Background/Objectives: The volume computed tomography (CT) dose index (CTDIvol) is a scanner output, not an organ dose, and cannot express how tube-current modulation varies along a patient. An organ-specific weighted CTDIvol addressing this has been reported before, in single-institution cohorts and often from inputs routine archives do not retain. New here is not the quantity but what an open, multi-vendor operationalisation reveals: whether its inputs survive archive curation and what the fallback costs when they do not. Methods: Forty abdominal CT series, ten per manufacturer, were drawn from the Cancer Imaging Archive and twelve organs segmented with TotalSegmentator at inference. Of 480 requested organ–series combinations, 455 were produced. A rule-based acquisition-constancy criterion admitted 39 series. Results: Modulation weights spanned 0.59 to 1.69, so the index departs from the whole-scan CTDIvol by up to 70% within one acquisition. A recorded CTDIvol survived in 29 of 40 archived headers and was reconstructable in 5 and unavailable in 6, availability differing markedly between manufacturers. Forcing that reconstruction on series that did retain a value agreed to within 12% on three scanner models and diverged by 58% and 84% on two others. Estimated organ mass was broadly consistent with International Commission on Radiological Protection (ICRP) Publication 89 for liver and kidneys. Conclusions: This index is not an absorbed dose; the implementation is open.
Background: Lung surveillance in cystic fibrosis (CF) relies on chest radiography (CR) despite cumulative radiation. Low-field MRI may be a radiation-free alternative for children and young adults. Methods: We prospectively compared 0.55 T MRI and CR in 28 same-day examinations of 22 people with CF (mean age 13 ± 5 years). Three raters scored the disease with the MRI CF score and a modified Chrispin–Norman score, expanded from two to three zones per hemithorax to match the six MRI lobes. Analyses used per-examination values averaged across the raters (Wilcoxon tests, intraclass correlation coefficients, ICC). Results: MRI yielded higher global scores (median 8.2 vs. 6.7, p = 0.002) and higher centrilobular opacity scores (p = 0.005, adjusted p = 0.03). Air trapping was also higher on MRI (p = 0.025) but not after correction for multiple comparisons (adjusted p = 0.12). The other categories did not differ. Both correlated inversely with FEV1 (p < 0.001), without a significant modality-by-FEV1 interaction (p = 0.08). Interobserver agreement was higher for MRI (ICC = 0.93) than CR (0.82), difference 0.12 (95% bootstrap CI 0.06 to 0.18). The findings held in a per-participant sensitivity analysis. Conclusions: 0.55 T MRI is a feasible, radiation-free alternative to CR in CF, yielding higher modality-specific scores and more consistent interobserver agreement. Because the instruments were only category-mapped and no CT reference was available, higher scores do not establish superior lesion detection.
Objective: Our objective was to describe morphologic features and longitudinal changes observed on high-resolution vessel wall imaging (HR-VWI) in a selected retrospective case series of patients evaluated for suspected inflammatory intracranial vasculopathy. Methods: This retrospective case series included patients who underwent 3T intracranial HR-VWI for suspected inflammatory intracranial vasculopathy. Clinical and imaging data were collected from medical records and imaging archives. Three radiologists reviewed baseline and follow-up examinations for vessel wall enhancement (VWE) pattern and grade, wall thickness, luminal narrowing, arterial distribution, diffusion-restricted lesions, and susceptibility-sensitive findings. Analyses were descriptive and patient-level. Each patient contributed one baseline and one final examination to longitudinal summaries; intermediate scans characterized individual trajectories only. Results: Nineteen patients were included, and 17 underwent serial HR-VWI. Median age was 46 years (IQR, 35.5–60.0 years), and 11 patients were female. Follow-up HR-VWI was available in 17 patients over a median of 391 days (IQR, 237–1819 days; range, 15–4034 days). VWE was present in all patients, with a purely concentric pattern in 15/19 (78.9%). The middle cerebral and internal carotid arteries were involved in 17/19 and 13/19 patients, respectively. Among the 17 patients with paired examinations, VWE grade decreased in 13 patients (76.5%) and remained stable in four patients (23.5%), with the median decreasing from 3.0 to 1.0. Maximum wall thickness decreased in 16 patients and remained stable in one, with the median changing from 2.0 mm to 1.4 mm. Conclusions: In this selected retrospective case series of patients evaluated for suspected inflammatory intracranial vasculopathy, 3T HR-VWI commonly demonstrated concentric VWE and provided descriptive longitudinal information regarding enhancement evolution. These findings should not be interpreted as evidence of diagnostic performance.
