BACKGROUND:Frailty is a risk factor for adverse outcomes in patients undergoing mitral valve transcatheter edge-to-edge repair (M-TEER) and is interrelated with sarcopenia. This study aimed to investigate the prognostic value of fatty muscle fraction (FMF), measured from routine preinterventional computed tomography, as an imaging biomarker for frailty in patients undergoing M-TEER. METHODS:In patients undergoing M-TEER between 2010 and 2022 at the Heart Center Bonn who received preprocedural contrast-enhanced computed tomography, skeletal muscle quantity and quality at the L3 vertebral level were assessed on arterial-phase images using a deep learning-based segmentation algorithm. Fatty muscle fraction (FMF) was quantified and correlated with anthropometric, clinical, echocardiographic, and laboratory parameters. The primary end point was all-cause mortality at 1 year following the procedure, with additional follow-up at 30 days, 2 years, and 3 years. RESULTS:A total of 197 patients (median age, 79.3 years [interquartile range, 75.0-84.0]; 54.3% men; median EuroScore II, 4.1 [2.6-5.8]) were investigated. One-year survivors had significantly lower FMF values than nonsurvivors (49.0% [39.0-58.7] versus 57.6% [49.6-69.2]; P<0.001). Patients were divided into tertiles according to the extent of FMF: low FMF (<44.42%), medium FMF (44.42-56.69%), and high FMF (>56.69%). Following M-TEER, high FMF was associated with a higher all-cause mortality at 30 days (16.7% versus 6.1% versus 1.5%; P=0.005), 1 year (30.3% versus 19.7% versus 7.7%; P=0.004), 2 years (36.4% versus 22.7% versus 12.3%; P=0.005) and 3 years (37.9% versus 22.7% versus 13.8%; P=0.005) follow-up. Multivariable Cox regression analysis for 1-year mortality revealed age (hazard ratio, 0.96 [95% CI, 0.93-1.00]; P=0.023), male gender (hazard ratio, 3.65 [95% CI, 1.67-7.975]; P<0.001), and FMF (hazard ratio, 1.05 [95% CI, 1.02-1.08]; P<0.001) as independent predictors of 1-year all-cause mortality. CONCLUSIONS:Computed tomography-derived FMF, as a potentially new frailty marker, is independently associated with all-cause mortality in patients undergoing M-TEER and may serve as a promising parameter for risk stratification.

BACKGROUND:To assess cardiac involvement in patients with newly diagnosed giant cell arteritis (GCA) using cardiovascular magnetic resonance (CMR). METHODS:In this prospective single-center study, patients with newly diagnosed GCA underwent CMR at baseline and under therapy 6 months later. The imaging protocol enabled evaluation of cardiac function and volumes, edema, late gadolinium enhancement (LGE), and T1 and T2 mapping, including extracellular volume fraction (ECV). Healthy controls were included for comparison. Group comparisons were performed using t-tests, Mann-Whitney U, and chi-square tests. Paired t-tests assessed longitudinal changes. RESULTS:A total of 45 GCA patients (mean age 73±9 years; 42.2% female) and 30 healthy controls were included. Active inflammatory cardiac disease was found in 3/45 (6.7%) of patients, comprising active pericarditis in 2/3 (66.7%) and active myocarditis in 1/3 (33.3%). LGE was observed in 12/45 (26.7%) of patients (5/12 [41.7%] with ischemic, 5/12 [41.7%] with non-ischemic, and 2/12 [16.7%] with pericardial pattern). Compared to controls, GCA patients had significantly elevated native T1 relaxation times (987±29 vs. 968±22 ms; p=0.003) and ECV values (27.3±3.8% vs. 25±2.1%; p=0.003). Patients with large-vessel involvement (24/45 [53.3%]) showed higher myocardial ECV values (28.7±4.2% vs. 25.8±2.5%, p=0.008). Follow-up CMR was performed in 35 patients; here, no significant changes were observed compared to baseline CMR (e.g., native T1: 984±25 vs. 980±20 ms, p=0.397). CONCLUSION:In patients with newly diagnosed GCA, CMR reveals subclinical cardiac involvement, including active inflammatory disease (myocarditis and pericarditis), post-ischemic scars, and signs of non-ischemic myocardial fibrosis.

