Objective: Pulmonary arterial hypertension is a severe complication of systemic lupus erythematosus (SLE). Current screening methods often miss early vascular changes. This study aimed to characterize subclinical pulmonary hemodynamic alterations in SLE patients without known pulmonary arterial hypertension using four-dimensional (4D) flow cardiovascular magnetic resonance (CMR) and to investigate their association with left ventricular diastolic function.
Materials and methods: Twenty-five SLE patients without known pulmonary arterial hypertension and 25 age-matched healthy controls were enrolled. All participants underwent 3-T 4D flow CMR to quantify hemodynamic parameters, including wall shear stress (WSS), flow volume, and relative pressure in the pulmonary arteries. SLE patients were further stratified based on echocardiographic assessment of diastolic function to analyze hemodynamic coupling.
Results: Compared to controls, SLE patients exhibited significantly lower maximum WSS in the main pulmonary artery (0.29 versus 0.33 Pa, p = 0.040) and asymmetric flow redistribution, characterized by higher relative pressure in the left pulmonary artery (0.54 versus 0.30 mmHg, p = 0.008) and increased flow rate in the right pulmonary artery (3.51 versus 2.90 L/min, p = 0.015). Qualitative analysis revealed vortical flow patterns in SLE patients. Subgroup analysis demonstrated that the reduction in WSS was primarily driven by patients with diastolic dysfunction (p = 0.006 versus controls).
Conclusion: SLE patients without pulmonary arterial hypertension exhibit distinct subclinical pulmonary hemodynamic alterations, including lower WSS and flow asymmetry. These alterations are intimately coupled with left ventricular diastolic dysfunction, suggesting that 4D flow CMR serves as a sensitive noninvasive tool for early risk stratification in this population.
Relevance statement: 4D flow CMR identifies subclinical pulmonary hemodynamic alterations coupled with diastolic dysfunction in SLE patients, serving as a sensitive noninvasive tool for early risk stratification before irreversible vascular remodeling occurs.
Key points: SLE patients without known pulmonary arterial hypertension show early pulmonary blood flow changes. 4D flow CMR detected asymmetric pulmonary flow redistribution in SLE patients. SLE patients exhibited altered left atrial function despite normal ventricles. Pulmonary flow changes correlated with left atrial remodeling in SLE. 4D flow CMR detects subclinical pulmonary hemodynamic differences in SLE.
Radiomics is a growing field in medical imaging that transforms images into high-dimensional quantitative data, offering insights into disease diagnosis, prognosis, and treatment planning. Using advanced computational techniques, radiomics uncovers patterns invisible to the human eye, playing a key role in precision medicine. However, the adoption of radiomics faces several barriers, including a lack of standardization, reproducibility challenges, and difficulties in clinical implementation. To address these challenges, a practical 10-step recipe is proposed to guide researchers in conducting effective radiomic studies: (1) identify a genuine clinical need and application; (2) establish a comprehensive database; (3) implement robust quality assurance and preprocessing; (4) ensure accurate image segmentation; (5) extract quantitative imaging features; (6) prioritize feature selection and dimension reduction; (7) consider integration of clinical and multi-omics data; (8) construct predictive models with machine learning techniques; (9) evaluate model performance using appropriate metrics; (10) translate models into clinical practice and workflow integration. This recipe emphasizes research rationale and methodologies, ensuring that the studies are aligned with real clinical needs, employing advanced techniques, and promoting reproducibility. By addressing these challenges through a structured approach, radiomics can transition from a research discipline to a clinical tool, contributing to more personalized and effective patient care. RELEVANCE STATEMENT: A structured 10-step framework is proposed to guide radiomic research, addressing key challenges in standardization and implementation. This practical guide supports any professional aiming to start in radiomics or adopt best practices, promoting reproducibility and clinical relevance in precision imaging workflows. KEY POINTS: Radiomics extracts quantitative data from medical images for improved diagnosis and treatment. Reproducibility, standardization issues, and clinical implementation barriers are among the main challenges of the technique. Data quality, feature selection, and machine learning are key to meaningful analysis. A structured 10-step guide for conducting reliable radiomic studies is proposed, taking a step toward a standardized workflow.
