Quantitative Imaging in Medicine and Surgery最新文献

Background: Right ventricular (RV) functional abnormalities significantly impact the clinical presentation and prognosis of hypertrophic cardiomyopathy (HCM) patients. The aim of this study was to assess RV systolic function across the three myocardial layers using two-dimensional speckle tracking echocardiography (2D-STE) and its derived layer-specific strain (LSS) in HCM patients with or without RV hypertrophy (RVH).

Methods: This cross-sectional study consecutively enrolled 102 HCM patients (HCMs) and 50 healthy controls (HCs). Patients with HCM were divided into two groups based on the presence or absence of RVH, according to RV wall thickness (RVWT). RV and left ventricular (LV) echocardiographic parameters were assessed using conventional echocardiography, tissue Doppler imaging (TDI), and 2D-STE.

Results: Significant differences were observed in RV global longitudinal strain (RVGLS), RVGLSendo, RVGLSmid, RVGLSepi, RV free wall strain (RVFWS), RVFWSendo, RVFWSmid, and RVFWSepi among the three groups (all P<0.05). The HCMs with or without RVH showed significantly lower absolute values of RVGLS, RVGLSendo, RVGLSmid, RVGLSepi, RVFWS, RVFWSendo, RVFWSmid, and RVFWSepi compared to the HCs (all P<0.05). Moreover, HCM patients with RVH exhibited significantly lower absolute values of RVGLS, RVGLSendo, RVGLSmid, RVGLSepi, RVFWS, RVFWSendo, RVFWSmid, and RVFWSepi compared to the HCMs without RVH (all P<0.05). Furthermore, RVWT was significantly associated with tricuspid annular plane systolic excursion (TAPSE), RVGLS, RVGLSendo, RVGLSmid, RVGLSepi, RVFWS, RVFWSendo, RVFWSmid, or RVFWSepi (all P<0.05). LV end-systolic volume index (LVESVi) was significantly associated with RVGLS, RVGLSmid, RVGLSepi, RVFWS, RVFWSendo, RVFWSmid, and RVFWSepi (all P<0.05). LVGLS was significantly associated with TAPSE, RVGLS, RVGLSendo, RVGLSmid, RVGLSepi, and RVFWSepi (all P<0.05). There were positive correlations between chest pain or New York Heart Association (NYHA) grade, and RVGLS, RVGLSendo, RVGLSmid, and RVGLSepi, and positive correlations between arrhythmia and RVGLS, RVGLSendo, RVGLSmid, RVGLSepi, RVFWS, RVFWSendo, RVFWSmid, and RVFWSepi (all P<0.05).

Conclusions: This study revealed that absolute values of the RV systolic function decreased significantly in HCM patients, as characterized by reduced absolute values of RVGLS and RVFWS across all three myocardial layers, particularly in HCM patients with RVH. LSS of RVGLS and RVFWS could serve as indicators of increased risk for chest pain, NYHA grade, and arrhythmias.

Background: Lung adenocarcinoma is a leading subtype of lung cancer, and accurate subclassification is critical for guiding clinical management. However, existing computed tomography (CT)-based approaches often fail to capture how lesion characteristics evolve across adjacent slices. This study aims to address this gap by integrating central-slice anatomy with dynamic radiomics information.

Methods: In this multi-center retrospective study, CT lesions were allocated into training, internal testing, and external validation cohorts. For each lesion, a 64×64 patch from the automatically selected central slice (IMG1) and a multi-layer radiomic sequence derived from PyRadiomics descriptors were extracted. After feature filtering and least absolute shrinkage and selection operator (LASSO) selection, 14 key descriptors were arranged slice-by-slice and modeled with a BiLSTM-Attention network. The resulting temporal representation was transformed into a two-dimensional Markov Random Field (MRF) map and fused with IMG1 to form a dual-channel input for the proposed dynamic radiomics fusion network (DRFN). DRFN was compared with two baselines: a convolutional neural network (CNN) using only IMG1 and a BiLSTM-Attention model using only radiomics sequences. Performance was assessed using area under the curve (AUC), accuracy, precision, recall, and F1-score, and interpretability was explored with class activation mapping (CAM).