Swarm Learning has been introduced as a decentralized alternative to Central Learning. We aimed to evaluate whether Swarm Learning (SL) is equivalent to Central Learning (CL) for training AI models to detect middle cerebral artery (MCA) occlusions on CT angiography (CTA). This retrospective study included CTAs of patients suspected of ischemic stroke and negative controls from two institutions and three scanners between October 2010 and March 2024. After manual segmentation of occluded vessels by two raters, two AI algorithms (MedNeXt‑S, Swin UNETR) were trained using CL and SL. Sensitivity, specificity, positive/negative predictive values (PPV, NPV), accuracy, and AUC were calculated. Significance testing was performed using the exact McNemar test and comparisons of relative predictive values. For external validation, CTAs from a fourth scanner from a third institution were used. 470 CTAs of 470 patients (mean age, 72.6 years ±14.3; 247 women) were included, with 356 (75.7

AIMS:Foetal circulation undergoes complex changes in congenital heart disease (CHD) that are challenging to assess with foetal echocardiography. This study aimed to assess clinical feasibility and diagnostic value of 4D flow cardiovascular magnetic resonance (CMR) in foetal CHD. METHODS AND RESULTS:Pregnant women in advanced third trimester pregnancy with foetal CHD were prospectively recruited for foetal CMR between August 2021 and November 2024. CMR protocol included cine and 4D time-resolved PC acquisitions. 4D flow data were evaluated for additional diagnostic information beyond cine imaging. Postnatal examination reports were used to confirm CMR findings. Net flow was quantified in the ascending aorta, distal aortic arch, main pulmonary artery, ductus arteriosus, and descending aorta and normalized to combined ventricular output (CVO). Student's t-test and analysis of variance with Tukey post hoc test were applied. Among 77 participants, 4D flow was of diagnostic quality in 62 (81%) foetuses [11/62 (18%) without CHD, 51/62 (82%) with CHD]. Comparison of net flow distributions revealed significant differences between foetuses without CHD and those with hypoplastic left heart complex (HLHC), dextro-transposition of the great arteries, and coarctation of the aorta. CVO was higher in normal foetal circulation compared with CHD (6.4 ± 1.5 mL vs. 5.1 ± 1.9 mL, P = 0.049). 4D flow provided diagnostic findings beyond cine imaging in 26/51 (51%) foetuses with CHD (e.g. retrograde aortic perfusion in HLHC) that were confirmed postnatally. CONCLUSION:Foetal 4D flow CMR characterizes flow patterns and provides relevant diagnostic information beyond cine imaging for prenatal CHD assessment.

Coronary CT angiography (CCTA) is recommended as a first-line test for evaluating suspected coronary artery disease in many patient cohorts. Photon-counting detector CT (PCD-CT) introduces direct energy discrimination, higher geometric dose efficiency, and ultrahigh spatial resolution compared to energy-integrating detector CT, enabling improved coronary lumen visualization and intrinsic spectral imaging. This multi-institutional guidance summarizes technical principles of PCD-CT relevant to cardiac imaging and synthesizes contemporary evidence across key applications: (1) CCTA protocol selection and reconstruction (including ultrahigh-resolution modes); (2) stent and heavily calcified plaque assessment; (3) coronary physiology using CT-derived fractional flow reserve; (4) structural heart evaluation including pre-TAVR planning; and (5) myocardial tissue characterization with late enhancement and extracellular volume mapping. We provide pragmatic protocol pathways and contrast-dose strategies, with compact tables that standardize scan modes, reconstruction kernels, iterative reconstruction levels, and spectral outputs.

