Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 18 appraisals
Physics in medicine and biology
Anatomically adaptive feature-wise linear modulation for deep learning-based low-dose CT denoising
Objective.Low-dose computed tomography (LDCT) reduces radiation dose but, introduces heterogeneous noise due to different photon attenuation based on anatomical tissue. Most deep learning techniques assume uniform noise in LDCT and perform equal noise removal across different regions, leading to sub-optimal performance across different tissues. This work aims to design a physics-based framework that explicitly models region-dependent noise characteristics to improve LDCT noise removal.Approach.We propose an anatomically adaptive noise reduction framework. The proposed anatomically adaptive feature-wise linear modulation (FiLM) model consists of a U-Net architecture that integrates two complementary units: the global FiLM unit in the encoder, which modifies features globally based on global image statistics to remove overall noise, and the local FiLM unit in the decoder, which modifies features locally based on the type of anatomical tissue. This dual design enables the modeling of overall image noise characteristics in addition to the removal of local noise associated with each anatomical tissue.Main results.The model performance was evaluated using AAPM-Mayo Clinic LDCT dataset, and the trained model tested on the TCIA dataset. The proposed model outperformed all competing methods. Local noise analysis showed that noise removal was consistent across different anatomical regions, achieving 37.14% in the lung, 48.75% in soft tissue, and 35.18% in bone. Visual results also confirmed significant improvements in noise removal, preservation of structural details, and reduction of non-residual distortions. Furthermore, the model demonstrated its ability to generalize under domain shift.Significance.This work presents a framework for anatomically adapted noise removal by linking feature modification with the physical properties of noise in LDCT through a dual-modulation process for both general and tissue-related noise. The model achieves a balance between noise removal and preservation of anatomical detail, making it a robust approach to LDCT noise removal.
28 July 2026
Read appraisal →Science advances
Interpretable dynamic quantitative vascular morphometry features using SHAP for anti-angiogenic therapy response prediction
Anti-angiogenic therapy benefits vary, with response rates of 40 to 70%, highlighting the need for early biomarkers to identify responders. We developed an automated machine learning framework that uses delta quantitative vascular morphometry features from standard contrast-enhanced CT to evaluate treatment response. This workflow combines automated tumor and vessel segmentation with feature extraction from routine scans for clinical use. Shapley additive explanations (SHAP)-based attributions identify key vascular and clinical features, providing meaningful, imaging-visible evidence aligned with therapy targets beyond traditional radiomics. Using baseline and follow-up CTs from 163 patients with lung cancer, we built three models using fivefold cross-validation, with the delta-merge model achieving high accuracy (area under the receiver operating characteristic curve = 0.842 internally, 0.806 externally). SHAP analysis uncovered an "arterial-dominant, venous-adaptive" pattern, where arterial involvement and venous recovery distinguish responders. This automated workflow and visualization support early, imaging-based response assessment and personalized treatment.
27 July 2026
Read appraisal →Pediatric surgery international
Pediatric colonic diverticulitis: clinical presentation, management, and review of the literature
Colonic diverticulitis is rare in the pediatric population and often presents with nonspecific symptoms that may be misdiagnosed as other causes of acute abdominal pain, such as appendicitis, leading to diagnostic uncertainty and variability in management. We report a case of pediatric colonic diverticulitis and present a systematic review to characterize its clinical presentation, diagnostic challenges and management. Clinical data for the index patient were retrospectively reviewed, and parental consent was obtained. A structured literature search was performed using PubMed, Scopus, and Embase. Studies reporting patients younger than 18 years were included. A 16-year-old male presented with right lower quadrant pain. CT suggested cecal diverticulitis. Worsening symptoms prompted diagnostic laparoscopy, revealing an inflamed diverticulum adjacent to a normal appendix. The postoperative course was uneventful. The review identified 27 studies including 101 patients. Median age was 14 years (IQR 12-16; range 3-17), with 57% male preponderance. Most cases involved a solitary right-sided diverticulum. CT correctly identified diverticulitis in 85%. Complicated disease occurred in 17%. In conclusion, in the pediatric age diverticulitis often arises from a solitary cecal diverticulum and can closely mimic appendicitis. Awareness may improve diagnosis, and uncomplicated cases can be treated conservatively with careful follow-up and appropriate clinical monitoring.
