BACKGROUND:Early diagnosis of hip osteoarthritis (OA) is hindered by the insensitivity of radiography to early biochemical changes. Ultrashort echo time (UTE) MRI allows for quantitative assessment of deep calcified cartilage, potentially identifying tissue degeneration before irreversible structural damage. PURPOSE:To quantitatively characterize spatial topological patterns of cartilage degeneration in early hip OA (Tönnis grade 1) using UTE-T2⁎ mapping and to evaluate the loss of joint biochemical homeostasis. STUDY TYPE:Prospective case-control study. POPULATION:71 hip joints from 36 participants, classified into a control group (Tönnis grade 0, n = 22) and an early OA group (Tönnis grade 1, n = 49). FIELD STRENGTH/SEQUENCE:3.0-T; multi-echo UTE sequence for T2⁎ mapping. ASSESSMENT:T2⁎ relaxation times were quantified in the acetabular superior, femoral head superior, and femoral head inferior regions by a musculoskeletal radiologist (X.X.X., 10 years of experience). Joint homeostasis was characterized by spatial heterogeneity (joint coefficient of variation [CV]) and topological balance (Acetabular Ratio: acetabular superior-to-femoral inferior relaxation time). STATISTICAL TESTS:Mann-Whitney U test, Chi-square test, Pearson correlation, multivariable logistic regression, and Variance Inflation Factor (VIF). P < .05 was considered statistically significant. RESULTS:The early OA group exhibited significantly elevated T2⁎ values across all regions compared with controls (P < .001). Regression analysis indicated that cartilage T2⁎ elevation in early OA represented a pathological trajectory distinct from physiological aging (P = .089 for T2⁎ after age adjustment). Crucially, the early OA group demonstrated a significant loss of joint homeostasis: (a) Spatial Heterogeneity: The joint CV was markedly increased in early OA (0.36 ± 0.23 vs 0.19 ± 0.10; P < .001), reflecting a "patchy" disruption of biochemical uniformity. (b) Topological Imbalance: The physiological acetabular-to-femoral "unity ratio" (baseline ∼1.0) showed a trend toward acetabular-dominant degeneration (1.50 ± 0.91 vs 1.09 ± 0.36; P = .094). DATA CONCLUSION:Early hip OA is characterized by a fundamental loss of biochemical homeostasis, manifested as acetabular-predominant degeneration and significantly increased spatial heterogeneity. Quantitative UTE-MRI provides a sensitive metric for monitoring this disruption of joint equilibrium beyond simple mean value elevation. LEVEL OF EVIDENCE: 2:Technical efficacy stage: 2.

Background:Overweight and obesity are significant risk factors for carotid atherosclerosis in patients with metabolic syndrome and type 2 diabetes mellitus, and carotid computed tomography angiography (CTA) plays a critical role in assessing vascular health. However, obese patients often require higher doses of radiation and contrast agents, which can pose risks. The deep learning image reconstruction with high setting (DLIR-H) algorithm offers the potential to enhance image quality while minimizing exposure. The objective of this study was to evaluate the effectiveness of the DLIR-H algorithm in improving CTA image quality under a triple-low scan protocol (low radiation dose, low contrast agent usage, and low injection rate) for overweight and obese patients [body mass index (BMI) >25 kg/m2], using dual-energy CTA (DE-CTA) and virtual monoenergetic images (VMIs) at 50 keV. Methods:A prospective study was conducted involving 62 patients who were randomly assigned to either the control or experimental group. The experimental group used the adaptive statistical iterative reconstruction-V (ASIR-V) 50%, deep learning image reconstruction with low setting (DLIR-L), deep learning image reconstruction with medium setting (DLIR-M), and DLIR-H algorithms with reduced radiation exposure and contrast agent. Both objective and subjective image quality evaluations were conducted. The effective dose (ED), contrast agent dose, computed tomography values (CTV), standard deviation of the carotid artery vessels (SDV), contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR) were calculated and compared at four anatomical regions: the aortic arch (AA), common carotid artery (CCA) origin, carotid bifurcation (CB), and internal carotid artery (ICA) origin. Results:The DLIR-H algorithm demonstrated image quality comparable to that of the ASIR-V algorithm. The experimental group exhibited a 49.4% reduction in ED (calculated from the dose length product, DLP) and a 13.5% reduction in contrast agent usage compared to the control group. At the AA level, the DLIR-H group had a significantly lower CTV than the control group [561.90 (516.90, 661.00) vs. 649.30 (572.60, 745.50), P<0.05]. At the CCA level, the DLIR-H group demonstrated a significantly lower SDV than the control group [35.90 (29.20, 43.80) vs. 41.70 (35.90, 54.70), P<0.05]. Except for the CCA level, at other anatomical levels, the DLIR-H group showed significantly lower SDV compared with the ASIR-V 50%, DLIR-L, and DLIR-M groups (P<0.05). Additionally, the DLIR-H group exhibited higher CNR and SNR than the ASIR-V 50%, DLIR-L, and DLIR-M groups at several anatomical levels (P<0.05). Conclusions:The DLIR-H algorithm significantly enhances image quality in CTA, reducing both radiation exposure and contrast agent usage in overweight and obese patients.