Objectives: Quantitative postprocessing of perfusion-weighted magnetic resonance imaging, including fractional tumor burden (FTB) maps, provides better visualization of the heterogeneous nature of glioblastomas. This study aimed to determine whether FTB maps help in distinguishing tumor progression (TP) from treatment-related abnormalities (TRA) in post-treatment glioblastoma patients.
Materials and methods: Unenhanced and contrast-enhanced T1-weighted and perfusion-weighted sequences of patients with new contrast-enhancing lesions were retrospectively included. Semiautomatic segmentation of these lesions was performed. Using predefined relative cerebral blood volume (rCBV) thresholds, voxels within this segmentation were classified as FTBlow, FTBmid, or FTBhigh. Patient outcome was determined by clinical and radiological follow-up. Non-parametric statistics were used to compare the FTB quantification. Diagnostic accuracy was evaluated with the area under the receiver operating characteristic curve (AUROC) and Youden's J. The difference between AUROCs was tested using bootstrapping.
Results: Fifty-nine patients were included, 35 of them showing TP (59%). The percentages of voxels classified as FTBlow and FTBhigh were significantly different between the groups (p = 0.031 and p = 0.010, respectively). Using the percentage of voxels classified as FTBhigh as a cutoff to differentiate TP from TRA yielded an AUROC of 0.70 (95% confidence interval: 0.56‒0.84), while FTBlow yielded 0.67 (0.52-0.82), without a significant difference (p = 0.466). The highest sensitivity and specificity based on the cutoff of 24% of voxels classified as FTBhigh coverage, were 63% and 79%, respectively.
Conclusion: FTB quantification yielded fair accuracy in the early detection of glioblastoma TP. Future research is needed to investigate how to use FTB maps in clinical practice.
Relevance statement: Early discrimination between TP and TRA, even with fair accuracy, can help in alleviating some uncertainty in glioblastoma patients. A clear visualization of lesion heterogeneity provided by FTB-maps could allow for more targeted treatment options and targeted follow-up.
Key points: Follow-up of patients with glioblastoma is complicated by the similar appearance of treatment effects and tumor growth on MRI. Perfusion imaging provides a basis for the creation of FTB maps. These visualize the heterogeneity of brain lesions. Quantitative analysis of FTB maps can help differentiate tumor growth from treatment effect with reasonable accuracy.
Objective: This study evaluates the feasibility of photon-counting detector CT (PCD-CT)-based coronary CT angiography (CCTA) using ultra-low flow contrast rate while maintaining diagnostic image quality.
Materials and methods: In this prospective trial, 292 patients underwent CCTA assigned to one of three protocols: ultra-low (1.5-1.8 mL/s) or routine (4.0-5.0 mL/s) contrast injection with PCD-CT, or routine injection with EID-CT. All scans utilized a high-pitch prospective electrocardiogram-triggering acquisition. PCD-CT images were reconstructed at 45 keV (ultra-low) or 60 keV (routine). Objective image quality was quantitatively assessed by measuring vessel attenuation, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). Subjective image quality parameters (vascular contrast, image noise, artifacts, and vessel clarity) were independently evaluated by two blinded readers using a 4-point Likert scale (1: non-diagnostic; 2: adequate; 3: good; 4: excellent).
Results: Objective image quality demonstrated comparable attenuation, CNR, and SNR in proximal coronary segments across all groups (all p > 0.05). The ultra-low PCD-CT protocol significantly lowers attenuation in the distal LAD (373.20 ± 49.58 HU) compared to routine protocols (PCD-CT: 393.52 ± 49.38 HU; EID-CT: 396.72 ± 47.55 HU; p = 0.01). While distal vessel clarity scores were modestly reduced in distal vessel clarity (ultra-low PCD-CT: 2.91 ± 0.81 versus routine PCD-CT: 3.58 ± 0.50 versus routine EID-CT: 3.54 ± 0.50; p < 0.01).
