Deep learning-based radiomics of computed tomography angiography to predict adverse events after initial endovascular repair for acute uncomplicated Stanford type B aortic dissection.

PURPOSE: This study aimed to construct a predictive model integrating deep learning-derived radiomic features from computed tomography angiography (CTA) and clinical biomarkers to forecast postoperative adverse events (AEs) in patients with acute uncomplicated Stanford type B aortic dissection (uTBAD) undergoing initial thoracic endovascular aortic repair (TEVAR). METHODS: We retrospectively evaluated 369 patients treated with TEVAR for acute uTBAD from January 2015 to December 2022. A three-dimensional (3D) deep convolutional neural network (CNN) automated radiomic feature extraction from CTA images. Feature selection, using Analysis of Variance (ANOVA) and the Least Absolute Shrinkage and Selection Operator (LASSO) algorithms, refined a radiomic score (Rad-Score). This score, alongside clinical parameters, was modelled via Extreme Gradient Boosting (XGBoost) analysis. Model calibration was assessed by calibration curves. RESULTS: The integration of the Rad-Score with clinical factors including albumin and C-reactive protein levels moderately enhanced predictive efficiency, exhibiting an area under the curve (AUC) of 1.000 (95%CI, 1.000-1.000) in the training cohort and 0.990 (95%CI, 0.966-1.000) in the internal validation cohort. In an independent validation cohort from another hospital, the combined model yielded an AUC of 0.985 (95%CI, 0.965-1.000), with an accuracy, precision, sensitivity, and specificity of 0.92, 0.92, 0.94, and 0.91, respectively. CONCLUSIONS: The synergistic application of deep learning-based radiomics from CTA and clinical indicators holds promise for anticipating AEs post-initial thoracic endovascular aortic repair in patients with acute uTBAD. The clinical utility of the constructed combined model, offering prognostic foresight during follow-up, has been substantiated.

Application of quantitative lung ultrasound instead of CT for monitoring COVID-19 pneumonia in pregnant women: a single-center retrospective study.

BACKGROUND: Computed tomography (CT) is the preferred imaging technique for the evaluation of COVID-19 pneumonia. However, it is not suitable as a monitoring tool for pregnant women because of the risk of ionizing radiation damage to the fetus as well as the possible infection of others. In this study, we explored the value of bedside lung ultrasound (LUS) as an alternative to CT for the detection and monitoring of lung involvement in pregnant women with COVID-19. METHODS: Clinical and LUS data of 39 pregnant women with COVID-19 were retrospectively reviewed. All LUS and CT images were analyzed to summarize the findings and calculate LUS scores and CT scores for each patient. LUS findings were compared with CT, and correlation between LUS scores and CT scores was evaluated. RESULTS: Among the 39 pregnant women, there were 6 mild-type cases, 29 common-type cases, 4 severe-type cases, and no critical-type cases. The most common LUS findings of COVID-19 pneumonia in pregnant women were various grades of multiple B-lines (84.6%), thickened and irregular pleural lines (71.8%), pleural effusion (61.5%) and small multifocal consolidation limited to the subpleural space (35.9%). The mean LUS score at admission was 0 points in mild-type cases, 10.6 points in common-type cases and 15.3 points in severe-type cases (P < 0.01). The correlation between LUS scores and CT was 0.793. All patients were clinically cured and each underwent an average of three LUS follow-ups during hospitalization. The mean LUS score at discharge was 5.6 points lower than that at admission. The consistency of LUS and chest CT during follow-up was 0.652. CONCLUSIONS: Quantitative LUS scoring can effectively instead of CT for detecting and monitoring of COVID-19 pneumonia in pregnant women and protect fetuses from the risk of ionizing radiation.

Semiquantitative lung ultrasound scores in the evaluation and follow-up of critically ill patients with COVID-19: a single-center study.