21 July 2026
Read appraisal →NeuroImage
Rapid multi-parametric quantitative MRI via deep learning-based synthetic-to-real reconstruction and 3D SSFP-MOLED imaging
Multi-parametric quantitative magnetic resonance imaging (mqMRI) holds significant clinical potential through multi-parametric tissue characterization, yet its adoption is hindered by prolonged scan time and sensitivity to non-ideal signal conditions, especially in high-resolution whole-brain protocols. To address these challenges, we propose a novel signal encoding method integrating phase-modulated three dimensional steady-state free precession with multiple overlapping-echo detachment (3D SSFP-MOLED). This method simultaneously encodes six physiological parameters (M0, T1, T2, T2*, B1+, ΔB0) into k-space by controlling overlapping echo detachment in signal acquisition. A physics-constrained synthetic data pipeline was developed to simulate MR signal evolutions with realistic field variations (ΔB0, B1+ inhomogeneities), enabling robust training of network for real-time parameter mapping. Whole-brain parametric maps (1×1×2 mm³ resolution) can be delivered within 3 minutes with only 2x parallel acquisition acceleration. Validation was performed on phantom, healthy volunteers, and clinical cases with tumors/hemorrhage. Experimental results show that our method can achieve rapid multi-parametric quantitation with high accuracy and reproducibility. By synergizing adaptive signal encoding, physics-informed synthetic training, and reproducible deep learning reconstruction, this work establishes a new paradigm for efficient and reliable mqMRI in clinical signal processing applications.
17 July 2026
Read appraisal →European respiratory review : an official journal of the European Respiratory Society
Prediction of PD-L1 expression in nonsmall cell lung cancer using artificial intelligence models based on radiomics: a systematic review and meta-analysis
BACKGROUND: Since 1992, the incidence of lung cancer has shown an overall downward trend, with survival rates significantly improving in recent years due to the application of immune checkpoint inhibitors (ICIs). Assessing programmed death ligand 1 (PD-L1) expression status in lung cancer patients is crucial for predicting the efficacy of ICIs and formulating treatment plans. Therefore, there is an urgent need to assess PD-L1 expression status in lung cancer patients, particularly those with nonsmall cell lung cancer (NSCLC). While numerous studies have demonstrated promising results using artificial intelligence (AI) models based on radiomics, current understanding in this area remains incomplete. METHOD: We conducted a search of medical databases and screened studies that meet the criteria. We performed feature extraction and data extraction from these studies, analysed them using R software, and generated forest plots. We conducted subgroup analyses based on imaging modality and algorithm model. RESULTS: A total of 35 studies were included in the analysis. For the Tumor Proportion Score (TPS)1 group, the pooled area under the curve (AUC) for the validation set was 0.782. For the TPS50 group, the pooled AUC for the validation set was 0.798. Subgroup analysis revealed that machine learning performed better when predicting TPS50, while deep learning was superior for predicting TPS1. When predicting TPS, prediction models constructed using 2-fluoro-2-deoxy-d-glucose (18F) positron emission tomography/computed tomography (CT) and contrast-enhanced CT outperformed those based on noncontrast CT. Conversely, when predicting a TPS, the noncontrast-CT-based prediction model significantly outperformed the other two groups. CONCLUSIONS: Imaging modality and algorithm model are key factors influencing AI prediction models for PD-L1 expression in NSCLC patients.