Abstract Objective To evaluate the impact of constraining proprietary and open large language models (LLMs) to structured outputs in processing radiology request forms (RRF). Materials and methods We evaluated five LLMs—two proprietary (GPT-5-Thinking, Gemini 2.5 Pro) and three open (Qwen3-235B-A22B-Thinking, gpt-oss-120b, medgemma-27b-it)—on 100 RRFs (50 computed tomography, 50 magnetic resonance imaging). Each model processed cases with and without constraints to structured outputs. Endpoints included accuracy for modality, anatomical region, contrast phase, urgency, “all correct” (all four categories correct), and “indication improved” (clarity of rewritten text). Outputs were evaluated against a reference standard defined by board-certified radiologists and compared with two radiology residents (first-year and third-year). Accuracies with 95% confidence intervals were calculated. Results Constraining to structured outputs had model-dependent effects: it improved Gemini 2.5 Pro (all correct: from 53.0% [43.3–62.5] to 66.0% [56.3–74.5]) but reduced GPT-5-Thinking accuracy (from 76.0% [66.8–83.3] to 53.0% [43.3–62.5]), with minimal influence on open models. Both proprietary LLMs outperformed the best open models (up to 41.0% [31.9–50.8]). All LLMs exceeded the unassisted first-year residents’ performance (19.0% [12.5–27.8]). LLM assistance improved first-year residents’ accuracy to 65.0% [55.3–73.6], approaching the third-year residents’ performance (80.0% [71.1–86.7]), who performed comparably to the best LLMs. Across models, performance was highest for modality and anatomical region, and lowest for urgency. Indication reformulation was judged clearer in > 90% of cases across all models without hallucinations. Conclusion Constraining to structured outputs exerted model-specific effects. Proprietary LLMs achieved the highest accuracy in RRF-based protocol selection and improved first-year resident performance to an experienced-resident level. Relevance statement LLMs may serve as valuable decision-support tools for radiology workflow. Constraining LLMs to structured outputs produced divergent, model-specific effects in radiology protocol selection—improving Gemini 2.5 Pro, reducing GPT-5-Thinking, and minimally affecting open models—highlighting the need for model-specific prompting strategies before adopting LLMs in radiology decision support. Key Points Structured output constraints affect LLM performance differently. Gemini 2.5 Pro benefits from structured prompting, while GPT-5-Thinking declines. Open-weight models show minimal impact from output constraining. Proprietary models outperform open models in radiology protocol selection. Graphical Abstract

Continuous efforts are made to reduce contrast media, improving patient safety, reducing environmental risks, and addressing recurring supply shortages. The aim of this study was to evaluate contrast agent-reduced CT protocols for transcatheter aortic valve reconstruction (TAVR) planning in photon counting detector CT (PCDCT). 162 BMI-matched examinations with standard dose contrast media (SCD; 80 mL; Iohexol 300 mg/mL; 81 examinations) and reduced contrast media dose (RCD; 50 mL; 81 examinations) for TAVR planning on a PCDCT were included in this retrospective monocentric study. Virtual monoenergetic reconstructions (VMI) at 70 keV, 60 keV and 50 keV of contrast agent-reduced examinations were compared with polyenergetic images. Quantitatively, regions-of-interest (ROIs) were placed in the abdominal aorta, iliac bifurcation, femoral artery, left ventricle and trapezius muscles. Signal-to-noise-ratio (SNR) and contrast-to-noise-ratio (CNR) were calculated. Qualitatively, diagnostic quality and contrast were assessed on a visual grading scale of 1 (non-diagnostic) – 5 (excellent) and contrast agent dose was estimated. Averaged, SNR and CNR decreased by 8.71

Assessment of the clinical feasibility of robot-assisted endovascular visceral interventions to reduce physical strain caused by prolonged standing and enabling remote interventions. Between 05/2024 and 09/2024, 45 patients were included in this prospective, single-center study. Patients scheduled for elective endovascular abdominal and pelvic interventions with superselective catheterization were assigned to manual (27 patients) or robotic-assisted treatment (18 patients). Radiation dose of the interventionalist, examination time (including preparation and follow-up), procedure duration and fluoroscopy time were compared between procedures using the CorPath GRX platform (Corindus, Waltham, MA) and conventional procedures. Technical success of robotic interventions was defined as achieving stable microcatheter positioning at the predefined target treatment point in the target vessel under robotic navigation, allowing execution of the planned therapy without conversion to manual navigation. 18 patients underwent robotic-assisted interventions (mean age 68 ± 12 years; 15 male), transarterial chemoembolization (TACE) (n = 9), 99mTc-MAA simulation (MAA)/transarterial radioembolization (TARE) (n = 2) and prostatic artery embolization (PAE) (n = 7). 27 comparable procedures were performed manually (mean age 68 ± 10 years; 21 male): TACE (n = 13); MAA/TARE (n = 7); PAE (n = 7). 16/18 (88.9

