Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 8 appraisals
Neural networks : the official journal of the International Neural Network Society
MFBD: Model-free backdoor defense based on vision-language pre-trained models
Backdoor attacks pose serious security threats to deep neural networks (DNNs), causing networks to output predetermined target labels when specific trigger patterns are embedded in input images. While existing backdoor defense methods utilize additional clean data or victim network models, their effectiveness in defending against various trigger injection methods is limited. We find that vision-language models, such as BLIP-2 and CLIP, are largely unaffected by backdoor triggers in poisoned images. Based on this observation, we propose MFBD, an effective model-free backdoor defense method leveraging the semantic consistency of dual vision-language descriptors. Specifically, we employ BLIP-2 and CLIP to generate descriptions of an image and its saliency masked image, respectively. We then embed both the descriptions and the image labels into a semantic space using SBERT. To distinguish clean and poisoned images, we compute the cosine similarity between their semantic representations. The image exhibiting a low cosine similarity value is classified as a poisoned image. We evaluate a suiute of backdoor defense methods against various attack methods on three datasets. Experimental results demonstrate that our method can effectively mitigate various complex backdoor threats while maintaining high performance across different datasets.
2 Aug 2026
Read appraisal →Scientific reports
Safety design guidelines for clinician-AI interaction in computer-aided diagnosis systems using system-theoretic framework with explainability validation
Ensuring the safety of Artificial Intelligence-enabled Computer-Aided Diagnosis systems is critical because diagnostic errors can have serious consequences for patient care. However, existing regulatory and risk management frameworks often do not sufficiently address the complex socio-technical interactions between clinicians and Artificial Intelligent systems, leaving key human-centered safety challenges underexplored. This paper presents a systematic and human-centered approach to deriving safety design guidelines for clinician-Artificial Intelligence interaction in Computer-Aided Diagnosis systems using System-Theoretic Process Analysis. Through this analysis, we identify critical hazards associated with clinician-Artificial Intelligence collaboration, including automation bias on system recommendations, and misinterpretation of explanations. Based on the identified unsafe control actions, we formulate a set of actionable and traceable safety design guidelines that promote transparency, coherent explanations, and calibrated trust in Artificial Intelligence-assisted decision-making. To bridge safety analysis and system design, the proposed guidelines are operationalized within a Computer-Aided Diagnosis interaction framework. The framework includes a safety-oriented Graphical User Interface that integrates multiple explanation methods and interactive mechanisms to promote clinician engagement. Furthermore, we introduce a safety-oriented evaluation approach that uses consistency across multiple explanation methods as a quantitative indicator of potentially unreliable or ambiguous explanations. By linking System-Theoretic Process Analysis, interaction design, and explainability evaluation, this work provides a unified and reusable framework for improving the safety and reliability of Artificial Intelligence-driven Computer-Aided Diagnosis systems.
31 July 2026
Read appraisal →Journal of breast imaging
Contrast-Enhanced Mammography for Breast Cancer Screening: A Systematic Review and Meta-Analysis
OBJECTIVE: To evaluate pathway-specific cancer detection and summarise screening outcomes (recall/biopsy/PPV3) for contrast-enhanced mammography when reported, and to compare them with those for low-energy mammography and breast MRI. METHODS: This systematic review and meta-analysis followed PRISMA guidelines. PubMed (MEDLINE), Embase, Scopus, and Web of Science were searched from January 1, 2011, through August 31, 2025, for studies evaluating contrast-enhanced mammography in screening or surveillance populations (including supplemental/add-on screening and surveillance of patients with a personal history of breast cancer). Diagnostic, recall, and problem-solving studies were excluded. Screening pathways were categorized as: (1) direct comparison with low-energy mammography, (2) direct comparison with MRI, and (3) add-on CEM to mammography-based screening. Random-effects meta-analyses were performed when designs were comparable. Comparisons between CEM and MRI were summarized qualitatively because of heterogeneity and limited cancer events across studies. Single-arm screening studies were summarized descriptively. Risk of bias was assessed using Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2). RESULTS: Fourteen studies met eligibility criteria for qualitative synthesis; seven screening and surveillance studies contributed to the meta-analysis. Compared with low-energy mammography, CEM detected 9.3 additional cancers per 1,000 women screened (risk difference [RD], +9.3 per 1,000; 95% CI, 4.0-14.6; P < 0.01; I2 = 0%). Three studies directly compared CEM with MRI in screening settings and were summarized qualitatively because of heterogeneity and limited cancer events. CONCLUSION: In screening and surveillance populations, contrast-enhanced mammography increases cancer detection compared with low-energy mammography. Evidence from available comparative studies suggests that cancer detection with CEM may be broadly comparable to MRI in selected screening contexts; however, the limited number of studies and heterogeneous designs preclude definitive comparative conclusions. These pathway-specific estimates may inform clinical implementation and the design of prospective screening trials.