Conclusion: For patients with difficulty establishing venous access routes, ultra-low contrast agent flow rates in PCD-CT maintain objective image quality comparable to that of standard protocols, with acceptable diagnostic performance despite slight reductions.
Relevance statement: Photon-counting detector CT (PCD-CT) maintains objective coronary CT angiography image quality comparable to standard protocols even at ultra-low contrast flow rates (1.5-1.8 mL/s), offering a clinically acceptable and safer alternative for patients with challenging venous access.
Key points: First validation of ultra-low flow contrast rate CCTA using photon-counting CT (PCD-CT). Ultra-low flow rates maintain objective image quality (CNR/SNR) versus routine protocols. PCD-CT enables 50% contrast reduction without diagnostic compromise.
Objective: There is no satisfactory model for predicting the therapeutic response to chemotherapy of nasopharyngeal carcinoma (NPC). We developed a nomogram using tumor-stroma ratio (TSR) and histogram features from pretreatment synthetic magnetic resonance MRI (SyMRI) to assess induction chemotherapy (IC) response in NPC.
Materials and methods: Data from 185 NPC patients were retrospectively collected from July 2022 to November 2023 (training cohort), and 82 NPC patients were prospectively enrolled from December 2023 to July 2024 (test cohort). A nomogram was developed to predict IC response using logistic regression based on clinicopathological and imaging features from SyMRI T1-, T2-, and proton density (PD)-weighted images, and apparent diffusion coefficient (ADC) maps. The nomogram was validated in the test cohort.
Results: Among the 267 patients (187 males, 80 females), with a mean age of 52.2 years (ranging 43.5-58.7), 181 were responders. Histogram features from ADC and T2-map did not differentiate non-responders (all p ≥ 0.220). A clinicopathological model based on TSR and a SyMRI model using T1map_mean and PDmap_Kurtosis were developed. In the test cohort, The nomogram, combining TSR, T1map_mean, and PDmap_Kurtosis, achieved an area under the curve (AUC) of 0.836 (95% CI: 0.690-0.914), outperforming the clinicopathological model (AUC of 0.711, 95% CI: 0.577-0.809, p = 0.015) and SyMRI model (AUC of 0.774, 95% CI: 0.623-0.822, p = 0.003).
Conclusion: The nomogram combining TSR and histogram parameters from pretreatment SyMRI showed a good performance in predicting IC response for NPC, superior to those of clinicopathological and SyMRI models.
Relevance statement: A nomogram based on pretreatment synthetic MRI and clinicopathological features can help in selecting patients as candidates for IC.
Key points: NPC patients with high TSR demonstrated sensitivity to IC. The nomogram, integrating TSR and synthetic MRI parameters, achieved a significantly high predictive performance. The nomogram may be a reliable tool for predicting the response to IC.
Objective: Irreversible electroporation (IRE) is a non-thermal ablation technique suitable for tumors near critical structures, but its widespread use is limited by technical complexity and the need for multiple electrodes. This study aimed to evaluate the feasibility, safety, and efficacy of a stereotactic percutaneous two-needle IRE approach for small liver tumors in anatomically challenging locations.
Materials and methods: In this retrospective study, 17 consecutive patients with 18 primary or secondary liver tumors (≤ 2.0 cm) adjacent to critical anatomical structures underwent CT-navigated stereotactic two-needle IRE between December 2021 and May 2025. Ablation was performed with a high-dose protocol (2 × 90 pulses, 90 µs, > 20 A). Primary endpoints were primary technique efficacy (PTE) and local tumor progression (LTP); secondary endpoints included complications. Needle placement was assessed through geometric analysis.