RATIONALE AND OBJECTIVES: Chest CT is not suitable for critically ill patients with COVID-19 and lung ultrasound (LUS) may play an important role for these patients. In this study, we summarized the findings of LUS and explore the value of semiquantitative LUS scores in evaluation and follow-up of COVID-19 pneumonia. MATERIALS AND METHODS: Retrospectively studied the LUS and chest CT imaging of 128 critically ill patients with COVID-19. The imaging data were reviewed to acquire the LUS and CT scores. The correlation between LUS scores and CT scores were made to evaluate the accuracy of LUS. A cut-off point of LUS score was calculated to distinguish critical-type patients from severe-type patients. LUS follow-up of 72 patients were compared with the gold standard chest CT. RESULTS: The most common LUS features of COVID-19 pneumonia were crowded or coalescent B-lines with multifocal small consolidations in multi-zone. The mean LUS score was 8.1 points in severe-type patients and 15.7 points in critical-type patients (P<0.05). The correlation between LUS scores and CT scores was high (r=0.891, p<0.01) and it was higher in critical-type patients than that in severe-type patients. The LUS score higher than 10.5 points had a 97.4% sensitivity and 75.0% specificity to distinguish critical-type patients. The consistency of LUS and chest CT in follow-up was 0.596, with higher consistency in diagnosis of lesion progression (Kappa values was 0.774). CONCLUSION: Our scoring system provides a more quantitative use of LUS findings and accurate evaluation of lung damage for critically ill patients with COVID-19.

High Resolution CT Imaging Dynamic Follow-Up Study of Novel Coronavirus Pneumonia.

Objective: To explore the clinical characteristics and dynamic follow-up changes of high resolution CT (HRCT) in 270 patients with Coronavirus Disease 2019 (COVID-19) pneumonia. Methods: Two hundred seventy COVID-19 pneumonia patients were retrospectively analyzed, including 146 males and 124 females, with median age of 51 (9,89). The clinical features, laboratory examination indexes and HRCT evolution findings of 270 COVID-19 pneumonia patients were analyzed. Results: 264 cases (95.74%) were positive at the first time nucleic acid test, 6 cases (2.22%) were negative, after multiple inspections, 270 cases (100%) were positive. According to the number of lung segments involved in the lesion, the lesions range from <30% of the lung area (Common type), 30-50% (Severe type), and> 50% (Critical type). At the first CT exam, 136 cases (50.37%) of the common type, 89 cases (32.96%) of the severe type and 45 cases (16.67%) of the critical type. At the second CT exam, 84 cases (31.11%) of the common type, 103 cases (38.15%) of the severe type and 83 cases (30.74%) of the critical type. In the third CT exam, there were 151 cases (55.93%) of the common type, 86 cases (31.85%) of the severe type and 33 cases (12.22%) of the critical type. The differences in image typing were statistically significant (P < 0.05). During this study, a total of 173 patients (64.08%) were recovered after treatment. Conclusion: In some epidemiological backgrounds, CT imaging manifestations and evolutionary characteristics are of great significance for early warning of lung injury, assessment of disease severity, and assistance in clinical typing and post-treatment follow-up.

Correlation Between Diabetic Cognitive Impairment and Diabetic Retinopathy in Patients With T2DM by 1H-MRS.

