ObjectivesTo characterize pleuroparenchymal abnormalities identified on computed tomography pulmonary angiography (CTPA) in cancer patients investigated for suspected pulmonary embolism (PE) and to compare imaging findings between patients with and without confirmed PE.MethodsThis single-center retrospective observational study included 113 consecutive cancer patients who underwent CTPA for suspected acute PE between January 2022 and January 2026. Radiology reports and electronic medical records were reviewed for demographic data, malignancy type and stage, PE-related imaging characteristics, and pleural and pulmonary parenchymal abnormalities. Imaging findings were compared between PE-positive and PE-negative groups. An exploratory multivariable model with a restricted set of imaging predictors was constructed to limit overfitting, with correction for multiple comparisons and bootstrap-based internal validation. Exploratory subgroup analyses by malignancy type and disease burden were also performed.ResultsPE was confirmed in 32 patients (28.3%). The median age was 59 years (IQR, 51–65), and 42.5% were female. The highest PE prevalence was observed in gastrointestinal, gynecologic, and thoracic malignancies. At nominal testing, infarct-like wedge-shaped opacity was more frequent in PE-positive patients (12.5% vs. 1.2%; OR, 11.43; p = 0.022), whereas ground-glass opacity (15.6% vs. 35.8%; OR, 0.33; p = 0.041) and interstitial/septal change (3.1% vs. 25.9%; OR, 0.09; p = 0.007) were less frequent. None remained statistically significant after false-discovery-rate correction (q = 0.133, 0.166, and 0.082, respectively). The parsimonious model had an apparent AUROC of 0.707 and an optimism-adjusted AUROC of 0.699, with acceptable grouped calibration (Hosmer–Lemeshow p = 0.522). No exploratory cancer-subgroup comparison met q < 0.05.ConclusionsPleuroparenchymal abnormalities are highly prevalent in cancer patients undergoing CTPA for suspected PE; however, most findings are nonspecific. Infarct-like wedge-shaped opacity generated the strongest exploratory PE-associated signal, but no individual imaging comparison survived multiplicity correction and model estimates remained imprecise. These findings should therefore be considered hypothesis-generating and require validation in larger cohorts.
PurposeGlioblastoma (GBM) exhibits variable treatment response and patients shown a median survival of 15 months. A major challenge lies in the difficulty of reliably assessing treatment response at an early stage, as it can usually only be evaluated on anatomical MRI performed up to three months after the start of treatment, when early interpretation may be complicated by the risk of pseudoprogression. Accurately predicting short-term and long-term survivors before radiotherapy and chemotherapy is therefore crucial. The objective of this study is to develop an AI model designed to predict survival outcomes based exclusively on pre-chemoradiotherapy MRI and clinical data.Material and methodsA total of 1,575 retrospective T1-Gd brain images were obtained from 175 patients treated for GBM between 2018 and 2023. A ResNet-51Q model pretrained on ImageNet was fine-tuned on pre-chemoradiotherapy MRI using transfer learning. To incorporate clinical data, deep hybrid learning (DHL) models were trained.ResultsAt patient-level, the AI model achieved 81.08% (CI 95%: 67.57–91.89), 68.42% (CI 95%: 47.06–88.89) and 94.44% (CI 95%: 81.82–1.00) for accuracy, sensitivity, and specificity, respectively, for the classification of short-term and long-term survivors using only pre-chemoradiotherapy MRI. At patient-level, the three-class model using only pre-chemoradiotherapy MRI, achieved 87.18% (CI 95%: 76.92– 97.44), 87.18% (CI 95%: 76.68–96.97), and 93.59% (CI 95%: 88.03–98.41) for accuracy, sensitivity, and specificity, respectively. In comparison at slice-level, the three-class model using only pre-chemoradiotherapy MRI, achieved 81.77% (CI 95%: 70.66–91.45), 81.77% (CI 95%: 71.07–91.33), and 90.88% (CI 95%: 85.25–95.74) for accuracy, sensitivity, and specificity, respectively. In comparison the DHL model with clinical data achieved an accuracy of 84.62% (CI 95%: 71.79–94.87), a sensitivity of 84.62% (CI 95%: 72.54–95.24) and a specificity of 92.31% (CI 95%: 86.07–97.53) and 76.32% (CI 95%: 63.16–89.47), 76.11% (CI 95%: 61.63–88.75) and 92.06% (CI 95%: 87.51–96.23) for four overall survival classification. For multiclass tasks, metrics were calculated using a macro-averaged approach.ConclusionThe combination of pre-chemoradiotherapy MRI and clinical data using the DHL model for early survival prediction; however, adding clinical variables yielded comparable prognostic performance rather than a demonstrable enhancement over imaging features alone. The use of AI models to predict survival outcome in GBM from pre-chemoradiotherapy data may support pre-chemoradiotherapy prognostic stratification.