Results: On an independent test cohort, DRFN achieved per-class AUCs of 0.97 (minimally invasive adenocarcinoma), 0.99 (adenocarcinoma in situ), and 0.95 (invasive adenocarcinoma). Grad-CAM heatmaps confirmed that DRFN consistently attended to lesion cores, spiculated margins, and adjacent vascular structures-mirroring radiologists' diagnostic reasoning. Compared to a single-channel CNN (AUCs: 0.77-0.88) and a BiLSTM-Attention-only model (AUCs: 0.88-0.94), DRFN demonstrated superior sensitivity, specificity, and generalizability.

Conclusions: By fusing static anatomical information with dynamic radiomics evolution maps, DRFN offers both high classification accuracy and transparent interpretability. Our framework thus holds promise as an intelligent diagnostic aid for lung adenocarcinoma subtyping in clinical practice.

Background: Static functional networks of the brain are disrupted in minimal hepatic encephalopathy (MHE), but their dynamic alterations are unknown. This observational study utilized multilayer network analysis to investigate dynamic network characteristics in hepatitis B cirrhosis (HBC) with or without MHE and assess their association with neurocognitive function (registry: https://ctms.xyeyy.com/iit/project/index; trial registration number: LYF20240134; date: 2024-07-31).

Methods: A total of 33 HBC patients [15 non-MHE (NMHE; HBC patients without MHE) and 18 MHE individuals], as well as 36 matched healthy controls (HCs), underwent neurocognitive assessments, resting-state functional magnetic resonance imaging (rs-fMRI), and clinical examinations. Dynamic network variations were quantified using network switching rates, and their relationships with clinical and neurocognitive parameters were evaluated.

Results: Both HBC patients with and without MHE status exhibited a range of altered network switching rates compared to HCs. Specifically, differences were observed in subnetworks including somatomotor network (SMN), dorsal attention network (DAN), ventral attention network (VAN), frontoparietal network (FPN), and subcortical network (SUB), as well as in nodal regions such as the right precentral gyrus (rPrG), left fusiform gyrus (lFuG), left inferior parietal lobule (lIPL), and right hippocampus (rHipp). Furthermore, altered global level and lFuG switching rates positively correlated with Psychometric Hepatic Encephalopathy Score (PHES) (r=0.341, 0.339; P=0.004, 0.004, respectively, Bonferroni corrected).

Conclusions: This study firstly revealed that HBC patients exhibited imbalanced functional dynamics in subnetworks and nodes, suggesting a potential mechanism underlying cerebral dysfunction in MHE.

Background: Endovascular recanalization has gained increasing application in the treatment of patients with chronic carotid artery occlusion (CCAO). Despite its common use in the endovascular treatment of CCAO, it is unclear whether local anesthesia is associated with low peri-procedural complications. This study aimed to investigate the safety and efficacy of endovascular recanalization under local anesthesia for patients with symptomatic CCAO.

Methods: Consecutive patients with symptomatic CCAO who underwent endovascular recanalization attempts under local anesthesia were retrospectively enrolled in this study. The recanalization attempts were discontinued if the patient reported intolerance. The recanalization success rate and peri-procedural complications were recorded. The underlying reasons for recanalization failure were investigated. A meta-analysis of the recanalization success and peri-procedural complication rates of the CCAO patients who underwent recanalization under local anesthesia was also conducted.

Results: In total, 69 patients were enrolled in the study. The recanalization success and peri-procedural complication rates were 68.1% and 2.9%, respectively. No iatrogenic vascular complications occurred. Recanalization failed in 22 patients, of whom, the cavernous segment could not be navigated in 14 patients, and the extracranial segment could not be transversed in eight patients. A total of seven studies, including this study, were included in the meta-analysis. Overall, the recanalization success rate was 66.2% [95% confidence interval (CI): 61.2% to 70.1%, I²=41%]. The peri-procedural and vascular complication rates were 5.5% (95% CI: 4.4% to 8.2%, I²=36%) and 4.9% (95% CI: 2.8% to 8.0%, I²=0%), respectively.

Conclusions: Endovascular recanalization for CCAO under local anesthesia was feasible and was associated with a low rate of peri-procedural complications.