Timely diagnosis of mesenteric vascular diseases, especially acute mesenteric ischemia (AMI) due to embolism in the superior mesenteric artery (SMA), is crucial for effective intervention. Dual-energy computed tomography angiography (DE-CTA) is a key diagnostic tool; however, concerns about contrast-induced nephropathy and radiation exposure persist. This study assesses the diagnostic performance of DE-CTA for detecting mesenteric vascular diseases and compares the effectiveness of image reconstruction algorithms, specifically ASIR-V and DLIR, in enhancing image quality while minimizing the risks associated with contrast agents and radiation exposure. DE-CTA with virtual monoenergetic imaging (VMI) at 40 keV was performed on 50 patients. Image quality was evaluated using contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), and standard deviation (SD). The performance of DLIR and ASIR-V algorithms was compared, with an emphasis on minimizing radiation dose and contrast agent use through optimized low-dose protocols. DLIR-H significantly outperformed both ASIR-V 50

Dual-energy CT (DECT) at 50 keV increases iodine attenuation but exponentially amplifies image noise. The feasibility of using deep learning image reconstruction (DLIR) to counteract this noise under a “dual-low” (low-radiation and low-contrast medium) thyroid CT protocol remains underexplored. To investigate the performance of a dual-low DECT protocol combined with DLIR in contrast-enhanced thyroid CT compared with a standard-dose protocol using adaptive statistical iterative reconstruction-Veo (ASIR-V). In this prospective study (August–December 2025), patients were randomly assigned to a standard-dose group (120 kVp, 1.0 mL/kg iodine, ASIR-V 50

Current diagnostic methods, such as serological tests and polymerase chain reaction (PCR), are limited by slow turnaround time and sample quality, making early diagnosis challenging. Although mNGS offers higher sensitivity, its expense and specialized equipment limit clinical applicability. These limitations underscore the need for more reliable, rapid, and cost-effective diagnostic tools. Between June 2023 and March 2024, 191 consecutive pediatric patients were enrolled. The most recent laboratory test results prior to bronchoalveolar lavage (BAL) were included. After LASSO screening, seven machine learning classifiers (LR, SVM, KNN, RF, ET, XGBoost, LightGBM) were tested, and the optimal one was selected for model construction. The nomogram model combined the radiomics (rad) signature and the clinical signature. The ROC curves were drawn to evaluate the diagnostic efficacy of different models. The calibration efficiency of the nomogram was evaluated by drawing calibration curves, and the Hosmer-Lemeshow test was used to evaluate the calibration ability of the models. Decision curve analysis (DCA) was utilized to evaluate the clinical utility of the models. A p-value < 0.05 was considered statistically significant. A total of 1,834 handcrafted radiomics features were extracted, including 360 first-order features, 14 shape features, and texture features. The LR classifier achieved the best AUC, reaching 0.922 and 0.867 for distinguishing Co-MPP from MPP in the training and test cohorts, respectively. For building the clinical signature, LR was selected as the base model. The univariate analysis results of all clinical laboratory and CT imaging features showed that only reticulation and bronchial lumen occlusion were significantly different between MP and Co-MPP patients (p = 0.011 and p < 0.001, respectively). The clinical signature achieved AUC values of 0.729 and 0.706 in the training and test cohorts, respectively. A nomogram based on the LR algorithm was constructed to combine the clinical signature and the Rad signature. The DeLong test showed that the performance of the nomogram and the Rad signature was significantly higher than that of the clinical signature (p < 0.05), whereas no significant difference was observed between the nomogram and the Rad signature. Both the Rad signature and the nomogram demonstrated substantial clinical benefit. Our findings indicate that machine learning can assist clinicians in distinguishing Co-MPP from MPP in children. Furthermore, both the Rad signature and nomogram model showed higher clinical benefit than the clinical signature.