OBJECTIVES:The diagnostic value of conventional T1 mapping for noninvasive assessment of liver fibrosis is limited in the presence of hepatic steatosis. To evaluate the diagnostic value of water T1 (wT1) with continuous inversion-recovery Look-Locker (CIR-LL) method, integrating spiral readout, Dixon, and dictionary-based processing, for the fat-corrected assessment of hepatic fibrosis in patients with chronic liver disease (CLD). MATERIALS AND METHODS:In this prospective study, consecutive participants with CLD underwent liver magnetic resonance imaging (MRI), which included assessment of MR-elastography (MRE)-derived liver stiffness, proton density fat fraction (PDFF), T1 relaxation times using modified Look-Locker inversion recovery (T1-MOLLI) and extracellular volume fraction (ECV), and wT1 relaxation times. MRE served as the reference standard to evaluate the diagnostic performance of MRI-based mapping parameters. Significant fibrosis (≥F2) was defined as MRE-derived liver stiffness >3.66 kPa in patients with PDFF≤5%, or >3.14 kPa in patients with PDFF>5%. Statistical analysis included Student t test, receiver operating characteristic (ROC) analysis, and Spearman correlation coefficient. RESULTS:A total of 81 CLD patients (mean age, 50±14 y; 32 female; 40 patients with PDFF>5%) were included. All measured mapping values were significantly higher in patients with significant fibrosis compared with those without (eg, wT1: 628±82 vs. 546±41 ms, P <0.001). wT1 showed a strong correlation with MRE-derived liver stiffness, outperforming T1-MOLLI and ECV mapping [whole cohort: r =0.67 (wT1) vs. 0.53 (T1-MOLLI) vs. 0.48 (ECV); cohort with PDFF>5%: r =0.69 (wT1) vs. 0.44 (T1) vs. 0.49 (ECV); P <0.05 in each case, respectively]. wT1 had a superior diagnostic performance for the detection of significant fibrosis [whole cohort, area under the curve (AUC): 0.82 (wT1); 0.77 (T1-MOLLI); 0.73 (ECV), P <0.001 in each case; cohort with PDFF>5%, AUC: 0.84, P =0.002 (wT1), 0.70, P =0.04 (T1-MOLLI), 0.70, P =0.04 (ECV)]. CONCLUSION:Compared with T1-MOLLI and ECV mapping, the proposed fat-corrected CIR-LL wT1 method proved to be a more robust marker of hepatic fibrosis in CLD, also in the presence of hepatic steatosis.