10 July 2026
Read appraisal →Journal of gastrointestinal cancer
Radiomics Models as Tools for Predicting Genetic Mutations in Colorectal Cancer: A Systematic Review and Meta-Analysis
BACKGROUND: The evaluation of genetic mutations is crucial for personalized therapy in colorectal cancer (CRC), but the invasive tissue biopsy is subject to sampling bias and other complications. Radiomics has emerged as a non-invasive tool to predict these mutations from standard medical images. In this systematic review and meta-analysis, we aimed to evaluate the diagnostic accuracy and methodological quality of radiomics models for predicting key genetic mutations in CRC. METHODS: A comprehensive search of PubMed, Scopus, Web of Science, and Embase was conducted in accordance with PRISMA guidelines. Studies evaluating radiomics models for predicting genetic mutations in CRC patients using pre-operative CT, MRI, or PET/CT were included. A meta-analysis of diagnostic accuracy was performed to calculate the pooled sensitivity, specificity. Methodological quality was assessed using the Radiomics Quality Score (RQS) and QUADAS-2 tools. RESULTS: Sixteen studies were included in the quantitative analysis. The pooled sensitivity and specificity were 0.75 (95% CI, 0.67-0.81) and 0.78 (95% CI, 0.70-0.85), respectively, with an overall AUC of 0.79. Subgroup analyses revealed that radio-clinical models integrating both clinical and radiomics features achieved superior sensitivity compared to models with only radiological input. However, the overall methodological quality of the included studies was low, with a mean RQS of 45%. CONCLUSION: Conventional radiomics models demonstrate promising results for the non-invasive prediction of genetic mutations in CRC, with sensitivity enhanced by the integration of clinical data. Despite this potential, significant methodological shortcomings and heterogeneity across studies highlight the need for standardized protocols and large-scale, prospective validation before these models can be translated into routine clinical practice.
10 July 2026
Read appraisal →PloS one
Deep learning based two-way feature depiction model for brain tumor detection
Brain tumors are one of the most fatal disorders that cause one of the highest mortalities in the world. Gliomas are the most common primary brain tumors originating from glial cells in the central nervous system. Traditionally, a tissue sample is extracted and examined for its genetic and characteristic properties. This method is invasive, painful, and takes a longer period to produce results. Various automatic Deep learning (DL) based schemes have been presented for the brain glioma detection, but they lack due to poor explainability, lower generalization, poor feature depiction, class imbalance problem and lower detection rate. This paper presents a deep learning based brain tumor detection using two way feature depiction model (TWFDM) that combines the 2D-Deep Convolution Neural Network (DCNN) and 1D-DCNN. The 2D-DCNN accepts the raw MRI images and the 1D-DCNN accepts the handcrafted local binary pattern (LBP), gray level cooccurrence matrix (GLCM), and Histogram of Oriented Gradient (HOG) features. Furthermore, improved particle swarm optimization (IPSO) is used for feature selection to minimize the computational complexity of the TWFDM system. The proposed TWFDM achieves an overall accuracy of 96.25%, a recall of 96.34%, a precision of 96.31%, and an F1-score of 96.32% on the Brain MRI dataset for four-class classification, representing an important improvement over traditional techniques.
3 July 2026
Read appraisal →Dento maxillo facial radiology
JADE: jawbone lesion diagnosis and decision supporting system
OBJECTIVES: To develop and evaluate JADE, a proof-of-concept retrieval-augmented generation (RAG) diagnostic assistive system, designed to enhance large language model (LLM) reasoning for jawbone lesion assessment. This study examined whether RAG improves diagnostic accuracy and stability compared with standalone LLMs and ORAD, a supervised learning-based system. METHODS: JADE was developed as a cloud-based RAG system integrating an expert-curated oral radiology database embedded using text-embedding-3-large and indexed in Qdrant Cloud. Structured clinical inputs were encoded as prioritized vector queries. Hybrid semantic and keyword-based retrieval appended relevant evidence to prompts for differential diagnosis generation across LLM backbones. Performance was evaluated in 25 validation cases and compared with GPT-5, Claude Sonnet 4.5, DeepSeek-R1, Gemini 2.5 Flash, their RAG configurations, and ORAD. Diagnostic accuracy was analyzed using Cochran's Q test, with post-hoc McNemar's tests and Bonferroni correction. Intra-model stability and response time were assessed. RESULTS: RAG-GPT-5 achieved the highest diagnostic accuracy (20/25), followed by RAG-Claude Sonnet 4.5 (18/25), RAG-DeepSeek R1 (17/25), and RAG-Gemini 2.5 Flash (15/25). Standalone models achieved 9-13/25 correct diagnoses, ORAD achieved 17/25. No significant differences were observed among standalone models or RAG-based models and ORAD. A significant improvement was observed for GPT-5 when integrated with RAG (P = .002). RAG configurations showed higher intra-model stability, with RAG-GPT-5 achieving mean stability of 0.90 ± 0.11. Mean response times ranged from 3 to 10 s. CONCLUSIONS: JADE improved diagnostic accuracy and stability compared with standalone LLMs, underscoring value of RAG reasoning in jawbone lesion assessment and marking the first RAG application in dentomaxillofacial radiology.