PURPOSE:To evaluate the accuracy, radiation dose, and clinical feasibility of a high-pitch low-dose chest CT (Sa36LDCT) scanning protocol with calcium-aware reconstruction for coronary artery calcium scoring (CAC). MATERIALS AND METHODS:In this prospective study, 90 patients underwent both standard CAC scanning (Qr36CACS, 120 kVp) and Sa36LDCT (70 to 120 kVp, high-pitch, ATVS). CAC scoring was quantified as Agatston score, calcium volume, and equivalent mass. Risk stratification was assessed using Agatston categories. Agreement between protocols was evaluated using intraclass correlation coefficients, Bland-Altman analysis, and weighted kappa. Subgroup analyses were performed according to heart rate (≤75 vs. >75 bpm). Radiation dose and subjective image quality were also compared. RESULTS:CAC metrics showed no significant difference between protocols, with excellent agreement (ICC = 0.983 to 0.996). Subgroup analysis revealed that heart rate did not significantly influence CAC quantification or risk classification, with Agatston score ICC values of 0.98 for ≤75 bpm and 0.99 for >75 bpm, demonstrating similar accuracy in both subgroups. Radiation dose reduction was consistently observed across both subgroups, with Sa36LDCT reducing effective radiation dose by 62% (0.74 vs. 0.28 mSv, P < 0.001), without compromising image quality. Risk stratification based on Agatston categories showed near-perfect consistency across heart rate subgroups (weighted k = 0.957). CONCLUSION:High-pitch Sa36LDCT provides accurate and reliable CAC scoring equivalent to standard CAC scoring, with substantial radiation dose reduction and preserved image quality. The protocol's robustness across heart rate subgroups demonstrates its potential as a safe and practical approach for combined pulmonary and cardiovascular screening, pending validation in larger, more diverse populations.

Background Deep learning image reconstruction (DLIR) has gained recognition as a promising technique to improve image quality in low-dose CT imaging. However, its performance in dual-energy CT portal venography (DE-CTPV), particularly under reduced contrast medium volume and radiation dose (dual-low dose) conditions, remains underexplored. Objective This study aims to compare the performance of DLIR and adaptive statistical iterative reconstruction (ASIR-V) in DE-CTPV, with a focus on image quality across multiple vascular segments of the portal venous (PV) system under dual-low dose protocols. Methods Patients undergoing DE-CTPV were reconstructed using DLIR medium (DLIR-M) and high strength (DLIR-H) and ASIR-V (50%). Image quality was assessed both subjectively and objectively in the main portal vein (MPV), left and right portal veins (LPV, RPV), splenic vein (SV), and superior mesenteric vein (SMV). Objective metrics, including image noise, contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR), were calculated. Additionally, radiation dose parameters (CTDIvol, DLP, ED) and contrast medium volume were compared with data from previous studies. Results In this study, the mean CTDIvol, DLP, and ED were 9.79 ± 2.13 mGy, 326.26 ± 84.58 mGy·cm, and 4.89 ± 1.27 mSv, respectively. The mean contrast medium volume was 79.5 ± 11.4 mL. DLIR-H significantly enhanced image quality across all vascular segments, achieving substantial reductions in image noise and notable increases in CNR and SNR (P < 0.05). It also received the highest subjective ratings for overall image quality, image noise, vascular edge sharpness, and diagnostic confidence compared to ASIR-V 50%. The use of 55 keV virtual monoenergetic imaging (VMI) further enhanced iodine contrast effectiveness, while DLIR effectively reduced noise, ensuring clearer and more consistent vascular delineation across all assessed vascular segments. Conclusion DLIR substantially improves image quality in DE-CTPV compared with ASIR-V 50%, even when utilizing dual-low dose protocol. By providing consistent, high-quality imaging across multiple portal venous segments, DLIR may offers a safer and more reliable approach for preoperative evaluation and postoperative monitoring in liver transplantation.