Visualization of coronary computed tomography angiography (CCTA) in standard cardiac planes requires reorientation of the three-dimensional image volume because the heart is double-oblique relative to the axial scan plane. To standardize this process across multi-phase cine CCTA, we developed an automated framework that segments cardiac structures and detects anatomic landmarks to define a patient-specific cardiac coordinate system. Landmark-derived left ventricular (LV) long-axis and LV-to-right ventricular transverse vectors were used to compute a rotation matrix for reorienting all 20 cine CCTA volumes. Six fixed slice planes were then defined to extract three long-axis (LAX) and three short-axis (SAX) slices at each phase. Matched CCTA and cardiac magnetic resonance (MR) slices were compared in 25 patients at diastasis, end-diastole (ED), and end-systole (ES), using automated segmentation-derived measurements of the combined LV cavity and LV myocardium area (LV + LVM). Agreement was evaluated with intraclass correlation coefficients (ICCs), Bland–Altman analysis, and repeated-measures mixed-effects models. ICCs were 0.927 (95% CI, 0.895 to 0.950) at diastasis, 0.950 (95% CI, 0.928 to 0.966) at ED, and 0.919 (95% CI, 0.887 to 0.943) at ES. Bland–Altman analysis showed small CT-positive biases of 1.05, 0.47, and 1.86 cm2 at diastasis, ED, and ES, corresponding to relative biases of 2.8%, 1.2%, and 5.8% of the phase-specific mean MR LV + LVM area. Mixed-effects models showed no statistically significant phase-level bias at diastasis or ED, whereas ES showed a small but significant CT-positive bias. These findings support automated CCTA reorientation as an interpretable framework for reproducible LAX/SAX slice extraction and future quantitative functional cine CCTA analysis.
Background: Accurate fracture interpretation on plain radiographs is critical for both trauma care and medico-legal decision-making, where reproducibility is as important as point accuracy. Although vision language models (VLMs) have shown promising diagnostic performance, their temporal stability in forensic radiography remains unclear. Methods: We analyzed 300 forensic radiographs (150 fracture-positive, 150 fracture-negative) from six long bones, independently evaluated by three emergency medicine physicians, three forensic medicine physicians, and three VLMs (ChatGPT-5.2, Gemini 3 Pro, Claude Sonnet 4.5) using an identical task format. Assessments included fracture presence and structured fracture subtype description (bone, morphology, displacement). VLM evaluations were repeated after one month under identical conditions. Results: Physician accuracy ranged from 79.7% to 98.7%. Emergency physicians reached the higher median sensitivity (92.0% against 76.7%), while specificity among the forensic readers was the more tightly clustered (median 92.0%, range 87.3–100.0%). ChatGPT-5.2 was the most accurate model (83.0%; sensitivity 70.0%, specificity 96.0%), followed by Gemini 3 Pro (77.7%), whereas Claude Sonnet 4.5 reached only 46.3% because of an extreme false-positive tendency (specificity 11.3%). Over one month, accuracy changed by −5.2, −4.1 and +8.6 percentage points, but these net figures concealed considerable case-level movement: within-model agreement ranged from near chance to substantial (mean Cohen κ 0.131 to 0.622), and the F1-score of Claude Sonnet 4.5 fell by 13.1 points despite its higher accuracy. Subtype descriptions were frequently correct once a fracture had been detected, but end-to-end subtype accuracy remained low. Conclusions: Current vision language models demonstrate encouraging diagnostic performance; however, their temporal reproducibility remains inadequate for independent medico-legal fracture interpretation. These findings highlight that reproducibility, in addition to diagnostic accuracy, should be considered a core benchmark when evaluating VLMs for high-stakes clinical and forensic use. Larger multicenter studies using independent external datasets are needed before forensic application is considered.