30 July 2026
Read appraisal →Journal of medical Internet research
Evaluation Frameworks for Clinical AI Incorporating Validation Strategies, Real-World Applicability, and Ethical Principles: Scoping Review
BACKGROUND: AI shows substantial potential in health care; however, the absence of standardized evaluation frameworks limits its safe and effective clinical implementation because of inconsistent validation requirements and fragmented ethical principles. Existing guidelines vary in structure, methodological rigor, and ethical integration, creating uncertainty. OBJECTIVE: This study aimed to systematically map, characterize, and critically analyze existing evaluation frameworks for clinical AI, focusing on three core dimensions: methodological rigor, validation strategies (internal validation, including reporting of technical and clinical performance; external validation, including real-world applicability), and alignment with the United Nations Educational, Scientific and Cultural Organization (UNESCO) AI ethical considerations. METHODS: A scoping review was conducted following PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. Six databases (PubMed, Embase, BVS, EBSCOhost, ProQuest, and Sage) and the Enhancing the Quality and Transparency of Health Research Network were searched without language or date restrictions up to February 2026. Eligible documents included peer-reviewed papers, gray literature, and organizational guidelines describing evaluation or reporting frameworks for clinical AI. Editorials, commentaries, and conference abstracts lacking a clearly defined evaluative framework or clinical applicability were excluded. Two reviewers independently screened records and extracted data. Data were extracted across three domains: (1) general characteristics, (2) methodological rigor and validation parameters, and (3) ethical integration and were synthesized using a dot plot-based gap map. Ethical adherence was assessed using a 10-domain UNESCO-based scoring matrix. No formal risk-of-bias assessment was conducted, consistent with scoping review methodology. RESULTS: From 3363 records, 46 frameworks met the inclusion criteria. Mapping revealed a rapidly expanding but fragmented landscape. Most frameworks targeted investigational use (88%), with limited focus on clinical applicability. Frameworks varied in structure, methodology, and scope, with a predominance of reporting guidelines and few validated tools. Most (63%) were developed through multi-institutional collaborations, and 32.6% incorporated transdisciplinary participation. Only 31.8% reported technical metrics (commonly area under the curve, sensitivity, and specificity), and 15.9% provided clinical indicators (eg, predictive values or calibration). Only 11.4% achieved methodological rigor, incorporating validation aligned with intended use, while most relied on partial validation strategies, highlighting a gap between model development and clinical evaluation. Ethical integration was heterogeneous: only 5 frameworks achieved high compliance (≥80%), whereas 4 scored <10%. The most frequently addressed UNESCO principles were awareness and education (71.1%) and transparency and explainability (70%), while human oversight (24.4%) and adaptive governance (33.3%) were least represented. Findings indicate a misalignment between framework design, validation requirements, and clinical implementation. CONCLUSIONS: Evaluation frameworks for clinical AI remain heterogeneous and oriented toward investigational contexts. Critical gaps persist in methodological rigor, validation aligned with intended use, and fragmented ethical coverage. These findings highlight the need for standardized, robust, and ethically grounded frameworks to enable safe, reliable, and scalable integration of AI into clinical practice.
24 July 2026
Read appraisal →European journal of trauma and emergency surgery : official publication of the European Trauma Society
The association between pancreatic enzyme serum concentrations and blunt pancreatic trauma: a systematic review.
PURPOSE: Blunt pancreatic trauma is uncommon. When suspected, serum amylase and lipase are measured, but their diagnostic value remains uncertain. This systematic review evaluated the diagnostic performance of serum amylase and serum lipase in detecting blunt pancreatic trauma in adults. METHODS: MEDLINE, Embase, Scopus and Cochrane Central were searched for studies evaluating serum amylase and/or lipase in adults with blunt pancreatic trauma. Observational studies reporting diagnostic accuracy or clinical outcomes were included. Two reviewers independently screened studies and assessed risk of bias using the Newcastle-Ottawa Scale. RESULTS: Four observational studies met inclusion criteria. A total of 4937 patients with blunt abdominal trauma, including 150 with pancreatic injury were investigated. Two studies evaluated serum lipase alone and two compared amylase and lipase. Pancreatic injuries were graded according to the American or Japanese Association for the Surgery of Trauma. Early serum amylase measurement (< 90 min post-injury) demonstrated low sensitivity. In contrast, testing performed more than 3 h after injury has shown improved diagnostic accuracy with one study suggesting combined serum amylase and lipase testing can achieve a sensitivity of 85% and specificity of 100%. CONCLUSION: Early serum pancreatic enzymes measurement does not appear to aid initial management decisions following blunt abdominal trauma. Delayed testing (> 3-6 h), particularly using lipase or combined lipase and amylase showed improved diagnostic performance. Prospective multicentre validation is needed to define their role in guiding further imaging after equivocal or negative initial CT.