Background: Low back pain (LBP) is a leading cause of disability worldwide, yet the contribution of paraspinal muscle fat infiltration remains uncertain. This study aimed to quantitatively assess paraspinal muscle fat fraction using the Iterative Decomposition of water and fat with Echo Asymmetry and Least-squares estimation for Imaging Quantification (IDEAL-IQ) sequence in patients with LBP and to explore its relationships with pain severity, disability, and demographic factors such as age, sex, and body mass index (BMI).

Methods: This retrospective cross-sectional study included 124 patients (64 males, 60 females; mean age, 38.3±12.6 years). Fat fraction values of the multifidus, erector spinae, and psoas muscles were measured bilaterally at L1-L2 through L4-L5 levels using IDEAL-IQ. Pain and disability were assessed with the Visual Analog Scale (VAS) and Oswestry Disability Index (ODI). Correlation and multiple regression analyses were conducted to determine associations, adjusting for age, sex, and BMI. Interobserver agreement between two radiologists was evaluated using the intraclass correlation coefficient (ICC).

Results: Women demonstrated consistently higher muscle fat fractions than men across all lumbar levels. In females, multifidus fat fraction ranged from 10.1% to 10.5% (vs. 5.3-5.7% in males), erector spinae ranged from 8.8% to 12.1% (vs. 5.0-7.1% in males), and psoas ranged from 5.4% to 5.7% (vs. 4.1-5.0% in males). These differences were statistically significant for all lumbar levels except psoas at L2-L3 and L4-L5. BMI showed a positive correlation with fat fraction across all lumbar levels (r=0.29-0.50, all P<0.001). VAS scores were positively correlated with multifidus (r=0.44-0.54), erector spinae (r=0.41-0.56), and psoas (r=0.35-0.42) fat fractions (all P<0.001). Similarly, ODI scores demonstrated significant positive correlations with multifidus (r=0.44-0.54), erector spinae (r=0.42-0.56), and psoas (r=0.42-0.52) fat fractions at every lumbar level (all P<0.001). Regression analysis confirmed fat fraction as an independent predictor of both VAS and ODI (adjusted R2=0.22-0.32, P<0.001). Interobserver agreement was excellent (ICC =0.92-0.98).

Conclusions: Quantitative assessment of paraspinal muscles and the psoas muscle fat fraction using IDEAL-IQ revealed significant associations with pain severity and disability, independent of demographic factors. If validated by larger prospective studies, incorporating IDEAL-IQ sequences into routine lumbar magnetic resonance imaging (MRI) protocols and reporting paraspinal muscle fat fraction may contribute to a more comprehensive and quantitative evaluation of patients with LBP.

Background: Clear visualization of the portal vein is essential for managing liver diseases such as those requiring transplantation and portal hypertension. However, abdominal computed tomography angiography (CTA) is often challenged by poor vascular contrast and lingering arterial signals, complicating manual segmentation. This study assesses the feasibility and effectiveness of using contrast-enhancement boost (CE-boost) and black-blood computed tomography (BBCT) techniques to improve image quality and reduce the manual segmentation time of the portal venous system in abdominal CTA.

Methods: In this retrospective study, 50 patients undergoing abdominal CTA were processed by three methods: conventional portal venous phase (Group A), CE-boost images (Group B), and combined CE-boost with BBCT images (Group C). Subjective image quality was scored by two experienced radiologists, whereas objective metrics [signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR)] were calculated from standardized region of interest (ROI) measurements. Manual segmentation times were recorded. Statistical analyses employed intraclass correlation coefficient (ICC) for interobserver agreement, Kruskal-Wallis test for subjective scores, and one-way analysis of variance (ANOVA) for objective metrics, with P<0.05 denoting significance.

Results: Group C demonstrated significantly higher subjective image quality scores (median 5.0 vs. 3.7 and 4.4; P<0.001) and superior objective metrics (SNR and CNR improvements, all P<0.001) compared to Groups A and B. Moreover, median segmentation time in Group C (2.2 min) was markedly shorter than it was in Group A (5.1 min) and Group B (4.2 min) (P<0.001).

Conclusions: The synergistic integration of CE-boost and BBCT markedly enhances portal vein visualization and substantially accelerates manual segmentation.