3 July 2026
Read appraisal →Physics in medicine and biology
Hybrid learning: a combination of self-supervised and supervised learning for joint MRI reconstruction and denoising in low-field MRI
Objective.Deep learning has demonstrated strong potential for magnetic resonance imaging (MRI) reconstruction. However, conventional supervised learning requires high-quality, high-signal-to-noise-ratio (SNR) reference data for network training, which are often difficult or impossible to obtain, particularly in low-field MRI. Self-supervised learning (SSL) eliminates the need for reference training data but may suffer from degraded performance under low-SNR conditions. To address these limitations, we propose hybrid learning, a new training framework that integrates self-supervised and supervised learning for joint MRI reconstruction and denoising when only low-SNR training data are available.Approach.Hybrid learning is implemented in two sequential stages. In the first stage, SSL is applied to fully sampled low-SNR data to generate higher-quality pseudo-references. In the second stage, these pseudo-references are then used as targets for supervised learning to reconstruct and denoise undersampled, noisy data. The proposed method was evaluated in four experiments using simulated and real noisy MRI data of the breast, lung, and brain across different field strengths (0.3 T to 3 T), sampling trajectories (Cartesian, spiral, and radial), noise levels, and undersampling ratios.Main Results.Hybrid learning consistently improved reconstruction quality relative to both supervised and self-supervised baselines under different acceleration rates, noise levels, and sampling patterns in all experiments. Compared with standard supervised learning using noisy references, it achieved up to 167.70% higher structural similarity index measure (SSIM), 95.41% lower normalized mean squared error (NMSE), and 90.70% lower high-frequency error norm (HFEN). Compared with standard SSL, it achieved up to 23.88% higher SSIM, 60.85% lower NMSE, and 49.13% lower HFEN.Significance.Hybrid learning enables improved MRI reconstruction under low-SNR imaging conditions by jointly addressing noise and undersampling. It provides a practical solution for robust deep learning-based reconstruction and is particularly well suited for applications such as low-field MRI, where image quality is limited by reduced SNR.
26 June 2026
Read appraisal →Physics in medicine and biology
Domain-specific adaptation for MR image synthesis with text-guided diffusion
Objective.Deep learning in medical imaging is severely constrained by data scarcity. Data synthesis offers a promising solution, but existing generative models have difficulty in restoring pathological texture features when trained on small-scale datasets. To address this, we propose a domain-specific, partition-based parallel text-guided latent diffusion model (LDM) for medical image synthesis.Approach.Each LDM operates on a defined image domain and is fine-tuned to reproduce specific texture characteristics. Diseased regions are identified from segmentation masks, while healthy regions are further subdivided using Voronoi-grayscale adaptation, enabling localized texture preservation. The fine-tuned LDMs independently synthesize corresponding image partitions, which are subsequently merged and denoised to form complete synthetic images with paired segmentation masks.Main results.We evaluated the approach on glioma MRI data, achieving a Fréchet Inception Distance of 13.65, demonstrating high perceptual realism. Texture fidelity was further supported by SSIM of 0.9674 and radiomic feature distribution analysis, both confirming close alignment between real and synthetic images. In a blinded visual Turing test, three radiologists achieved an average sensitivity of only 25.5% when identifying synthetic MRI slices, resulting in a 74.5% deception rate, and 41% of the synthetic samples were universally misclassified as real by all experts. In downstream experiments, U-Net trained on the synthetic-augmented dataset improved DSC by 14% on average.Significance.These results demonstrate that the proposed domain-specific adaptation framework can generate perceptually plausible, structure-preserving synthetic MRI slices in data-constrained environments, while improving downstream segmentation performance. The method therefore shows potential as an augmentation-oriented tool for AI model development, clinical teaching, assisted diagnosis, and rare-disease research.
