinically significant hilar lymphadenopathy, underscoring its potential clinical utility.

Cardiopulmonary long-term effects 6, 18 and 30 months after severe COVID-19 infection

Abstract Background SARS-CoV-2 infection affects the cardiopulmonary system in both the acute and long-term phase. This study aimed to comprehensively assess symptoms and potential longterm impairments 6, 18 and 30 months in patients previously hospitalized for severe Covid-19 infection. Methods This prospective registry included patients hospitalized for PCR-confirmed Covid-19 infection. Approximately 6 months post-discharge, follow-up examination included patient history, clinical examination, echocardiography, electrocardiogram, cardiac magnetic resonance imaging (cMRI), chest computed tomography (CT) scan, pulmonary function test (PFT), six-minute walk test (6MWT) and a comprehensive laboratory panel. Patients with pathologic findings during the first visit underwent a second (at 18 months) and third (at 30 months) follow-up examination. Those without pathologic findings or who refused further medical examinations were contacted via phone to inquire about symptoms. Results Between July 2020 and April 2022, 200 patients (91% general ward, 9% intensive care unit) were recruited. Due to dropouts, the second visit was conducted in 170 patients and the third visit in 139 (74 in person, 65 via telephone). Long Covid criteria were fulfilled by 73% at 6 months, 52% at 18 months and 49% at 30 months post-discharge, with fatigue being the most common symptom (Figure 1). Echocardiography at 6 months showed impaired left ventricular function in 15 patients, with normalization in 80% at 18 months and further 66% at 30 months (Figure 2). cMRI revealed pericardial effusions in 28 patients at 6 months, which resolved in 47% at 18 months and in further 60% at 30 months. Signs of peri- or myocarditis were present in 7 patients at 6 months and were resolved in all 4 patients who attended control studies at 18 months. PFT at 6 months detected reduced vital capacity in 17 patients, of whom 10% returned to normal at 18 months and a further 66% showed normalization at 30 months. Chest CT scans at 6 months identified post-infectious residues in 41 patients, with full recovery in 20% at 18 months without further normalization after 30 months. The length of in-hospital stay was identified as a significant predictor for persisting Long Covid 6 months after discharge (95% CI: 1.005 - 1.12, p=0.03). No predictors have been found for Long Covid at 18 and 30 months. Conclusion While the prevalence of Long Covid decreased over time, a significant symptom burden persisted at 6, 18 and even 30 months after severe Covid-19 infection. Structural and functional abnormalities were less frequent compared to reported symptoms, posing a challenge in substantiating the causes of these symptoms.