Objectives: To evaluate associations between shrinkage of ablated liver area on computed tomography (CT) over time after radiofrequency ablation (RFA) and clinical parameters related to liver function and fibrosis. Methods: Patients with hepatocellular carcinoma who underwent RFA and follow-up CT were retrospectively reviewed. The ablated area volume (AAV) on CT obtained within 1 week and around 6 months after RFA was measured, and reduction rate of AAV was calculated. The AAV reduction rate was compared among Child-Pugh classification, modified albumin-bilirubin (mALBI) grades, FIB-4 index categories, and lesion locations using generalized estimating equations (GEE). Univariable and multivariable GEE analyses were performed to evaluate associations between the AAV reduction rate and clinical parameters. Results: Fifty-three lesions in 41 patients (median age, 76 [range, 38–88] years, 24 men) were evaluated. The AAV reduction rate was significantly lower in Child-Pugh class B than class A (p < 0.001) and in mALBI grade 2b than grade 1 (p < 0.001) or grade 2a (p = 0.004). Significant differences were also observed among FIB-4 index groups (p < 0.001) and among lesion locations, with lower AAV reduction rates in medial and anterior segments than in lateral (p < 0.001) and posterior (p = 0.017) segments. Univariable GEE analyses showed significant associations between AAV reduction rate and cholinesterase, albumin, total bilirubin (T-Bil), prothrombin time, platelet count, and FIB-4 index (p < 0.05). Multivariable GEE analysis demonstrated that both albumin and T-Bil remained independently associated with AAV reduction rate (p < 0.001). Conclusions: The AAV reduction rate tended to be lower in patients with impaired liver function and advanced liver fibrosis.
Background/Objectives: Reliable uncertainty quantification (UQ) is a prerequisite for deploying automated systems in safety-critical medical imaging workflows, yet existing approaches either sacrifice computational efficiency or provide poorly calibrated confidence estimates. We present UQ-Mamba, a lightweight architecture that embeds uncertainty quantification natively into a Mamba state space model via linearized error propagation. Methods: UQ-Mamba yields per-prediction approximate epistemic and aleatoric uncertainty estimates in a single deterministic forward pass at only 9.5% additional inference overhead. We note that these components are heuristic approximations derived under three explicit assumptions (diagonal covariance, first-order linearization, and scalar mean-activation reduction) and have not been empirically validated as true Bayesian posteriors. By propagating learnable log-variance parameters through the SSM state transition matrix, UQ-Mamba bridges the gap between parameter efficiency and principled calibration without requiring stochastic sampling or multiple forward passes. Results: Evaluated across four medical imaging modalities—CT organ classification, colorectal histopathology, dermoscopy, and chest radiography—UQ-Mamba achieves 89.71% accuracy with ECE = 0.0217 on OrganMNIST using only 466K parameters (3.3× lower ECE than ResNet-50 at 50× fewer parameters; note that UQ-Mamba optimizes NLL, whereas ResNet-50 uses standard cross-entropy, which is a confounding factor in the ECE comparison), improves the Mamba baseline by 2.42 percentage points on PathMNIST (ECE = 0.1188 after temperature scaling), achieves 68.88% test accuracy with ECE = 0.0597 on HAM10000 dermoscopy (matching EfficientNet-B0 at 9× fewer parameters), and reaches mAUC = 0.8196 on CheXpert chest radiographs. Conclusions: Ablation studies confirm that the SSM propagation mechanism is necessary for meaningful uncertainty decomposition. These results establish uncertainty-aware SSMs as a promising proof-of-concept direction for calibrated, parameter-efficient medical image classification, with potential relevance to resource-constrained deployment settings pending further clinical validation.