BACKGROUND:To compare the diagnostic quality of deep learning (DL) super-resolution reconstructed breath-hold (BH) and free-breathing (FB) single-shot (SSH) black-blood T2-weighted short tau inversion recovery (STIR) imaging with standard BH T2-STIR in cardiovascular magnetic resonance (CMR). METHODS:In this prospective study, short-axis BH and FB SSH T2-STIR were added to a standard cardiomyopathy CMR protocol at 1.5T, and DL super-resolution reconstruction was performed. Two readers evaluated diagnostic quality and certainty using a five-point Likert scale. Presence of focal edema was assessed on T2-weighted sequences, including standard T2-STIR and T2 mapping (both used for reference clinical assessment) as well as SSH T2 STIR and DL-SSH T2-STIR. Friedman test and one-way ANOVA were performed. RESULTS:Eighty-one participants (mean age: 54 ± 20 years; 50 men) were included. No difference was found in edema detection between reference assessment and DL-SSH T2-STIR (both 26%, (21/81 participants)). Scan time was reduced by 63% for BH and 86% for FB DL-SSH T2-STIR compared to standard T2-STIR (90 ± 6 s vs. 35 ± 3 s vs. 243 ± 16 s; p<.0001). BH and FB DL-SSH T2-STIR achieved lower artifact burden (5 [IQR, 4-5] vs. 4 [IQR, 4-5] vs. 4 [IQR, 3-5]; p<.0001), superior image contrast and sharpness compared to standard T2-STIR, especially in non-cooperative or arrhythmic participants. BH and FB DL-SSH T2-STIR imaging provided higher diagnostic certainty than standard T2-STIR (5 [IQR, 5-5] vs. 5 [IQR, 5-5] vs. 4 [IQR, 4-5]; p<.0001). Edema visibility was superior in BH DL-SSH compared to BH-SSH and standard T2-STIR (5 [IQR, 4.8-5] vs. 4 [IQR, 3.3-5] vs. 4 [IQR, 3-4.8]; p<.0001). Inter-rater agreement was substantial to excellent in the rating of edema visibility (BH DL-SSH T2-STIR, κ: 0.73 [95% CI: 0.44-1.0]; BH SSH T2-STIR, κ: 0.79 [95% CI: 0.66-0.97]; standard T2-STIR, κ: 0.86 [95% CI: 0.71-1.0]). Slice level-analysis showed that BH DL-SSH T2-STIR consistently provided superior image quality in apical slices compared to BH SSH and standard T2-STIR (4 [IQR, 4-5] vs. 4 [IQR, 4-4] vs. 4 [IQR, 3-4]; p<.0001). CONCLUSION:DL-SSH imaging enabled ultrafast T2-STIR acquisition and robust edema assessment in routine clinical CMR.

Purpose To evaluate the impact of dual-phase three-dimensional (3D) modified Dixon (mDixon) steady-state MR angiography (MRA) on phase-dependent image quality and vessel measurements in pediatric patients with congenital heart disease (CHD). Materials and Methods In this retrospective single-center study, pediatric patients with CHD underwent cardiac MRI at 3 T, including free-breathing dual-phase 3D mDixon steady-state MRA with respiratory navigator gating and interleaved acquisition of diastolic and systolic images. Vessel diameters were measured independently by two readers. Overall image quality, including vessel border delineation and breathing-, motion-, flow-, and susceptibility-related artifacts, was evaluated using a five-point Likert scale (1 = nondiagnostic; 5 = excellent). Paired t tests or Wilcoxon signed rank tests were used for phase comparisons. Results A total of 47 patients (mean age, 6.4 years ± 3.5 [SD]) were included. Arterial vessel diameters were larger in systole than in diastole (eg, main pulmonary artery: mean, 23.2 mm ± 5.3 [SD] vs 19.1 mm ± 3.9; P < .01). Systolic image quality was higher for coronary arteries and pulmonary veins (eg, pulmonary veins: median, 5.00 [IQR, 4.00-5.00] vs 4.00 [IQR, 3.00-4.00]; P < .01), whereas diastolic image quality was higher for most other vessels (eg, main pulmonary artery: median, 5.00 [IQR, 4.00-5.00] vs 3.00 [IQR, 2.00-4.00]; P < .01). In two (4%) of 47 patients with nondiagnostic diastolic image quality due to arrhythmia or high heart rate variability, systolic imaging yielded diagnostic image quality. Conclusion Dual-phase 3D mDixon steady-state MRA enabled phase-dependent assessment of vessel dimensions and image quality in pediatric CHD. Additional systolic imaging improved diagnostic adequacy and visualization of selected structures within a single acquisition. Keywords: Pediatric Congenital Heart Disease, Modified Dixon MR Angiography Supplemental material is available for this article. © RSNA, 2026.