Results: PTE was obtained in 17/18 tumors (94.4%, 95% confidence interval (CI): 72.7-99.9%). At a median follow-up of 12.4 months, LTP occurred in 1/18 tumors (5.6%, 95% CI: 0.1-27.3%). No complications or procedure-related mortality were observed. Geometric analysis showed high accuracy of stereotactic needle placement, while treatment failure was associated with suboptimal geometry.
Conclusion: Stereotactic percutaneous two-needle IRE seems to be technically feasible with a favorable safety profile for small liver tumors in anatomically challenging locations and may offer a simplified alternative to multielectrode approaches. However, given the small, retrospective single-center design, these findings are preliminary and require prospective multicenter validation to establish oncologic effectiveness and generalizability.
Relevance statement: Stereotactic two-needle irreversible electroporation offered a simplified, safe, and effective alternative to multielectrode ablation, potentially broadening treatment options for liver tumors near critical structures and improving accessibility, reproducibility, and outcomes in interventional oncology.
Key points: First systematic clinical evaluation of stereotactic two-needle irreversible electroporation (IRE) for liver tumors. Two-needle configuration with high-dose protocol simplifies IRE compared with standard multielectrode approaches. This proof-of-concept study demonstrates high efficacy and absence of complications in small liver tumors near critical structures. Two-needle IRE may broaden clinical applicability in anatomically challenging locations.
Objective: Retropharyngeal edema (RPE) on MRI in patients with acute neck infection is associated with disease severity. We explored the potential role of RPE volume as a quantitative marker and developed a convolutional neural network (CNN) for automated RPE volume segmentation.
Materials and methods: Volumes of RPE were manually segmented from T2-weighted fat-suppressed Dixon magnetic resonance (MR) images from 244 patients. These volumes were correlated with clinical variables, such as the need for intensive care unit (ICU) admissions, C-reactive protein (CRP) levels, maximal abscess diameter, and length of hospital stay (LOS). Manually segmented masks were used to train a CNN.
Results: Patients who required ICU admission had significantly higher RPE volumes than those who did not, and RPE volume outperformed the binary RPE (presence/absence) in classification analysis of ICU admissions. Furthermore, RPE volume correlated positively with LOS, CRP, and maximal abscess diameter. At the slice level, the deep learning (DL)-based model achieved its highest area under the receiver operating characteristic curve (AUROC) in sagittal slices (98.2%) and its highest Dice similarity coefficient in axial slices (0.534).
Conclusion: RPE volume is a promising quantitative imaging biomarker associated with relevant clinical outcomes in acute neck infections. Our DL-based model enables automated quantification of RPE volume.
Relevance statement: RPE volume provides clinically meaningful information in acute neck infections, outperforming binary classification in predicting disease severity and correlating with key clinical outcomes. Automated DL-based segmentation accurately locates the RPE and provides a moderate quantitative measurement of RPE volume, supporting its potential as a clinical imaging biomarker.
Key points: RPE volume correlated with markers of severe illness and outperformed binary RPE classification. We developed a DL-based algorithm for slice-wise classification and automatic segmentation of RPE. The classification model achieved excellent performance, while segmentation yielded modest Dice similarity coefficients consistent with prior imaging-based tumor segmentation algorithms.
Background: T2-weighted imaging (T2WI) of the liver suffers from prolonged scan times and respiratory motion artifacts. Deep learning (DL)-based reconstruction can accelerate acquisition while maintaining diagnostic quality. We compared respiratory-gated (RG) and breath-hold (BH) DL-T2WI to radial k-space sampling acquisition and reconstruction with motion suppression (ARMS)-T2WI and evaluated how respiratory characteristics affect image quality.
Materials and methods: We prospectively enrolled 120 participants who underwent 3-T RG DL-, BH DL-, and ARMS-T2WI. Three radiologists evaluated image quality and lesion conspicuity using a 5-point scale. Respiratory characteristics were extracted from breathing curves.