Virtual monoenergetic imaging (VMI) at 40 keV improves iodine attenuation in colon cancer CT but is constrained by severe image noise. Deep learning image reconstruction (DLIR) may address this limitation, but its effect on anatomical edge preservation across multiple targets requires investigation. To evaluate the impact of DLIR on objective and subjective image quality of 40-keV VMIs in colon adenocarcinoma, with emphasis on the trade-off between noise reduction and edge definition. In this completed retrospective study (patient enrollment window: May 2024 to February 2025), 60 patients (mean age, 62.8 years ± 15.1; 34 men) with confirmed colon adenocarcinoma underwent dual-energy CT using a low-iodine protocol (1.0 mL/kg). Portal venous phase data were reconstructed at 40 keV using adaptive statistical iterative reconstruction-V (ASIR-V) 50

Emerging evidence indicates that stroke may impair glymphatic function by disrupting cerebrospinal fluid clearance. The coupling between global blood-oxygen-level-dependent signals and cerebrospinal fluid flow has been proposed as a non-invasive biomarker of glymphatic activity. However, it remains unclear whether this coupling can be modulated through rehabilitation. This study investigated whether language therapy in patients with post-stroke aphasia enhances global blood-oxygen-level-dependent- cerebrospinal fluid coupling, thereby reflecting potential restoration of fluid-brain interaction. This longitudinal observational study was conducted at a single centre and included 20 patients with post-stroke aphasia and 35 age- and sex-balanced healthy controls. All participants underwent MRI scanning, including resting-state blood-oxygen-level-dependent and structural imaging. Among the post-stroke aphasia group, 14 patients completed both pre- and post-treatment assessments after undergoing a standardized 4-week speech-language therapy programme. The Western Aphasia Battery was used to quantify language deficits and monitor treatment-related changes. Global blood-oxygen-level-dependent-cerebrospinal fluid coupling was quantified using cross-correlation analysis across the whole brain and four predefined language-related resting-state networks. At baseline, patients exhibited significantly reduced global blood-oxygen-level-dependent-cerebrospinal fluid coupling compared to healthy controls (P < 0.05). Following therapy, coupling significantly increased within the language, salience and dorsal attention networks, whereas coupling within the default mode network significantly decreased (all P < 0.05). Notably, increased global blood-oxygen-level-dependent-cerebrospinal fluid coupling in specific networks was significantly correlated with improvements in targeted language functions, such as object naming, responsive naming and auditory word recognition (P < 0.05). These findings suggest that language rehabilitation enhances neurophysiological coupling between brain activity and cerebrospinal fluid flow, potentially reflecting restoration of fluid-brain interaction in post-stroke aphasia.