BackgroundCardiac magnetic resonance elastography (MRE) is a non-invasive technique that quantifies myocardial stiffness by tracking shear wave propagation through the heart. Shear waves are generated externally and delivered to the thorax through a passive driver placed on the chest. Passive driver position is a critical yet underexplored determinant of measurement quality.PurposeTo compare two passive driver positions in 3.0 T cardiac MRE — the apical surface projection (ASP) and the left lateral chest wall (LCW) — and to determine which yields more reliable left ventricular (LV) myocardial stiffness measurements.MethodsTwenty-three healthy volunteers (mean age 43.0 ± 17.0 years; 14 men, 9 women) underwent two cardiac MRE acquisitions, with the passive driver placed first at the ASP and then at the LCW. LV myocardial stiffness and the octahedral shear strain signal-to-noise ratio (OSS-SNR) were measured at both positions. Differences between positions were analyzed using paired t-tests with Cohen's d. Inter- and intra-observer reproducibility were assessed using intraclass correlation coefficients (ICC) and Bland–Altman analysis.ResultsMyocardial stiffness was significantly lower at the ASP than at the LCW (9.0 ± 1.7 kPa vs. 14.7 ± 3.0 kPa; mean difference −5.8 kPa, 95% CI −7.2 to −4.3; P<0.001; Cohen's d=−1.73). OSS-SNR was significantly higher at the ASP (1.72 ± 0.48 vs. 1.25 ± 0.29; P<0.001; Cohen's d=0.85). At the ASP, stiffness measurements demonstrated excellent intra- and inter-observer agreement, with ICCs of 0.962 and 0.963, respectively. At the LCW, stiffness measurements demonstrated good inter-observer agreement and excellent intra-observer agreement, with ICCs of 0.839 and 0.914, respectively.ConclusionPassive driver placement at the ASP resulted in significantly higher OSS-SNR and superior reproducibility for LV myocardial stiffness measurements compared with LCW placement. Therefore, the ASP should be considered the preferred driver position for cardiac MRE in clinical and research settings.
BackgroundSevere contrast media reactions during computed tomography (CT) are rare but potentially life-threatening and require rapid recognition and coordinated emergency response. CT technicians and radiographers are often frontline professionals during contrast-enhanced CT, yet their preparedness in Jordan remains underexplored.ObjectiveTo assess knowledge of severe contrast media reactions among CT technicians and radiographers in Jordan and identify factors associated with total knowledge scores.MethodsA descriptive cross-sectional study was conducted using an online questionnaire adapted from a previously published tool. The questionnaire assessed five domains: general knowledge of contrast media, risk factors, precautions, recognition of severe reactions, and management. Scores were converted to percentages. Data were analyzed using descriptive statistics, Mann–Whitney U tests, Kruskal–Wallis tests, Spearman correlation, and multivariable linear regression.ResultsThe analysis included 314 participants. The median total knowledge score was 76.3% (IQR: 55.3–84.2). General knowledge showed the highest median score (100%, IQR: 80–100), whereas risk factor knowledge was the lowest-performing domain (65%, IQR: 50–80). Radiographers achieved higher total knowledge scores than technicians (78.9% vs. 71.0%, p = 0.002). Total knowledge also differed significantly across educational levels (p = 0.043). In multivariable regression, total knowledge was independently associated with gender, age group, master's-level education, professional role, experience >25 years, and self-reported confidence.ConclusionCT technicians and radiographers in Jordan demonstrated moderate-to-high overall knowledge of severe contrast media reactions, but important gaps were evident, particularly in risk factor recognition. Structured, recurrent, competency-based training programs focusing on risk assessment, reaction recognition, emergency management, and simulation-based practice are needed to strengthen preparedness and patient safety during contrast-enhanced CT.
MotivationArtificial intelligence (AI) is reshaping radiology through advances in machine learning, deep learning, and generative AI. As these technologies become embedded in diagnostic workflows, radiology education must adapt to prepare learners for AI-enabled clinical practice. This review aimed to synthesize current evidence, identify educational priorities, and highlight challenges and opportunities for implementation.IntroductionThe growing adoption of AI in medical imaging requires radiology curricula to extend beyond image interpretation and encompass AI literacy, ethical reasoning, critical appraisal, and human–AI collaboration. Understanding how AI is currently incorporated into radiology education is essential for developing effective training frameworks. This review examined educational approaches, learner outcomes, and implementation challenges associated with AI in radiology education.MethodologyA scoping review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR), using the Population–Concept–Context(PCC) framework to define the review question and eligibility criteria. Searches were performed in Scopus, PubMed, and IEEE Xplore for studies published between 2020 and 2025. After screening and eligibility assessment, 29 original studies were included. Educational outcomes were classified into skill development, engagement, and performance domains.Results and discussionSkill development was the most frequently investigated outcome (55.2%), followed by performance (27.6%) and engagement (17.2%). Approximately 86% of studies reported positive or improved educational outcomes. AI-based interventions enhanced learner confidence, AI literacy, diagnostic reasoning, and readiness for clinical implementation. Generative AI tools showed promise for tutoring, assessment, and self-directed learning but raised concerns regarding reliability, hallucinations, bias, and ethical use. Common barriers included limited faculty expertise, insufficient formal training, curricular overcrowding, inadequate infrastructure, and governance challenges.ConclusionAI has significant potential to strengthen radiology education through personalized learning, competency development, and technology-enhanced training. Nevertheless, sustainable implementation requires standardized curricula, faculty development, competency frameworks, and robust ethical oversight. Future research should focus on longitudinal outcomes and evidence-based educational models that prepare radiology professionals for increasingly AI-integrated healthcare systems.