Background: Spinocerebellar ataxia type 2 (SCA2) is an inherited neurodegenerative disorder characterized by progressive cerebellar degeneration. One difficulty in treating this disease lies in identifying preclinical carriers: individuals who carry the pathogenic ATXN2 mutation but remain asymptomatic with respect to motor manifestations. Though magnetic resonance imaging (MRI) has proven valuable in supporting the diagnosis of ataxia, traditional univariate approaches using linear measurements have shown limited ability to capture the complex anatomical changes that occur across the disease spectrum, particularly during the preclinical phase. Methods: This study employed a comprehensive multivariate approach to improve the classification of individuals across the SCA2 spectrum. We developed a multinomial logistic regression model incorporating multiple linear measurements derived from magnetic resonance imaging to discriminate between healthy controls (n = 72), preclinical carriers (n = 17), and patients with manifest SCA2 (n = 61). To mitigate inherent class imbalance, particularly in the smaller preclinical subgroup, we implemented the Synthetic Minority Over-sampling Technique (SMOTE), generating a balanced dataset that enhances the model’s ability to discern the distinctive anatomical features. This was compared to the model applied to the unbalanced data. An improvement was observed when applying SMOTE. Results: The multivariate model demonstrated discriminatory performance, achieving an overall accuracy of 80.7%. The ability to identify healthy controls (AUC: 0.96), preclinical individuals (AUC: 0.75), and clinical individuals (AUC: 95%). This represents an advance over previous univariate approaches, which have had difficulty capturing the neurodegenerative changes characteristic of the preclinical stage. Conclusions: By integrating multiple neuroimaging biomarkers into a multivariable model, this study provides a tool for early identification of preclinical SCA2 carriers. The ability to accurately classify these individuals opens an opportunity for early therapeutic intervention before irreversible neurological deterioration occurs. This approach shows promise for optimizing clinical trial design and personalized care in SCA2.
Background: Multimodal breast ultrasound, including B-mode imaging, color Doppler flow imaging, and elastography, provides complementary information for lesion characterization. However, effectively integrating heterogeneous modalities remains challenging due to inconsistent feature distributions, limited cross-modal interaction, computational cost in existing methods, and sensitivity to noise and missing data. Methods: We presented an efficient Cross-Modal Interaction and Dynamic Fusion Network (CIDFNet) for multimodal breast ultrasound analysis. The framework integrates a multi-scale feature enhancement module to improve modality-specific representations, a cross-modal interaction module to enable early-stage feature exchange across modalities, and a dynamic fusion strategy to adaptively combine modality information based on feature reliability estimation. In addition, an invertible neural network is incorporated to reconstruct missing modality features during training. Results: Experiments on an internal dataset of 248 patients with 1532 images show that CIDFNet obtains an AUC of 85.69%, accuracy of 75.51%, recall of 50.00%, F1-score of 62.50%, and precision of 83.33%, while requiring 49.51 M parameters and 79.79 G FLOPs, respectively. Under a simplified Gaussian noise perturbation setting, performance degradation is observed. Conclusions: CIDFNet presents a framework for multimodal breast ultrasound analysis that reflects a trade-off between performance and computational efficiency.
Purpose: To investigate the agreement on perfusion parameters derived from two different commercially available solutions for dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in patients with prostate cancer (PCa). Methods: A total of 50 patients (mean age, 71.6; range 56–86) who had undergone radical prostatectomy between December 2021 and September 2022 were included in this retrospective study. All patients had undergone DCE-MRI on a single 3T-MR scanner. Tumor segmentation on MR images was performed by two radiologists in consensus after radiologic-pathologic correlation using topographic maps as a reference standard. Subsequently, four perfusion parameters were calculated by dedicated commercially available solutions from two different vendors. Both solutions adopted a population-based arterial input function and an extended Tofts model as the pharmacokinetic model. The perfusion parameters were as follows; volume transfer constant (Ktrans), rate constant (kep), volume fraction of extravascular extracellular space (ve), and volume fraction of plasma (vp). The differences between paired measurements were compared by Bland–Altman analyses and the reproducibility was evaluated using the intraclass correlation coefficient (ICC). Results: The study population consisted of Gleason score (GS) 6 (n = 12), GS 7 (n = 34), GS 8 (n = 1), and GS 9 (n = 3). Significant differences were found for all parameters (p < 0.0001). Mean differences were as follows: Ktrans, −0.2102 (95% confidence interval; −0.2687 to −0.1518); kep, −0.7632 (−0.9005 to −0.6258); ve, −0.1507 (−0.2422 to −0.05907); vp, −0.02929 (−0.03383 to −0.02476). ICCs for average measures were as follows: Ktrans, 0.2989 (−0.2355 to 0.6021); kep, 0.6883 (0.4507 to 0.8231); ve, −0.1331 (−0.9967 to 0.3570); vp, 0.2653 (−0.3106 to 0.5881). Conclusion: All perfusion parameters were significantly different between the two solutions. Therefore, comparison of perfusion parameters across different solutions is not recommended.