Background Dual-energy CT pulmonary angiography (DE-CTPA) facilitates dose reduction through virtual monoenergetic imaging (VMI). However, the optimal reconstruction strategy for balancing image noise and vascular contrast in low-dose protocols—particularly for evaluating subsegmental arteries—remains to be fully established. Purpose To determine the optimal VMI energy level and reconstruction algorithm for "double-low" (low radiation and contrast media) DE-CTPA and to compare the image quality of deep learning image reconstruction (DLIR) versus adaptive statistical iterative reconstruction-V (ASIR-V) in patients with pulmonary embolism (PE). Materials and Methods This study, conducted between January and August 2025, included 39 patients with suspected PE who underwent DE-CTPA using an optimized low-dose protocol. Images were reconstructed at VMI levels from 40 to 140 keV using ASIR-V (30%–70%) and DLIR (medium [DLIR-M] and high [DLIR-H] strength). Objective metrics (CT attenuation, image noise, signal-to-noise ratio [SNR], and contrast-to-noise ratio [CNR]) were evaluated in the main, lobar, segmental, and subsegmental pulmonary arteries. Subjective image quality was assessed by two radiologists using a 5-point Likert scale. Results Among all energy levels, 40 keV VMI yielded the highest vascular attenuation. At this energy level, DLIR-H achieved significantly lower image noise compared with ASIR-V and DLIR-M ( P < 0.05). The 40 keV DLIR-H protocol demonstrated the highest CNR across all vascular territories, significantly outperforming the standard 70 keV ASIR-V 50% protocol in the pulmonary trunk ( P = 0.013), lobar ( P < 0.001), segmental ( P = 0.010), and subsegmental arteries ( P < 0.001). Subjectively, 40 keV DLIR-H images were rated as excellent (median score, 5.0) and were noninferior to the standard 70 keV protocol for evaluating segmental and subsegmental arteries ( P > 0.05). Conclusion A low-dose DE-CTPA protocol combining 40 keV VMI with high-strength DLIR provides superior objective image quality and excellent subjective depiction of subsegmental arteries while substantially reducing radiation and contrast media exposure.

Spontaneous isolated superior mesenteric artery dissection (SISMAD) is traditionally diagnosed via CT angiography (CTA) structural features, but localized biological inflammation remains poorly characterized. We aimed to characterize the radial spatial gradient of perivascular adipose tissue (PVAT) in SISMAD and evaluate a composite Vascular Inflammation Index (VII) as an auxiliary indicator of local inflammatory activity. In this retrospective study (2016–2025), 55 patients with SISMAD and 110 symptomatic controls were evaluated. PVAT volume, mean attenuation, and standard deviation (SD) were quantified across cumulative radial shells (0–1 to 0–5 mm) from the SMA wall. A composite VII, integrating volume burden, expansion kinetics, and tissue texture, was developed. Multivariable logistic regression and bootstrap analysis (1000 iterations) assessed independent predictors and incremental diagnostic value. Participants (mean age, 50.4 years ± 15.9) included SISMAD patients who had higher white blood cell counts than controls (10.7 ± 5.2 vs. 8.7 ± 4.0 × 109/L; P = 0.017). PVAT volume was significantly higher in SISMAD (0–5 mm: 10.1 mm3 ± 5.1 vs. 6.8 mm3 ± 4.0; P < 0.001). SISMAD patients exhibited aggressive radial expansion (expansion ratio: 12.5 vs. 8.6; P = 0.032) and lower 3-mm heterogeneity (22.8 vs. 24.1 HU; P = 0.046), suggesting an edema-induced homogenization effect. The VII achieved an AUC of 0.800 (95

Background:Coronary magnetic resonance angiography (CMRA) is limited by respiratory motion artifacts, for which diaphragmatic navigation (dNAV) is commonly applied. Myocardial navigation offers a direct motion-tracking alternative, but its performance in different anatomical placements remains underexplored. This study aimed to compare the image quality and clinical feasibility of diaphragmatic and myocardial navigation approaches at 3.0 T. Methods:Thirty-three healthy volunteers underwent CMRA with four navigators: dNAV, left ventricle navigation (LvNAV), right atrial navigation (RaNAV), and apex navigation (ApNAV). Acquisition efficiency, scan duration, success rate, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were analyzed for major coronary segments: right coronary artery (RCA), left anterior descending artery (LAD), and left circumflex artery (LCX). Two observers independently scored image clarity and sharpness (4-point scale) for segments. Interobserver agreement was assessed via intra-class correlation coefficient (ICC). Results:LvNAV and RaNAV both achieved a 100% technical completion rate (30/30). This was higher than the technical completion rates for ApNAV (97%, 29/30) and dNAV (90%, 27/30), with the latter two groups experiencing failures due to navigator drift. Scan durations showed no significant differences (dNAV: 478.70±104.68 s; LvNAV: 469.62±102.64 s; P=0.332). LvNAV yielded the highest SNR and CNR (P<0.05). Proximal and mid segments of the RCA, LAD, and LCX showed comparable visualization between dNAV and LvNAV (P>0.05), whereas distal segment visibility was reduced mainly with RaNAV and ApNAV, with dNAV and LvNAV preserving better distal visualization. Coronary visualization rates differed significantly (χ2=76.563, P<0.001), with dNAV and LvNAV outperforming other techniques. Conclusions:Myocardial navigation, particularly LvNAV, yields superior image quality and visualization performance comparable to dNAV, representing a promising alternative under conditions requiring enhanced motion stability. Personalized navigator selection based on respiratory patterns may further optimize CMRA outcomes.