25 June 2026
Read appraisal →PloS one
Artificial intelligence in the diagnosis of deep vein thrombosis: A scoping review
Deep vein thrombosis (DVT) is the formation of thrombi in the deep venous system, most often in the lower extremities. Although usually not life-threatening, DVT requires timely diagnosis to prevent complications such as pulmonary embolism and post-thrombotic syndrome. The growing demand for image interpretation has generated interest in applying artificial intelligence (AI) to automated DVT detection. This scoping review analyzes the performance of artificial intelligence in diagnosing DVT using computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound (US). We conducted a search across seven databases from inception to May 2025 using terms related to deep vein thrombosis, artificial intelligence, and machine learning. Eligible studies were limited to those evaluating DVT diagnosis using CT, MRI, or ultrasound. Two independent reviewers selected eligible studies, and quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). Eleven studies published between 2021 and 2025 met the inclusion criteria. Some of the AI algorithms included RetinaNet, Deep R-Belief Neural Networks, and Sooty Tern Optimization. US-based models were the most studied algorithms, with sensitivities and specificities ranging from 68 to 100% and 70-100%, respectively. The MRI-based model achieved sensitivities, specificities, and accuracies of 95% to 97%. One CT-based model demonstrated a sensitivity of 83%. Studies evaluated across multiple imaging datasets showed high sensitivities, specificities, and precision of 96% or higher. Future research should prioritize multicenter validation and integration of clinical factors. In addition, explainable frameworks capable of integrating multiple imaging datasets must be developed with attention to workflow efficiency and cost-effectiveness to support clinical translation. The results indicate that AI is best situated as a supplementary tool rather than a replacement for expert interpretation in DVT diagnosis.
23 June 2026
Read appraisal →Journal of neuro-oncology
Clinical utility of liquid biopsy in distinguishing true progression from radiation necrosis and pseudoprogression in malignant brain tumor.
BACKGROUND: Distinguishing true tumor progression from pseudoprogression (PsP) and radiation necrosis (RN) following radiotherapy or radiosurgery for intracranial tumors remains a significant clinical challenge. Conventional imaging modalities, including MRI and PET, often lack sufficient diagnostic accuracy, while histopathological confirmation, although definitive, is invasive and not always feasible. Liquid biopsy (LB) has emerged as a promising noninvasive diagnostic approach. METHODS: A systematic review was conducted across PubMed, Embase, Scopus, and Cochrane databases. A total of 964 records were identified, including PubMed (n = 322), Embase (n = 280), Scopus (n = 297), and Cochrane (n = 65), along with three clinical trials identified from trial registers. After removal of 157 duplicates using Covidence, 807 studies were screened by title and abstract. Of these, 749 were excluded, and 58 articles underwent full-text review. Ultimately, 11 studies met the inclusion criteria and were included in the analysis. RESULTS: A diverse range of LB-based biomarkers has been investigated, including immune cell-based markers (e.g., HLA-DRneg/low and VNN2+ CD14+ monocytic myeloid-derived suppressor cells), immune-related proteins (e.g., CXCL11 and MUC-16), circulating tumor cells, cell-free DNA, mitochondrial DNA, B1-SINE elements, microvesicles, and RNA analytes. Plasma was the most used biofluid, with limited evaluation of urine and cerebrospinal fluid. Analytical techniques varied widely, including flow cytometry, PCR-based assays, immunostaining-FISH, and multiplex protein platforms. Several biomarkers showed statistically significant differences among RN, PsP, and tumor recurrence; however, these findings were largely derived from small, heterogeneous cohorts with inconsistent reporting of diagnostic performance metrics. Composite indices, such as the DR-VNN2 index (DVI) and the Necrosis Prediction Index (NPI), showed potential for improving diagnostic differentiation. CONCLUSIONS: LB-based biomarkers show promise for differentiating RN and PsP from tumor recurrence; however, current evidence is limited by small sample sizes, methodological heterogeneity, and lack of standardized diagnostic criteria. Larger prospective studies and validation of composite biomarker models are required to establish their clinical utility. CLINICAL TRIAL NUMBER: Not Applicable.