To compare the added benefit of optical see-through, the Magic Leap 2 (ML2), and video see-through, the Apple Vision Pro (AVP), head-mounted displays extended reality (XR) during percutaneous biopsies on an abdominal phantom. In this phantom-based prospective cohort study, sixteen radiologists (5 experienced and 11 beginners) performed six needle insertions: two without XR, two with ML2, and two with AVP. Two lesions of differing difficulty (depth and proximity to vessels) were targeted. Post-procedure CT measured accuracy (hitting the target and distance and angle to lesion center). Cognitive workload and user experience were assessed using a structured questionnaire with a rating of 0–20. Beginners benefited more from XR than experienced radiologists. Beginners improved their targeting accuracy with XR, especially for the more complex lesion closer to blood vessels, where success rates increased using XR (conventional: 1/11, ML2: 2/11, AVP: 4/11), with a decrease in mean distance to target (conventional: 17.95 ± 9.14 mm; ML2: 14.21 ± 6.27 mm; AVP: 12.02 ± 7.19 mm). Advanced radiologists had overall lower success rates (conventional: 1/5, ML2: 0/5, AVP: 1/5). XR also reduced puncture time for beginner (ML2: 68.1

New magnetic resonance imaging (MRI) gradient technology enables the acquisition of ultrahigh b-value diffusion-weighted imaging (DWI). We assessed its impact on image quality and Prostate Imaging Reporting and Data System (PI-RADS) scores in prostate MRI. Participants with cancer suspicion prospectively underwent 3-T prostate MRI (maximum gradient strength 200 mT/m). Sequences with b-values of 0/800, 1,500, 2,500, 3,500, and 4,500 s/mm² were acquired. Lesion conspicuity was rated from 1 (non-diagnostic) to 5 (excellent). Apparent signal-to-noise ratios (aSNR) and acquisition times were determined. Cumulative link mixed-effects models, repeated measures ANOVA, and Cohen/Fleiss κ statistics were used. A total of 107 participants, aged 67 ± 8 years (mean ± standard deviation), were included. Compared to DWI(b1500), the DWI(b2500), DWI(b3500), and DWI(b4500) acquisitions were worse regarding both lesion conspicuity (median score, 5 [interquartile interval 4–5] versus 4 [3–4] versus 2 [2–3] versus 2 [1–2], respectively; all p < 0.001) and aSNR (19.0 ± 7.5 versus 12.7 ± 4.8 versus 11.8 ± 4.1 versus 11.4 ± 2.6, respectively; all p < 0.001). Acquisition times increased from DWI(b1500) (107 ± 9 s) to DWI(b4500) (329 ± 26 s). Cohen κ for PI-RADS score agreement was good to moderate (DWI(b2500): 0.87 [confidence interval 0.81, 0.94]; DWI(b3500): 0.75 [0.65, 0.84]; DWI(4500): 0.61 [0.49, 0.72]). Acquired ultrahigh gradient DWI sequences with ultrahigh b-values in prostate MRI had worse image quality than standard b-values, while PI-RADS agreement between DWI(b1500) and DWI(b2500) was good. However, diagnostic estimates for clinically significant prostate carcinoma remained limited due to a small biopsy sample size (50/107 patients). Ultrahigh b-value DWI showed no improved diagnostic performance in comparison to standard b-value DWI regarding the identification of potential prostate cancer. Ultrahigh b-value should not replace standard high b-values (1,500 s/mm²) for imaging workup of patients with suspicion for prostate cancer.