Carotid CT angiography (CTA) is valuable for diagnosing carotid artery disease but involves radiation and contrast agent risks. Deep Learning Image Reconstruction (DLIR-H) shows potential for maintaining image quality in low-dose protocols. In this prospective study, 180 patients undergoing dual-energy CTA were divided into three groups: a control group (ASIR-V 50

ObjectiveTo develop and validate an integrated nomogram combining MRI radiomics features with clinical variables for predicting 30-day progression risk in patients with acute ischemic stroke (AIS), and to deploy a web-based visualization tool for clinical application.MethodsThis retrospective two-center study included 254 AIS patients. Radiomics features were extracted from DWI using MaZda, and key features were selected by LASSO logistic regression to build a radiomics signature. Clinical predictors were identified using logistic regression. Radiomics, clinical, and combined models were developed and evaluated using area under the curve (AUC), net reclassification index (NRI), and integrated discrimination improvement (IDI), calibration analysis, and decision curve analysis (DCA). Optimal cutoffs were determined in the training cohort using the Youden index and then directly applied to the external validation cohort without re-optimization. The corresponding training-derived thresholds were −1.2023 for the clinical model, −1.0368 for the radiomics model, and −2.3639 for the combined model. Bootstrap resampling (1,000 repetitions) was used for internal validation of the combined model.ResultsEight optimal radiomics features were selected from 300 extracted features. Multivariate analysis identified four independent predictors: Radiomics score (Radscore) (OR = 2.23, 95% CI: 1.28–5.11), National Institutes of Health Stroke Scale (NIHSS) score (OR = 1.57, 95% CI: 1.25–2.14), type 2 diabetes mellitus (OR = 9.28, 95% CI: 2.02–56.03), and monocytes (MO) (OR = 1.08, 95% CI: 1.02–1.16). The combined model showed favorable discriminative performance with AUCs of 0.945 (95%CI: 0.903–0.987) and 0.904 (95%CI: 0.854–0.954) in the training and validation cohorts, respectively. When the training-derived cutoff of −2.3639 was applied to the validation cohort, the combined model achieved 93.75% sensitivity, 73.95% specificity, and 83.85% balanced accuracy. Bootstrap internal validation yielded an optimism-corrected C-index of 0.912 and a corrected calibration slope of 0.769. The combined model outperformed the clinical-only and radiomics-only models, although the external validation should still be interpreted as an initial cross-center assessment.ConclusionThe integrated radiomics-clinical nomogram incorporating Radscore, NIHSS score, type 2 diabetes mellitus (T2DM), and MO showed potential for predicting 30-day progression risk in patients with AIS. This web-based tool may support early risk stratification, but further large-scale prospective validation is required before routine clinical application.