26 May 2026
Read appraisal →Journal of radiological protection : official journal of the Society for Radiological Protection
The integration of artificial intelligence in radiation medical physics: insights from an international survey with regional variability
Artificial intelligence (AI) is considered to be a leading technology in radiation medical physics, which has the potential for improving efficiency and precision in imaging, radiotherapy, and nuclear medicine. Nonetheless, its application in the clinical setting is hindered by education, regulation, and ethical issues. To evaluate the views of medical physicists on AI adoption, potential issues, perceived advantages, ethical issues, and training needs. A cross-sectional survey in the form of an online questionnaire was distributed internationally to practicing medical physicists, representing specialties in radiation therapy, diagnostic imaging, and nuclear medicine. The instrument was used to capture demographic variables and AI familiarity, perceived benefits and barriers, ethical issues, training preferences, and future expectations. Descriptive statistics, chi-square tests, and binary logistic regression were performed. Open-ended responses were analyzed using thematic analysis. Most respondents were moderately familiar with AI, but nearly half had not incorporated AI into their practice despite being interested. Adoption patterns differed significantly across geographic regions, with higher usage reported in developed countries. AI familiarity was strongly associated with adoption. The most prominent benefits noted were dose calculation accuracy, patient safety, and workflow efficiency. In contrast, the main obstacles were inadequate training, high implementation costs, absence of standardized protocols, and low availability of advanced equipment. Ethical apprehensions focused on accountability and reduced human oversight. Preferred educational strategies included hands-on workshops and on-the-job training. Regional analysis and modeling revealed variability in AI adoption and identified key predictors, including AI familiarity and specialization. Results indicate that AI is considered an important technology in radiation medical physics. However, implementation was influenced by regional and professional differences. These findings highlight the necessity to expand AI education, create standard training programs, improve infrastructure, and provide clearer governance frameworks to support safe and responsible AI integration.
22 May 2026
Read appraisal →Physics in medicine and biology
A novel deep-learning approach for monitoring gastrointestinal air variation during radiotherapy in young patients using radiographs
Objective.To develop deep-learning (DL) predictive models of internal gastrointestinal (GI) gas from radiographs, a surrogate for dose degradation in abdominal proton beam therapy (PBT).Approach.DL architectures (convolutional neural networks and U-Nets) were trained to predict GI air local path length and/or global volume from radiographs. Frontal/lateral radiographs with paired volume and path length labels were simulated from CTs and GI gas segmentations using the Tomographic Iterative GPU-based Reconstruction (TIGRE) toolbox. CTs from The Cancer Imaging Archive Paediatric-CT-SEG dataset (N= 349, 0-16 year, North America) were used for model development and internal evaluation; variation in GI filling, organ motion and body size were simulated for augmentation. External evaluation utilised data from two international radiotherapy institutions: one consisting mostly of children and young people (CYP) (N= 21, 2-19 year, Europe); the other predominantly of teenagers and young adults (TYAs) (N= 14, 10-24 year, Asia). Finally, real-life use case was demonstrated on two subjects treated with PBT.Main results.The optimal configuration was the dual U-Net, accurately predicting both volume (mean absolute error (MAE) = 24.0 ± 32.9 ml) and path length (MAE = 0.80 ± 0.59 mm) on the held-out Paediatric-CT-SEG dataset, while also generalising well to external radiotherapy datasets (volume MAE = {18.9 ± 22.4, 34.6 ± 35.2} ml, and path length MAE = {0.52 ± 0.38, 0.67 ± 0.55} mm on the CYP and TYA datasets, respectively). We found a lower Dice similarity coefficient for gas segmentations in TYA vs Paediatric-CT-SEG and CYP (0.56 ± 0.24 vs 0.87 ± 0.09 and 0.84 ± 0.16, respectively), reflecting cohort-specific gas morphology differences. When applied longitudinally, the network detected both monotonic and realistic intrasubject GI gas variation. Proof-of-concept evaluation on clinical radiographs showed promising performance (volume MAE = 29.5 ± 25.9 ml, and path length MAE = 1.10 ± 0.75 mm).Significance.DL-powered radiographs quantified internal GI gas changes with promising performance and generalisability to diverse ages and demographics. This was the first step towards a novel personalised image-guided radiotherapy traffic-light workflow for abdominal PBT with the goal of reducing the need for CBCT to required fractions.