Abstract Purpose To evaluate the occurrence of periaortic inflammation, including abscess formation, as a complication following endoleak repair after endovascular aortic repair. Materials and methods This retrospective single-center study analyzed 88 EVAR revisions for endoleak treatment performed in 59 patients between 2015 and 2025. Patients underwent one to three revision procedures. Endoleaks were classified by type, and treatment was performed via transarterial access or direct sac puncture with a posterior/translumbar access route. Embolization materials for type II endoleaks included n-butyl-2-cyanoacrylate, ethylene vinyl alcohol, and coils, while stent-graft extensions and transmural fixation were used for type I and III endoleaks. Results Of the 59 patients, 38 (64.4%) underwent one revision, 13 (22.0%) two revisions, and 8 (13.6%) three revisions. Treated endoleaks included type I (19.3%), type II (77.3%), type III (2.3%), and combined type II/III (1.1%). A total of 73 procedures (83.0%) were performed via transarterial access and 15 (17.0%) via direct sac puncture. Follow-up CT identified 6 cases of periaortic inflammation after type II endoleak repair, presenting with soft tissue mantle, fat stranding, or abscess formation. Symptom onset ranged from 33 to 108 days after the most recent intervention. No significant association was found between periaortic inflammation and access route or embolic material. Conclusion Periaortic inflammation and abscess formation are possible severe complications after endoleak repair following endovascular aortic repair. To date, these inflammatory changes have not been described systematically in the literature and warrant further investigation.

Background: Despite being a common gating method for cardiovascular magnetic resonance (CMR), electrocardiogram (ECG) gating has its disadvantages, and new gating strategies are desirable. An alternative CMR gating method is Doppler ultrasound (DUS) gating, which detects blood flow and ventricular movement. The aim of this study was to prove the feasibility of DUS gating as a novel CMR gating method in a clinical patient population. Methods: In this prospective study, patients underwent clinically indicated CMR. Balanced steady-state free precession two-dimensional cine sequences in short-axis and 4-chamber views were acquired using ECG and DUS gating. DUS and ECG signal were recorded simultaneously. Time difference between R-wave and DUS systolic trigger detection was defined as trigger delay, the standard deviation of trigger delays as trigger jitter. Left and right ventricular parameters were assessed: left and right ventricular ejection fraction (LVEF, RVEF) and left and right ventricular end-diastolic volume index (LVEDVI, RVEDVI). Overall image quality was assessed using a 5-point Likert scale (5 = excellent to 1 = non-diagnostic). For statistical analysis, paired t-test, Wilcoxon test, Pearson correlation, and intraclass correlation coefficient (ICC) were employed. Results: Twenty-one patients (7 female) were included (age: 45.4 +/- 19.7 years; body mass index: 27.6 +/- 5.5 kg/m2). DUS mean trigger delay was 128 +/- 28 ms. DUS mean trigger jitter was 23 +/- 13 ms. Overall image quality showed no difference between ECG and DUS gating (e.g., short axis: 5 [interquartile range (IQR) 3-5] vs 4 [IQR 3.5-5]; P = 0.21). Quantitative analysis revealed no differences between ECG and DUS gating: LVEF (53.2 +/- 9.2% vs 52.3 +/- 9.1%; P = 0.18; ICC 0.97 [95% confidence interval [CI] 0.93-0.99]), LVEDVI (84.5 +/- 15.8 mL/m2 vs 83.3 +/- 15.8 mL/m2; P = 0.06; ICC 0.99 [95% CI 0.98-1.00]), RVEF (52.8 +/- 8.0% vs 51.6 +/- 7.2%; P = 0.06; ICC 0.96 [95% CI 0.89-0.99]), and RVEDVI (80.8 +/- 17.6 mL/m2 vs 80.9 +/- 16.5 mL/m2; P = 0.91; ICC 0.98 [95% CI 0.96-0.99]). In one patient with a prominent lingula of the lung image quality was non-diagnostic with DUS gating. Conclusion: CMR gating with DUS is feasible and can offer an equivalent performance to ECG regarding image quality and quantitative parameter assessment.