Results: All sequences showed comparable lesion-to-liver contrast ratios (p = 0.139), detection rates (p = 0.106), and lesion conspicuity scores (p = 0.990). RG DL-T2WI showed higher overall image quality compared to BH DL-T2WI (p = 0.027), and similar scores to ARMS-T2WI (p = 0.106). A respiratory score calculated using four parameters predicted ARMS-T2WI image quality with an area under the receiver operating characteristic curve (AUROC) of 0.836 (95% confidence interval 0.638-0.968) in the validation set. For RG DL-T2WI, a respiratory score using seven parameters achieved an AUROC of 0.831 (0.652-0.967) in the validation set. Standard deviation of the respiratory amplitude (SDamp) was an independent factor for BH DL-T2WI image quality (validation set, odds ratio 0.297, p = 0.049). For patients with high SDamp, RG DL-T2WI provided better image quality compared to BH DL-T2WI (68.6% versus 14.3%, p < 0.001).
Conclusion: Both RG and BH DL-T2WI offer image quality comparable to ARMS-T2WI. Respiratory metrics derived from breathing curves may facilitate personalized liver imaging.
Relevance statement: Both respiratory-gated and breath-hold T2WI with deep learning reconstruction showed comparable image quality to T2WI based on radial k-space sampling strategies. Respiratory parameters enable personalized magnetic resonance liver imaging workflows.
Key points: Respiratory-gated and breath-hold deep learning T2WI exhibited satisfactory image quality. Respiratory curve traits variably impact T2WI quality, guiding personalized imaging workflows. Respiratory-gated deep learning-reconstructed T2WI benefits patients with breath-holding difficulties in liver MRI.
Objective: Accurate identification of the trapezium is crucial for trapeziometacarpal (TMC) arthroplasty but remains challenging on standard radiographs due to overlapping anatomy. Artificial intelligence has shown promise in musculoskeletal imaging, yet its application to small joints is limited.
Materials and methods: We retrospectively analyzed 624 thumb radiographs, of which 519 met the inclusion criteria. Radiographs of insufficient quality-blurred images or non-centered TMC joints-were excluded by consensus of two hand surgeons. Manual trapezium annotations performed by an expert surgeon were reviewed by two additional surgeons. Inter-observer agreement was assessed on 10% of cases using Cohen κ. We developed a two-stage deep learning pipeline combining You Only Look Once (YOLO)v8 for trapezium detection with U-Net for segmentation. Its performance was compared with the standalone U-Net, segment anything model (SAM), and Mobile-SAM. Detection accuracy was measured using mean average precision (mAP), while segmentation was evaluated with Dice similarity coefficient (DSC) and intersection over union (IoU).
Results: YOLOv8 achieved a detection mAP of 99.5%. The combined YOLOv8 + U-Net model yielded a DSC of 94.2% and an IoU of 89.1%, outperforming U-Net (DSC 89.5%, IoU 81.2%), SAM (Dice 88.8%, IoU 80.3%), and Mobile-SAM (Dice 88.9%, IoU 80.5%). Inter-observer agreement was excellent (κ = 0.89, DSC = 93.8%).
Conclusion: The proposed two-stage pipeline provides accurate, reproducible trapezium segmentation on radiographs, outperforming widely used models. This approach may enhance preoperative planning and intraoperative guidance in TMC arthroplasty.
Relevance statement: This two-stage AI pipeline enables precise trapezium segmentation on thumb radiographs, supporting improved surgical planning and intraoperative guidance in TMC arthroplasty, with potential to enhance implant placement accuracy and patient outcomes.
Key points: A two-stage AI pipeline (YOLOv8 + U-Net) accurately detects and segments the trapezium on thumb radiographs. The method outperforms popular segmentation models and achieves expert-level reproducibility. This tool may enhance surgical planning and intraoperative guidance for TMC arthroplasty.
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：