14 May 2026
Read appraisal →Journal of neuroradiology = Journal de neuroradiologie
Automated estimation of perivascular space and brain morphometry from deep learning-reconstructed three-dimensional T1-weighted MRI: comparison with the conventional technique
PURPOSE: To evaluate the utility of deep learning-based reconstruction (DLR) three-dimensional T1-weighted imaging (T1-WI) in improving fine structural delineation through the automated estimation of brain morphometry and perivascular space (PVS). MATERIALS AND METHODS: In this retrospective cohort study, participants who underwent T1-WI between August 2021 and June 2022 were included. Participants were categorized into normal and disease groups based on the presence of visible cerebral lesions. Cortical thicknesses and cerebral volumes from DLR T1-WI were estimated using FreeSurfer and compared to those derived from conventional T1-WI. Additionally, participants were grouped by MRI-visible PVS severity according to the number of enlarged PVS in centrum semiovale and basal ganglia. PVS volumes from DLR and conventional T1-WI were compared across these groups. Statistical comparisons were performed using a paired independent t-test with Benjamin-Hochberg adjustment. RESULTS: 240 participants (mean age: 40.8 ± 17 years) were included. In the normal group (n = 144), cortical thicknesses and volumes were significantly greater in DLR T1-WI compared to conventional T1-WI (P < 0.001). This trend was also observed in the disease group (n = 96) and in external datasets (n = 63). Across all three PVS severity groups (mild, n = 111; moderate, n = 71; severe, n = 58), DLR T1-WI significantly reduced PVS volumes and variability compared to conventional T1-WI. DLR T1-WI demonstrated PVS volumes (mild: 738.4 mm³ vs. 1408.1 mm³; moderate: 919.0 mm³ vs. 1895.8 mm³; severe: 1328.5 mm³ vs. 2711.7 mm³; all P < 0.001) and variability (mild: 224.4 mm3 vs. 420.9 mm³; moderate: 229.7 mm3 vs. 462.9 mm³; severe: 382.4 mm3 vs. 653.6 mm³) that were consistently smaller. CONCLUSION: DLR T1-WI yielded cerebral cortical thickness and volume measurements that were more consistent with expected anatomical values than those from conventional T1-WI.PVS volumes measured from DLR T1-WI exhibited less variability and better concordance with radiologists' visual assessments.
5 May 2026
Read appraisal →Journal of clinical ultrasound : JCU
Deep Learning Radiomics Model Based on Ultrasound Images Predicts Myometrial Infiltration of Endometrial Cancer
OBJECTIVE: This study aims to develop and validate a deep learning radiomics (DLR) model based on ultrasound images for non-invasively distinguishing the myometrial infiltration (MI) of patients with endometrial cancer (EC). METHODS: We retrospectively collected 310 patients with EC who underwent surgical resection from center 1 from September 2017 to January 2025, forming a training group and a validation group. An external testing group was comprised of 80 patients from center 2. We extracted deep learning (DL) features and radiomics features from ultrasound images, establishing a DLR model after dimensionality reduction. The receiver operating characteristic analysis was used to evaluate the practicality of the proposed model. The utility of the proposed model was evaluated using receiver operating characteristic, calibration, and decision curve analysis. RESULTS: A total of 390 EC patients were included in the study. In the validation set, the AUC of the radiomics model was 0.874 (0.774-0.949), and in the testing group, the AUC of the DL model was 0.844 (0.699-0.957). The DLR model was superior to radiomics models and DL models in interpreting images, and also outperformed the diagnostic performance of advanced ultrasound physicians. CONCLUSION: The DLR model based on ultrasound images can accurately and non-invasively distinguish the MI depth of EC patients, assisting doctors in formulating more favorable treatment plans for patients.