The integration of tumor biomarker sensing with real-time therapy monitoring within a single nanoplatform remains a significant challenge for precision medicine. Here, we report a programmed magnetic resonance tuning probe for concurrent imaging of tumor-associated microRNA-21 (miRNA-21) and drug release monitoring in vivo. The probe is engineered with a Gd-labeled, drug-loaded substrate strand (sDNA) as a signal enhancer, superparamagnetic Fe3O4 as a quencher, and an S-cDNA-inhibited DNAzyme walker (WDNAzyme) as the miRNA-responsive activator. In the presence of overexpressed miRNA-21 in the tumor tissues, hybridization of S-cDNA with the substrate-recognition region of WDNAzyme suppresses its catalytic activity. Recognition of miRNA-21 induces competitive strand displacement, thereby activating the WDNAzyme. This programmed cascade simultaneously releases the loaded drug and the Gd contrast agent, switching the magnetic resonance (MR) signal from an “off” to an “on” state. This design yields an activatable, miRNA-21-responsive probe for tumor-specific imaging. Furthermore, the recovery of the T1 signal exhibits a linear correlation with the amount of drug released, allowing for quantitative pharmacokinetic monitoring. The probe's dual functionality for miRNA-21-responsive imaging and quantitative drug release tracking is validated in both cellular systems and murine tumor models. This work presents a smart theranostic strategy that merges biomarker detection with non-invasive therapy feedback, offering a promising avenue for image-guided personalized treatment.

Carotid artery plaques are a major contributor to ischemic stroke, yet current assessment methods focusing on luminal stenosis often fail to fully capture plaque vulnerability. High-resolution magnetic resonance vessel-wall imaging (HR–MR-VWI) serves as the reference standard for plaque vulnerability assessment but is limited by cost and availability. Computed tomography angiography (CTA) offers a more accessible, non-invasive alternative for plaque evaluation. This study aimed to develop and validate a plaque-level predictive model for carotid plaque vulnerability using machine learning and CTA features, with HR–MR-VWI serving as the reference standard. The goal was to identify key imaging features that contribute to plaque vulnerability and assess the performance of machine learning models in predicting plaque instability. A retrospective analysis was conducted on patients who underwent both carotid CTA and HR–MR-VWI within one month. Plaque features were extracted from CTA, including plaque composition, vascular lumen geometry, and perivascular adipose tissue (PVAT). Data were randomly divided into training (70

To evaluate carotid computed tomography angiography (CTA) performance in assessing MRI-defined plaque vulnerability using high-resolution magnetic resonance vessel wall imaging (HR-MR-VWI) as the reference, and to develop a multidimensional predictive model integrating plaque composition, geometry, and perivascular adipose tissue (PVAT). Patients undergoing both CTA and HR-MR-VWI were retrospectively included. Plaques were classified by modified AHA criteria and co-registered to ensure spatial correspondence. Quantitative CTA features were extracted via semi-automated segmentation. A historical cohort (2018–2024) was partitioned into training (n = 130) and internal validation (n = 57) sets; a recent cohort (2024–2025) served as a prospective temporal test set (n = 78). A multivariate logistic regression model was developed and evaluated using the area under the receiver operating characteristic curve (AUC) and decision curve analysis. In 265 plaques, PVAT attenuation (OR = 1.05; p < 0.001) and maximum diameter stenosis (MDS) (OR = 1.03; p < 0.05) emerged as independent predictors of MRI-defined vulnerability. The combined model achieved robust discrimination with AUCs of 0.86, 0.80, and 0.85 in the training, internal validation, and prospective temporal test sets, respectively. Calibration and decision curve analysis demonstrated excellent agreement and clinical net benefit across all cohorts. CTA-derived MDS and PVAT attenuation are robust independent predictors of MRI-defined carotid plaque vulnerability. This supportive, proof-of-concept nomogram offers a tool for characterizing high-risk plaque phenotypes, highlighting CTA as a viable supplementary tool to MRI in routine practice. Question Can a multidimensional carotid CTA model, integrating perivascular adipose tissue and luminal geometry, accurately identify high-risk plaque phenotypes compared to high-resolution MRI? Findings CTA-derived perivascular fat attenuation and stenosis severity are independent predictors of MRI-defined plaque vulnerability, achieving robust diagnostic performance across internal and temporal validation. Clinical relevance This plaque-level CTA model provides a rapid, accessible tool for identifying MRI-defined vulnerable carotid plaques and personalized management in routine clinical practice where MRI access is limited.