Objectives The aim of this study was to assess the impact of an iterative metal artifact reduction (iMAR) algorithm combined with virtual monoenergetic images (VMIs) for artifact reduction in photon-counting detector computed tomography (PCDCT) during interventions. Materials and Methods Using an abdominal phantom, we conducted evaluations on the efficacy of iMAR and VMIs for mitigating image artifacts during interventions on a PCDCT. Four different puncture devices were employed under 2 scan modes (QuantumSn at 100 kV, Quantumplus at 140 kV) to simulate various clinical scenarios. Image reconstructions were initially performed without iMAR and subsequently with iMAR settings. The latter was tested with 7 different metal presets for each case. Furthermore, iMAR-reconstructed images were paired with VMIs at energy levels of 70 keV, 110 keV, 150 keV, and 190 keV. Qualitative assessments were conducted to evaluate image quality, artifact expression, and the emergence of new artifacts using a Likert scale. Image quality was rated on a scale of 1 (nondiagnostic) to 5 (excellent), whereas artifact severity was rated from 0 (none) to 5 (massive). Preferences for specific iMAR presets were documented. Quantitative analysis involved calculating Hounsfield unit (HU) differences between artifact-rich and artifact-free tissues. Results Overall, 96 different scanning series were evaluated. The optimal combination for artifact reduction was found to be iMAR neurocoils with VMIs at 150 keV and 190 keV, showcasing the most substantial reduction in artifacts with a median rating of 1 (standard: 4). VMIs at higher keV levels, such as 190 keV, resulted in reduced image quality, as indicated by a median rating of 3 (compared with 70 keV with a median of 5). Newly emerged artifact expression related to reconstructions varied among intervention devices, with iMAR thoracic coils exhibiting the least extent of artifacts (median: 2) and iMAR neurocoils displaying the most pronounced artifacts (median: 4). Qualitative analysis favored the combination of iMAR neurocoils with VMIs at 70 keV, showcasing the best results. Conversely, quantitative analysis revealed that the combination of iMAR neurocoils with VMIs at 190 keV yielded the best results, with an average artifact expression of 20.06 HU (standard: 167.98 HU; P < 0.0001). Conclusions The study underscores a substantial reduction in artifacts associated with intervention devices during PCDCT scans through the synergistic application of VMI and iMAR techniques. Specifically, the combination of VMIs at 70 keV with iMAR neurocoils was preferred, leading to enhanced diagnostic assessability of surrounding tissues and target lesions. The study demonstrates the potential of iMAR and VMIs for PCDCT-guided interventions. These advancements could improve accuracy, safety, efficiency, and patient outcomes in clinical practice.

Lower rates of hepatic encephalopathy (HE) following left-sided transjugular intrahepatic portosystemic shunt (TIPS) placement have been hypothesized to stem from a distinct ammonia distribution within the portal venous system. This prospective study investigated ammonia concentrations at five portal and splanchnic venous sites in 50 fasting cirrhotic patients (20 female [40%]; mean age: 60.4 years) prior to TIPS implantation for ascites (33/50, 66%) or variceal bleeding (17/50, 34%). While ammonia levels were significantly higher in the superior mesenteric vein (mean: 143 µg/dl) compared to the splenic vein (mean: 66 µg/dl; p < 0.001), mean ammonia concentrations in the right (104 µg/dl) and left (107 µg/dl) portal vein branches were found to be equivalent (p = 0.008 for equivalence). No systematic differences between right and left portal vein ammonia were observed. These findings suggest that in fasting cirrhotic patients, local ammonia levels in the main portal vein branches do not differ significantly. Therefore, other factors likely contribute to any observed differences in HE rates related to TIPS placement site, warranting further investigation into alternative mechanisms.

Large language models (LLMs) like Generative Pre-trained Transformer 4 (GPT-4) can assist in detecting errors in radiology reports, but privacy concerns limit their clinical applicability. This study compares closed-source and privacy-compliant open-source LLMs for detecting common errors in radiology reports. A total of 120 radiology reports were compiled (30 each from X-ray, ultrasound, CT, and MRI). Subsequently, 397 errors from five categories (typographical, numerical, findings-impression discrepancies, omission/insertion, interpretation) were inserted into 100 of these reports; 20 reports were left unchanged. Two open-source models (Llama 3-70b, Mixtral 8x22b) and two commercial closed-source (GPT-4, GPT-4o) were tasked with error detection using identical prompts. The Kruskall–Wallis test and paired t-test were used for statistical analysis. Open-source LLMs required less processing time per radiology report than closed-source LLMs (6 ± 2 s vs. 13 ± 4 s; p < 0.001). Closed-source LLMs achieved higher error detection rates than open-source LLMs (GPT-4o: 88