Objective: The objective of this study was to investigate the relationship between intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) and dynamic contrast-enhanced MRI (DCE-MRI) parameters in soft tissue tumors (STTs). Methods: This retrospective study included patients with histopathologically confirmed STTs who underwent both DCE-MRI and IVIM-DWI between March 2022 and February 2024. Patients with prior therapy and lipomatous tumors were excluded. DCE-MRI parameters (Ktrans, Kep, Ve, iAUC) were obtained from pharmacokinetic maps using manually placed regions of interest (ROIs) in the most perfused tumor areas, avoiding necrotic and cystic regions. Corresponding ROIs were applied to IVIM-DWI maps. IVIM parameters (D, D*, f) were calculated using 11 b-values. Results: Twenty-nine patients (mean age, 56 ± 18 years; 14 malignant, 15 benign) were included. Interobserver agreement was excellent for DCE-MRI parameters, whereas IVIM-DWI parameters showed moderate-to-good agreement, with D showing the lowest reproducibility. In malignant tumors, f demonstrated strong positive correlations with Ktrans (r = 0.81, p < 0.001) and iAUC (r = 0.79, p < 0.001), both of which remained significant after correction for multiple comparisons. fD* was higher in malignant than in benign lesions in the unadjusted group comparison; however, diagnostic performance was not evaluated in the present study. No significant differences were observed for DCE-MRI parameters between benign and malignant tumors. Conclusions: IVIM-DWI parameters demonstrated associations with DCE-MRI metrics in malignant STTs and may provide complementary information regarding tumor perfusion. However, the findings should be interpreted cautiously because ROI analysis was limited to a single representative slice. Further validation using larger cohorts and volumetric tumor assessment is required.
Background/Objectives: Virtual Bronchoscopic Navigation is used to guide bronchoscopes toward peripheral pulmonary lesions, but broad clinical and research adoption remains limited by the cost of proprietary software and by segmentation failures in small distal airways that can interrupt path planning. This study presents Virtual Bronchoscopic Pathfinder, an open-source, web-based system designed to provide automated airway segmentation, robust path generation, and browser-based three-dimensional visualization. Methods: The system integrates five components: a connectivity-aware deep learning model for pulmonary airway segmentation using Connectivity-Aware Surrogate and Local-Sensitive Distance modules; TotalSegmentator for automated tumor localization; a topology-preserving three-dimensional thinning algorithm implemented in C++ for centerline extraction; a bidirectional Dijkstra algorithm operating on a three-tier anatomical cost field with centerline, airway lumen, and parenchymal costs; and a zero-footprint visualization interface built on vtk.js with synchronized axial viewing and interactive volume rendering. VBP was validated on 306 thin-section CT series from 154 subjects in the public Lung-PET-CT-Dx dataset. Results: Among the 306 CT series, 33 series (10.8%) were excluded because of scanner-specific segmentation artifacts. In the remaining 273 anatomically valid series, the system successfully generated complete end-to-end navigation paths for all cases. The overall pipeline success rate was therefore 273 of 306 series (89.2%). The web-based interface was also confirmed to operate without client-side installation across desktop, laptop, and mobile device configurations. Conclusions: Virtual Bronchoscopic Pathfinder demonstrates that a reliable and accessible virtual bronchoscopic navigation workflow can be constructed entirely from open-source components. By combining connectivity-aware segmentation, cost-field path planning, and browser-based visualization, the system provides a practical foundation for imaging informatics research and future development of intra-procedural bronchoscopic guidance.
Aim: This study aimed to evaluate the impact of cone beam computed tomography (CBCT) on preoperative surgical decision-making and risk assessment for mandibular third molar (MM3) extractions in cases identified as high-risk by orthopantomography (OPG). Materials and Methods: This prospective observational diagnostic study utilized the purposive sampling method, recruiting 50 MM3s from 33 patients (21 females, 12 males; mean age 24.24 ± 6.77 years, range 16–42). Samples were categorized into five distinct radiographic groups based on the proximity of roots to the inferior alveolar nerve (IAN) on OPG. The methodology involved a comparative 3D analysis to determine neurovascular contact, spatial orientation, and the presence of a cortical border. Surgical strategies, specifically the necessity for coronectomy or the lingual split technique, were reassessed following 3D evaluation. Postoperative neurosensory outcomes were recorded. Statistical analysis was performed using the Fisher–Freeman–Halton and Kruskal–Wallis tests. Results: CBCT identified direct IAC contact in 74% of the cases. In 18% of the cases initially deemed high-risk by OPG, CBCT revealed a safe distance, thereby altering the surgical approach. Tooth angulation (p = 0.012) and Pell and Gregory classification (p = 0.024) were significant predictors of contact. Temporary neurosensory disturbances occurred in 4% (n = 2) of the sample, specifically in cases where CBCT had confirmed the loss of nerve canal cortication. Conclusions: In accordance with the study aim, CBCT provides essential 3D data that refines surgical planning in nearly one-fifth of high-risk cases. The findings justify selective CBCT use, guided by the ALADA principle, to minimize iatrogenic injury.