5 May 2026
Read appraisal →JMIR human factors
Perceptions of Health Care Professionals on the Integration and Use of AI in Clinical Cancer Care: Interview Study
BACKGROUND: Artificial intelligence (AI) is increasingly recognized for its potential to transform cancer care. However, much of the existing evidence of its efficacy comes from controlled settings. There remains a need to complement this knowledge with insights into how AI tools are perceived and used in real-world clinical settings, as well as how their use impacts clinical practice. OBJECTIVE: This study aimed to explore key factors influencing clinicians' acceptance of AI tools and examine how AI adoption and use impact clinical workflows in cancer care. METHODS: We used purposive sampling for recruiting oncology-related health care professionals and collected data using web-based semistructured interviews to gather their perceptions. Data were thematically analyzed and interpreted through the lenses of sociotechnical systems theory and the Unified Theory of Acceptance and Use of Technology. RESULTS: Participants largely accept and perceive AI tools as beneficial to clinical practice. Unified Theory of Acceptance and Use of Technology constructs were reflected in our data as determinants of intention to adopt AI tools. Trust appears as an influential factor in shaping attitudes toward AI tools. Acceptance is found to both precede AI tool use and to grow following successful integration. The use of AI tools is perceived to yield operational benefits, such as reduced workload and time savings, and clinical benefits, such as increased diagnostic reliability and reduced patient recall. Minimal disruption to clinical workflows following integration of AI tools was reported for some cancer screening applications and organ-at-risk segmentation, whereas greater disruption was anticipated for 3D cancer screening. Although accountability and lack of explainability are highlighted in literature as barriers to AI adoption, participants do not view these as significant obstacles in image-based diagnostic contexts. Additionally, negative impacts, such as overreliance on AI and reduced critical review of AI results, arise in association with the use of AI tools. CONCLUSIONS: Participants perceive AI tools to deliver benefits to clinical cancer care. However, their adoption relies on their alignment with clinical needs and seamless integration into clinical workflows. To encourage clinician acceptance, the identified concerns must be addressed. Future work should focus on training programs, co-design with clinicians, and exploration of mitigation strategies for emerging adverse effects, such as automation bias and potential skill erosion.
22 Apr 2026
Read appraisal →AJNR. American journal of neuroradiology
Temporomandibular Joint Lesions with Intracranial Extension: Illustrative Cases from a Systematic Review of the Literature and Our Institution
BACKGROUND: Intracranially extending temporomandibular joint (TMJ) lesions may be radiologically misinterpreted as primary intracranial or skull base pathologies, leading to diagnostic delays or inappropriate management. PURPOSE: This systematic review aimed to characterize the clinical and imaging features of such TMJ lesions and evaluate the impact of radiologic misclassification. We also aimed to develop a diagnostic framework for when to consider an intracranially extending TMJ lesion, based on clinical and radiologic features. DATA SOURCES: A comprehensive search of MEDLINE, Scopus, and EMBASE, conducted in accordance with PRISMA guidelines, yielded 2255 records. STUDY SELECTION: After screening with predetermined inclusion and exclusion criteria, 128 studies involving 152 patients were included in the final analysis. DATA ANALYSIS: Statistical analyses were performed using STATA software. We also identified 3 patient cases through our institutional neuroradiology practice who were clinically and radiologically assessed for intracranially extending TMJ lesions. DATA SYNTHESIS: Patients had symptoms for an average of 34 months before diagnosis (47% women, mean age 50 years). The most common pathologies were pigmented villonodular synovitis/tenosynovial giant-cell tumor (43%) and synovial chondromatosis (24%). Neurologic symptoms were reported in 48% of cases, most frequently hearing loss (70%). Nearly one-third (33%) of cases with an imaging differential did not list a TMJ pathology (18/55). In cases with accurate imaging diagnosis, 90% had both CT and MRI performed. Most lesions were nonenhancing (CT 83%, MRI 75%) and demonstrated no adjacent brain edema (96%). In 2 cases, a TMJ ganglion cyst and pseudogout were misdiagnosed as intracranial tumors, resulting in unnecessary intervention, including repeat craniotomy and radiotherapy. LIMITATIONS: There were inherent biases of case report literature, including variability in the reporting of the imaging and clinical features, management, and follow-up. CONCLUSIONS: TMJ lesions with intracranial extension often present with nonspecific symptoms and can mimic extra-axial tumors, leading to misdiagnosis on imaging. Recognition of hallmark imaging features, including lack of parenchymal invasion and distinct imaging patterns, may help improve radiologic accuracy and prevent overtreatment. We propose a diagnostic framework outlining when to suspect intracranially extending TMJ lesions based on clinical and imaging features, and how to avoid common diagnostic pitfalls.
4 Apr 2026
Read appraisal →