BMC Medical Imaging最新文献

Background: Epicardial adipose tissue (EAT) contributes to arrhythmogenic substrate development through electrophysiological and structural remodeling. Pathological EAT exhibits significant heterogeneity. We investigated whether EAT heterogeneity, quantified by entropy analysis on cardiac magnetic resonance (CMR), predicts the occurrence of ventricular and atrial arrhythmias post-myocardial infarction (MI).

Materials and methods: This cohort study enrolled 241 consecutive patients post-MI. CMR was performed to assess EAT volume, myocardial scar, biventricular function, and strain. EAT heterogeneity was quantified using an entropy algorithm. The primary endpoints were the occurrence of ventricular arrhythmias (VAs) or atrial tachyarrhythmias (AAs).

Results: Over a median follow-up of 31 months, 43 (17.8%) patients developed VAs and 30 (12.4%) developed AAs. EAT entropy was significantly higher in patients who developed either VAs or AAs. In multivariable Cox regression analysis, EAT entropy, LA reservoir strain (Es), and global longitudinal strain (GLS) were independent predictors of VAs. For AAs, EAT entropy, Es, and EAT thickness independently predicted the outcome. ROC analysis revealed that the model integrating these parameters have good efficacy for the prognosis evaluation of VAs (area under the curve [AUC] = 0.872) and AAs (AUC = 0.917). Myocardial fibrosis exhibited modest correlations with EAT entropy.

Conclusion: EAT heterogeneity, quantified by entropy, is an independent predictor of both VAs and AAs in post-MI patients. This novel imaging biomarker may enhance risk stratification and guide therapeutic strategies in this high-risk population.

Background: To investigate whether a machine learning (ML) model integrating CT-based radiomics and clinical features can noninvasively evaluate the aldosterone secretion and CYP11B2 status of the adrenal gland, using Shapley Additive Explanations (SHAP) for model interpretation.

Methods: Patients who underwent adrenalectomy with CYP11B2 immunohistochemistry (IHC) confirmation between January 2022 and February 2025 were analyzed retrospectively. Radiomics models were developed based on different radiomics features selected by ElasticNet. Clinical features were used to construct the clinical model. The radiomics score and clinical features were integrated into the combined model. The SHAP method was applied to calculate feature importance. An external validation cohort from a second center was additionally analyzed to assess generalizability.

Results: In total, 140 patients were enrolled in the study. The clinic-radiomics model exhibited superior performance, achieving average AUC values of 0.912, 0.923 and 0.958 in distinguishing classical histopathology, nonclassical histopathology, and non-functional adrenal adenoma (NFA), respectively. The decision curve analysis showed that the combined model performs best. The SHAP method ranked the five features according to their importance. In the external cohort (n = 50), the model achieved AUCs of 0.807, 0.735, and 0.722 for classical histopathology, nonclassical histopathology, and NFA, respectively.

Conclusion: A machine learning model integrating radiomics and clinical features enables noninvasive prediction of adrenal aldosterone secretion function and histopathological subtypes, with preliminary evidence of generalizability in an independent external cohort.

Background: HIV-associated neurocognitive disorders (HAND) are common complications in HIV-infected individuals, and working memory impairment is one of the core features. Although combination antiretroviral therapy (cART) has reduced the incidence of severe HAND, mild HAND remains prevalent. This study aims to explore the functional brain characteristics related to working memory in HIV-infected individuals with different cognitive states using resting-state and task-based functional magnetic resonance imaging (fMRI), and to identify candidate imaging markers for early diagnosis and inform future intervention targeting.

Methods: Fifty-nine HIV-infected individuals (30 with cognitive integrity [CI], 29 with asymptomatic neurocognitive impairment [ANI]) and 37 healthy controls (HC) were enrolled. Resting-state and task-based fMRI were acquired. Task-fMRI was performed using a spatial working memory task to analyze brain activation, functional connectivity (FC), and reconfiguration efficiency of FC from rest to task. FC networks were constructed as ROI-ROI Pearson correlation matrices (Fisher z-transformed) and significant group differences were identified using network-based statistics. Pearson and Spearman correlation analyses were used to explore the relationships between reconfiguration efficiency and clinical/cognitive variables.

Results: HC showed better task performance than both HIV groups, and ANI exhibited the poorest accuracy. Compared with CI, ANI had significantly lower neurocognitive domain T-scores in memory, attention/working memory, and abstraction/executive function. In task-fMRI analyses, ANI showed decreased activation in the bilateral orbital middle frontal gyri and the left middle temporal gyrus, alongside increased activation in the left cerebellum crus I relative to CI. Whole-brain analyses demonstrated widespread FC increases in both HIV groups at rest and during the task compared with HC. Reconfiguration efficiency differed across groups and showed stage-related associations with immune and cognitive measures.

Conclusions: Cognitive impairment in virally suppressed HIV is accompanied by altered working-memory network engagement, with greater cortico-cerebellar involvement in ANI. While static whole-brain FC showed widespread increases but limited CI-ANI separation under stringent correction, altered rest-to-task FC reconfiguration efficiency was associated with immune indices and neurocognitive/behavioral performance, suggesting that this cross-state metric may serve as a candidate marker for HAND phenotyping and risk stratification.

Background: Artificial intelligence (AI) has shown increasing potential in lung cancer imaging, particularly in detection, staging, prognosis, and recurrence prediction. However, there is limited synthesis of head-to-head comparative evidence between CT, FDG PET/CT, and multimodal fusion models within the same cohorts.

Objectives: To systematically review and critically appraise studies that directly compared CT-only, FDG PET/CT-only, and combined multimodal models in lung cancer, with emphasis on clinical setting, fusion strategy, validation design, and clinical utility.

Methods: This systematic review followed PRISMA 2020 and PRISMA-S guidelines. PubMed, Scopus, IEEE Xplore, and Google Scholar were searched for English-language human studies published between January 1, 2019, and September 8, 2025. Eligible studies reported same-cohort, head-to-head comparisons of CT, PET/CT, or multimodal models for lung cancer screening, staging, or prognosis. Risk of bias was assessed using PROBAST for prediction model studies and SANRA for narrative reviews. Data were extracted in duplicate and synthesized narratively, with meta-analysis performed where ≥ 3 studies were sufficiently homogeneous.

Results: From 2,417 records (PubMed 845, Scopus 920, IEEE Xplore 452, Google Scholar/manual 200), 31 studies met inclusion criteria (20 primary modeling studies, 11 reviews). In screening cohorts, low-dose CT deep-learning models consistently outperformed other modalities, with modest incremental value from clinical covariates. For nodal staging, integrated PET/CT radiomics-clinical models showed superior discrimination, calibration, and net-benefit compared with unimodal approaches. In prognostic and recurrence settings, fused PET/CT models outperformed CT- or PET-only models across institutions, with further improvement from clinical variables. Radiogenomics and pathology integration provided added value but were limited by small samples and lack of external validation.

Conclusions: Comparative evidence demonstrates that modality performance is context-dependent: CT dominates in screening, PET/CT fusion excels in staging and prognosis, and multimodal integration with clinical or biomarker data enhances discrimination and utility. Standardization, harmonization, and rigorous external validation remain critical for generalizability.

Clinical trial number: Not applicable.