Purpose: To develop and technically validate a reproducible multicentre MRI radiomics workflow for pancreatic cyst risk stratification using paired T1- and T2-weighted imaging from public datasets. Methods: Public datasets were screened and Cyst-X was selected as the primary cohort because it contained pancreatic MRI, risk labels, masks and metadata. A linked Cyst-X subset was enriched with metadata, filtered to an exact paired T1/T2 cohort, and processed through image–mask quality control, 1.0 mm isotropic resampling, intensity normalisation, PyRadiomics feature extraction, feature reduction and patient-level centre-held-out validation. The revised modelling strategy used a T2 + clinical all-patient primary analysis (n = 409) and a complete-case paired T1/T2 sensitivity analysis (n = 299). Results: The final cohort comprised 409 patients and 818 image-level rows across EMC, IU, MCF and NYU. All 818 image–mask pairs passed post-preprocessing QC. T2 radiomics were complete for all 409 patients; however, 110 T1 feature sets were missing, all from MCF. In the all-patient T2 + clinical model comparison, logistic regression achieved the highest macro-AUC (0.737). The T2 + clinical random forest comparator achieved macro-AUC 0.716 (95% CI 0.678–0.755), accuracy 0.545 (95% CI 0.496–0.592) and macro-F1 0.530 (95% CI 0.481–0.577). The paired T1/T2 complete-case random forest sensitivity model achieved macro-AUC 0.735 (95% CI 0.691–0.777), accuracy 0.575 (95% CI 0.520–0.632) and macro-F1 0.554 (95% CI 0.494–0.605). Conclusion: This study demonstrates the feasibility of constructing a reproducible public data MRI radiomics workflow for pancreatic cyst risk stratification. Model performance was modest, and independent external validation is required before clinical application.
Background/Objectives: To evaluate whether day-of-procedure ambient ozone exposure is associated with pneumothorax after CT-guided lung biopsy. Methods: This retrospective single-centre study included 160 CT-guided lung biopsies performed between January 2018 and February 2026. Environmental data from the day of biopsy were assigned from the nearest national monitoring station. The primary outcome was any pneumothorax on post-biopsy CT; the secondary outcome was drainage-requiring pneumothorax. Multivariable logistic regression included ozone exposure, emphysema, and access route through dependent lung area (ARDA). Ozone was analysed as a continuous variable per 10 μg/m3 and, exploratorily, using a ROC-derived threshold of ≥75.8 μg/m3. Restricted cubic splines assessed nonlinearity. Sensitivity models adjusted for needle size, biopsy system, operator identity, and season. Drainage-requiring pneumothorax was analysed using Firth logistic regression. Results: Pneumothorax occurred after 86 of 160 biopsies (53.8%), and 13 biopsies (8.1%) required drainage. Ozone was not associated with pneumothorax when modelled linearly (OR, 1.09 per 10 μg/m3; 95% CI, 0.97–1.23; p = 0.167). In exploratory threshold modelling, ozone ≥ 75.8 μg/m3 was associated with pneumothorax (OR, 2.76; 95% CI, 1.39–5.61; p = 0.004). Emphysema increased pneumothorax odds (OR, 2.16; 95% CI, 1.03–4.68; p = 0.047), whereas ARDA was protective (OR, 0.23; 95% CI, 0.11–0.45; p < 0.001). Spline analysis supported nonlinearity (p = 0.001). For drainage-requiring pneumothorax, only emphysema was significant. Conclusions: Ambient ozone showed an exploratory nonlinear association with pneumothorax after CT-guided lung biopsy, with a threshold signal around 70–80 μg/m3. ARDA was protective, whereas emphysema was associated with drainage-requiring pneumothorax.