22 July 2026
Read appraisal →The Journal of international medical research
Diagnostic performance of machine learning-based radiomics models for predicting epidermal growth factor receptor mutation status in lung adenocarcinoma in Chinese patients: A systematic review and meta-analysis
ObjectiveThis systematic review and meta-analysis evaluates the diagnostic performance of machine learning-based radiomics models for predicting epidermal growth factor receptor mutation status in Chinese patients with lung adenocarcinoma.MethodsFollowing the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines and prospectively registered in the International Prospective Register of Systematic Reviews (CRD420251273027), a systematic search of PubMed, Embase, Web of Science, the Cochrane Library, Scopus, China National Knowledge Infrastructure, Wanfang, VIP, and Chinese Biomedical Literature Database was conducted from inception to 31 October 2025. Two reviewers independently screened studies, extracted data, and assessed bias using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. A bivariate random-effects model was used to synthesize the data. Subgroup analyses were conducted for three factors: (a) imaging modality (computed tomography vs. positron emission tomography-computed tomography); (b) algorithm type (deep learning vs. conventional machine learning); and (3) validation strategy (external vs. internal).ResultsThirteen studies encompassing 6628 patients were included. The pooled sensitivity was 71% (95% confidence interval: 68-74), the pooled specificity was 81% (95% confidence interval: 78-84), and the summary area under the curve was 0.85 (95% confidence interval: 0.82-0.88). Deep learning models significantly outperformed conventional machine learning models (area under the curve: 0.871 vs. 0.798; P = 0.012). Computed tomography-based models yielded higher accuracy than positron emission tomography-computed tomography-based models (area under the curve: 0.879 vs. 0.828; P = 0.038). Models validated on independent external cohorts demonstrated superior performance compared with those relying solely on internal validation (area under the curve: 0.922 vs. 0.841; P = 0.006). Imaging modality was a significant source of heterogeneity (P < 0.05). No threshold effect or publication bias was detected.ConclusionMachine learning-based radiomics models exhibit promising diagnostic accuracy for the noninvasive prediction of epidermal growth factor receptor mutations in Chinese patients with lung adenocarcinoma. Computed tomography-based deep learning models subjected to independent external validation represent the current optimal approach. However, the retrospective nature and substantial heterogeneity of the included studies necessitate large-scale, prospective, multicenter trials with standardized workflows before clinical translation.
4 July 2026
Read appraisal →Diagnostic and interventional imaging
Artificial intelligence in emergency musculoskeletal imaging: A critical review of current applications.
Artificial intelligence (AI) is increasingly shaping emergency musculoskeletal imaging, where rapid and accurate diagnosis is often challenged by high imaging volumes and time pressure. These constraints increase the risk of missed injuries and underscore the need for tools that support faster and reliable assessments. AI systems show promise in improving workflow efficiency by prioritizing urgent studies, guiding modality selection, and reducing reporting delays. Deep learning models can enhance abnormality detection by identifying fractures, soft tissue injuries, and infectious processes, and can provide structured classifications that support clinical decision making. Pediatric-focused AI systems address age-specific developmental considerations and offer valuable support for clinicians with varying levels of pediatric musculoskeletal expertise. Large language models further expand the role of AI by improving report clarity, generating structured impressions, and facilitating communication with clinicians, patients, and families. Despite these advances, challenges remain, including limited external validation, dataset bias, and medicolegal considerations. This review summarizes current AI applications across these domains and highlights key strengths, limitations, and future directions for safe and effective integration into emergency musculoskeletal imaging.
2 July 2026
Read appraisal →Emergency medicine journal : EMJ
Interventions to reduce imaging in children with upper or lower extremity injuries: a systematic review and meta-analysis
BACKGROUND: Radiation exposure, transition delays and costs associated with unnecessary imaging in children have stimulated research into clinical decision rules and other interventions to reduce imaging in the emergency department (ED). The objective of this systematic review is to examine the effectiveness of implementing interventions to reduce imaging in children with upper/lower extremity injuries in the ED. METHODS: Seven databases and the grey literature were searched up to May 2024. Comparative studies assessing interventions to reduce imaging in children with upper/lower extremity injuries implemented in the ED were eligible. Two independent reviewers screened for study eligibility, quality assessment and data extraction, with disagreements settled via third-party adjudication. Changes in imaging are reported as ORs with 95% CIs, using a random effects model. RESULTS: From 9387 citations, eight unique studies enrolling 7793 children were included with the majority using a before-after design. Potential concerns for bias were documented due to a lack of reporting of key quality domains. Decision rules for ankle injuries successfully reduced radiography (OR=0.11; 95% CI 0.07 to 0.16, I2=38%). A decision rule for wrist injuries reduced imaging (OR=0.06; 95% CI 0.03 to 0.11); however, eight injuries were missed. Two studies implementing clinical guidelines reported decreases in radiographs per patient (p<0.001). One trial reported increased imaging in children assessed by triage nurses using an established clinical decision rule (OR=5.44; 95% CI 2.96 to 10.02), with 16 missed injuries identified. CONCLUSIONS: Guidelines incorporating clinical decision rules, particularly decision rules for ankle injuries, can reduce radiography for children with extremity injuries in the ED. Further investigations are warranted to identify other extremity injuries, the components of the intervention and the most efficient clinicians to target. PROSPERO REGISTRATION NUMBER: CRD42016042875.
1 July 2026
Read appraisal →