Objective: To explore the correlation between diabetic cognitive impairment (DCI) and diabetic retinopathy (DR) through examining the cognitive function and the metabolism of the cerebrum in Type 2 diabetes mellitus (T2DM) by 1H-MRS. Methods: Fifty-three patients with T2DM were enrolled for this study. According to the fundus examination, the patients were divided into the DR group (n = 26) and the T2DM without DR group (T2DM group, n = 27). Thirty healthy adults were selected as a control group (HC group, n = 30). Cognitive function was measured by Montreal Cognitive Assessment (MoCA). The peak areas of N-acetylaspartate (NAA), Cho-line (Cho), Creatine (Cr), and Myo-inositol (mI) as well as their ratios were detected by proton magnetic resonance spectroscopy (1H-MRS). The difference analysis between the three groups was performed by one-way ANOVA. When p < 0.05, LSD-t was applied. A partial correlation analysis (with age as a covariate) was used to analyze the correlation between metabolites in the DR group and MoCA scores. Among all T2DM patients, Chi-square test age, gender, education level, BMI, SBP, DBP, FPG, HbA1c, TC, TG, HDL-C, LDL-C, DR, and DCI correlation were measured. Differences were statistically significant while P < 0.05. Results: 1. The scores of MoCA in the DR group or in the T2DM group were significantly less than those in the HC group (F = 3.54, P < 0.05), and the scores of MoCA in the DR group were significantly less than those in the other groups (F = 3.61, P < 0.05). 2. There were significant differences for NAA in the bilateral hippocampus in DR patients, T2DM patients, and healthy controls (P < 0.05). 3. The NAA/Cr was significantly positively correlated with the score of MoCA in DR patients' left hippocampus (r = 0.781, P < 0.01). 4. Chi-square analysis found that there was a correlation between DR and DCI (x 2 = 4.6, df = 1, p = 0.032, plt: 0.05). There was no correlation between other influencing factors and DCI (P > 0.05). Conclusion: DCI is closely correlated with the DR in patients with T2DM. Hippocampal brain metabolism may have some changes in two sides of NAA in patients with DR, 1H-MRS may provide effective imaging strategies and methods for the early diagnosis of brain damage and quantitative assessment cognitive function in T2DM.

Quantitative Aortic Distensibility Measurement Using CT in Patients with Abdominal Aortic Aneurysm: Reproducibility and Clinical Relevance.

Purpose. To investigate the reproducibility of aortic distensibility (D) measurement using CT and assess its clinical relevance in patients with infrarenal abdominal aortic aneurysm (AAA). Methods. 54 patients with infrarenal abdominal aortic aneurysm were studied to determine their distensibility by using 64-MDCT. Aortic cross-sectional area changes were determined at two positions of the aorta, immediately below the lowest renal artery (level 1.) and at the level of its maximal diameter (level 2.) by semiautomatic segmentation. Measurement reproducibility was assessed using intraclass correlation coefficient (ICC) and Bland-Altman analyses. Stepwise multiple regression analysis was performed to assess linear associations between aortic D and anthropometric and biochemical parameters. Results. A mean distensibility of Dlevel 1. = (1.05 ± 0.22) × 10-5 Pa-1 and Dlevel 2. = (0.49 ± 0.18) × 10-5 Pa-1 was found. ICC proved excellent consistency between readers over two locations: 0.92 for intraobserver and 0.89 for interobserver difference in level 1. and 0.85 and 0.79 in level 2. Multivariate analysis of all these variables showed sac distensibility to be independently related (R2 = 0.68) to BMI, diastolic blood pressure, and AAA diameter. Conclusions. Aortic distensibility measurement in patients with AAA demonstrated high inter- and intraobserver agreement and may be valuable when choosing the optimal dimensions graft for AAA before endovascular aneurysm repair.

Correlation between epicardial adipose tissue and severity of coronary artery stenosis evaluated by 64-MDCT.

PURPOSE: The purpose was to investigate the correlation between epicardial adipose tissue (EAT) thickness, EAT volume, and severity of coronary artery stenosis. METHODS: We retrospectively enrolled 188 patients that underwent coronary computed tomography (CT) angiography for clinically suspected coronary artery disease using 64-MDCT. Images were reconstructed using a retrospective electrocardiogram-gated algorithm with 0.625-mm-thick sections. EAT thickness and volume were calculated. RESULTS: The coronary CT angiography showed 106 patients who had coronary artery pathology (178 lesions), 21 patients with moderate stenosis (27 lesions), 12 patients with severe stenosis (18 lesions), and 6 patients with complete occlusion (8 lesions). EAT thickness, EAT volume, and Gensini score were statistically different among groups (FT=32.306, FV=27.743, F=110.483, P=.000). Pearson correlation analysis showed that Gensini score had significantly positive correlation with EAT thickness and volume, respectively. CONCLUSION: EAT thickness and volume demonstrated a positive correlation with severity of coronary artery stenosis.