Background/Objectives: Artificial intelligence (AI) can support lesion detection in gadolinium-based contrast agent-enhanced (GBCA-enhanced) breast MRI. However, its effectiveness on virtual contrast-enhanced (vCE) images remains unclear. This feasibility study evaluated the publicly available MAMA-MIA nnU-Net model trained on GBCA-enhanced data using an independent cohort of both GBCA-enhanced and vCE breast MRI. Methods: This IRB-approved retrospective study included the publicly available nnU-Net model trained on n = 1506 MAMA-MIA breast MRI scans and a cohort of n = 2126 in-house 3T breast MRI scans. A generative adversarial network (Pix2Pix-GAN) was developed on n = 1870 of the in-house scans and used to generate vCE data on the remaining independent n = 256 in-house cases. The MAMA-MIA nnU-net was applied to both GBCA-enhanced (GBCA) and corresponding vCE images. Ground-truth segmentations of malignant lesions served to calculate the Dice score, Hausdorff distance, and lesion dimension differences. Results: The final test set comprised n = 250 cases (n = 69 malignant, n = 181 benign). Lesion detection rates were 91% (n = 63/n = 69; 95% confidence interval (CI): 82.3–96.0%) for GBCA and 84% (n = 58/n = 69; 95% CI: 73.7–90.9%) for vCE. Two lesions missed in GBCA were identified by vCE. The Hausdorff distances were similar (GBCA: 6.4 (IQR: 3.2–9.3; 95% CI: 5.2–7.8) mm; vCE: 6.7 (IQR: 3.9–9.7; 95% CI: 5.3–8.0) mm, p = 0.564). The Dice scores showed minor differences (GBCA: 0.829 (IQR: 0.723–0.900; 95% CI: 0.786–0.865) vs. vCE: 0.826 (IQR: 0.720–0.857; 95% CI: 0.770–0.836); p < 0.001). vCE images had slightly higher non-target tissue segmentation (median 6072 mm3 vs. 5754 mm3). Conclusions: A GBCA-trained algorithm demonstrated some cross-domain transferability to vCE images, albeit with a reduced case-level sensitivity of 84% (95% CI: 73.7–90.9%) vs. 91% (95% CI: 82.3–96.0%). Based on these preliminary results, further research, including larger cohorts and more diverse datasets, is warranted.
Purpose: This study aimed to assess MRI-related claustrophobia severity and patient-reported experiences among Saudi patients to examine their associations with selected demographic variables. Methodology: A cross-sectional study was conducted using a structured questionnaire administered to 200 Saudi patients who had previously undergone MRI examinations. The questionnaire comprised five sections covering demographic data, phobia severity and patient-reported experiences before, during and after MRI examinations. Statistical analysis was performed using SPSS statistical package (IBM SPSS Statistics version 26, IBM Corp., Armonk, NY, USA), applying chi-square tests to examine associations between demographic variables and questionnaire responses. Results: A significant majority of participants, 76.5%, reported a positive MRI experience, whereas only 6.5% reported a negative experience. Shortness of breath during the MRI examination was the most frequently reported source of discomfort (75%). Significant associations were identified between demographic characteristics and phobia severity. Age and gender were significantly correlated with sudden fear responses, while educational level was strongly associated with receiving adequate pre-scan information and overall examination experience. Conclusions: Despite the high percentage of positive experiences, a notable proportion of participants reported anxiety-related distress during MRI examinations. The observed associations between demographic variables and claustrophobia-related responses suggest the potential value of patient-centred approaches, particularly improved pre-scan education, to improve the MRI-related patient experience and reduce anxiety-related distress.
Background: Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) interpolation and segmentation are critical for clinical diagnosis, anatomical quantification and personalized treatment. Most existing methods perform these two tasks separately, leading to computational redundancy and insufficient mining of shared spatial features. This study aims to construct an integrated multi-task learning framework for the synchronous processing of medical image interpolation and segmentation. Methods: We propose a unified multi-task framework named TASC-SwinMT for joint interpolation and multi-frame segmentation of CT and MRI images. It employs a shared SwinUNet encoder to extract general spatial features, matched with two task-specific decoders for frame prediction and mask generation. Three functional modules are designed for cross-task synergistic learning, and a dynamic multi-task loss function is used to balance objective optimization. Experiments are performed on Medical Segmentation Decathlon Task02_Heart and Task06_Lung datasets. Results: Our method outperforms baseline models and ablation variants in both tasks with outstanding accuracy and significantly reduced computational overhead. It exhibits superior performance in lesion boundary depiction, small object segmentation and inter-slice consistency, and anatomical prior constraints with frequency-domain modeling further enhance prediction quality. Conclusions: The cross-task feature sharing and joint optimization strategy are validated effective. The proposed TASC-SwinMT framework has favorable stability and generalization ability, providing a reliable solution for clinical medical image analysis.