BACKGROUND:Accurate preoperative evaluation of rectal cancer is essential for staging and treatment planning. Low-energy virtual monoenergetic imaging (VMI) enhances iodine contrast in dual-energy computed tomography (DECT) but increases image noise. Deep learning image reconstruction (DLIR) may mitigate this issue, but its effectiveness for 40 keV VMI in rectal cancer is underexplored. OBJECTIVE:To evaluate the impact of DLIR on 40 keV VMI image quality and its diagnostic performance in assessing extramural venous invasion (EMVI) and T staging, compared to adaptive statistical iterative reconstruction (ASIR-V). METHODS:Sixty-two patients with rectal adenocarcinoma underwent preoperative DECT using a low-iodine contrast protocol (1 mL/kg, 300 mg iodine/mL). Images were reconstructed at 70 keV ASIR-V 40 %, 40 keV ASIR-V 40 %, and 40 keV DLIR (medium [DLIR-M] and high [DLIR-H] settings). Objective and subjective image quality were compared using repeated-measures ANOVA or Friedman tests. Pathological findings were used as the reference standard for EMVI and T staging. RESULTS:Both 40 keV ASIR-V 40 %, DLIR-M, and DLIR-H significantly improved image quality compared to 70 keV ASIR-V, with improvements in CT attenuation, image noise, contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), edge rise slope (ERS), and the area under the noise power spectrum (NPS) curve (all P < 0.001). DLIR-M and DLIR-H outperformed 40 keV ASIR-V in terms of image noise and CNR. Subjective image quality scores were highest with DLIR-H. In diagnostic performance, DLIR-H achieved slightly better results for EMVI (AUC = 0.882) and T staging (AUC = 0.592) compared to ASIR-V. CONCLUSION:DLIR, particularly DLIR-H, significantly improves 40 keV VMI image quality but offers mild improvement in diagnostic performance for EMVI and T staging. The combination of low-keV VMI and DLIR provides high-quality imaging with reduced iodine doses, making it a promising approach for optimized DECT protocols in rectal cancer.

Perineural invasion (PNI) is a key prognostic determinant in gastric adenocarcinoma but remains radiologically occult on conventional CT. To develop and validate an integrated model combining dual-energy CT (DECT)–derived functional spectral slopes and subregional habitat signatures for the preoperative prediction of PNI. In this retrospective study, 175 patients (median age, 66 years; interquartile range, 58–73 years; 131 men) with gastric adenocarcinoma who underwent preoperative DECT between 2023 and 2025 were evaluated. Patients were randomized into a training cohort (n = 122) and an internal test cohort (n = 53). Tumor subregions were parcellated into three biologically distinct habitats using voxel-wise clustering. Stable habitat features (ICC ≥ 0.75), three-phase spectral slopes, and a Habitat-Graph Neural Network (H-GNN)–derived Topological Interaction Score (TIS)—capturing spatial adjacency relationships between habitat subregions—were collectively submitted as candidate predictors to LASSO regression for construction of the final Spectral-Habitat Signature. Diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC). Clinical utility was assessed via decision curve analysis. PNI was pathologically confirmed in 46.9

Carotid artery plaques, especially those with intraplaque hemorrhage (IPH), are significant contributors to ischemic stroke. Although high-resolution magnetic resonance vessel-wall imaging (HR-MR-VWI) is the gold standard for assessing plaque vulnerability, its limited availability and high cost pose challenges. Computed tomography angiography (CTA) offers a more accessible, non-invasive alternative, but its ability to detect IPH and assess plaque instability remains underexplored. Machine learning techniques have shown promise in improving the prediction of carotid plaque vulnerability using CTA. This study aimed to develop and validate a machine learning model using CTA to predict IPH in carotid plaques. The model integrated key imaging features, including plaque composition, vascular lumen geometry, and perivascular adipose tissue (PVAT), with HR-MR-VWI serving as the reference standard. The goal was to evaluate the model’s potential for non-invasive plaque vulnerability prediction, particularly in clinical settings where MRI is not available. A retrospective analysis was conducted on patients who underwent both carotid CTA and HR-MR-VWI within one month. Key plaque features, including composition, vascular lumen geometry, and PVAT, were extracted from CTA. The dataset was split into training (70