Purpose: To evaluate how often history taking and physical examination are omitted before MRI referral and whether their omission is associated with clinical reasoning quality and MRI diagnostic yield. Materials and Methods: In this prospective study, adults undergoing MRI at a tertiary academic hospital were surveyed before imaging to determine whether the referring clinician had taken their history and performed a physical examination. Multivariable regression was used to assess determinants of omission and associations with clinical reasoning quality (defined as agreement between the suspected diagnosis and MRI findings) and MRI positivity (defined as findings relevant to the indication). Results: Among 275 patients (median age 61 years; 50.0% male), history taking was omitted in 18.2% of cases and physical examination was omitted in 70.9%. History taking was less likely during surveillance than during new/first visits (odds ratio (OR) 0.140, p < 0.001) and more likely when MRI was requested by residents rather than medical specialists (OR 4.645, p = 0.018). Physical examination was more likely when MRI was requested by residents (OR 3.174, p = 0.007) or nurse specialists/physician assistants (OR 3.145, p = 0.033), and less likely during follow-up visits (OR 0.183, p < 0.001) and surveillance visits (OR 0.061, p < 0.001). Omission of physical examination was not associated with clinical reasoning quality (p = 0.370). Neither omission of history taking nor omission of physical examination was associated with MRI positivity (p = 0.430 and p = 0.286, respectively). Conclusions: History taking and physical examination were often omitted before MRI referral. Although no statistically significant association was observed between omission of bedside assessment and clinical reasoning quality or MRI positivity, reduced bedside assessment may limit the clinical context informing referral and interpretation.
Background/Objectives: Left ventricular (LV) geometry reflects structural adaptation to aging and biological sex. While cardiac magnetic resonance (CMR) provides precise morphologic assessment, most prior studies have focused on volumetric and mass-based parameters rather than routinely reported linear indices. This study aimed to evaluate the influence of age and sex on LV geometry using wall thickness, LV end-diastolic diameter (LVEDD), and proportional indices derived from standard CMR reports. Methods: In this retrospective cross-sectional study, 95 adult patients who underwent clinically indicated CMR were included. LV wall thickness, LVEDD, relative wall thickness (RWT), and wall thickness-to- LVEDD ratio (WT/LVEDD) were recorded. Participants were stratified by sex and age groups (18–40, 41–60, >60 years). Group comparisons, correlation analysis, multivariable linear regression, logistic regression, and Age × Sex interaction testing were performed to evaluate independent associated parameters of LV morphology and concentric remodeling. Results: The mean age was 34.94 ± 16.00 years; 60.0% were male. Males had significantly larger LVED (43.12 ± 6.83 mm vs. 39.76 ± 6.11 mm, p = 0.014) and greater wall thickness measurements (p < 0.05 for septal and posterior wall thickness). Age showed a significant positive correlation with mean LV wall thickness (r = 0.275, p = 0.007) and WT/LVEDD ratio (r = 0.241, p = 0.019), but not with LVEDD (p = 0.414). In multivariable analysis, male sex was independently associated with larger LVED (B = 3.345, p = 0.017), whereas age was independently associated with WT/LVEDD ratio (B = 0.0018, p = 0.019). Logistic regression demonstrated that age independently increased the odds of concentric remodeling (OR = 1.041 per year, 95% CI: 1.011–1.072, p = 0.006). No significant Age × Sex interaction was observed. Conclusions: Advancing age was independently associated with proportional LV geometric remodeling, whereas male sex primarily influenced absolute ventricular dimensions. Routine CMR report-derived linear measurements were sufficient to detect these distinct structural patterns. These findings highlighted the feasibility of using standardized morphologic indices in daily clinical practice to identify early age-related concentric remodeling.
