Abstract Introduction Dysphagia is a frequent and clinically relevant complication after intracerebral haemorrhage (ICH) and is associated with increased morbidity and mortality. While radiologic severity and lesion characteristics are known to influence neurological outcome, their relationship with objectively measured dysphagia severity (DS) and swallowing management remains insufficiently defined. This study investigated the association between radiologic ICH burden and DS assessed by flexible endoscopic evaluation of swallowing (FEES), and examined the clinical impact of FEES-guided swallowing management in an neuro intensive care unit (ICU) cohort. Methods We conducted a retrospective analysis of consecutive ICH patients treated in neurological ICU who underwent standardized bedside dysphagia assessment, with FEES performed when clinically indicated and feasible. DS and functional oral intake were quantified using the Fiberoptic Endoscopic Dysphagia Severity Scale (FEDSS) and the Functional Oral Intake Scale (FOIS-G). Multivariable regression models adjusted for age and clinical severity were used to examine associations between radiologic parameters (hematoma volume, location, ICH score, intraventricular and subarachnoid haemorrhage) and dysphagia outcomes. Results Of 241 patients 120 (49.8%) were diagnosed with relevant dysphagia FEES was performed in 68 patients (28.2%) at a median of 13 days after admission. After FEES, the diet was modified in 67% of patients and the oral intake score improved from a median of one to four. Radiologic severity was independently associated with dysphagia outcomes: higher ICH score predicted lower odds of of favourable FOIS-G and FEDSS scores. Deep intracerebral haemorrhage, intraventricular haemorrhage and subarachnoid haemorrhage were associated with worse functional oral intake. FEES led to a modification of swallowing management in 67% of examined patients, most frequently allowing safer and more liberal oral intake compared with bedside assessment alone. Conclusion In ICU-treated patients with ICH, dysphagia is common and closely linked to radiologic severity and lesion characteristics. FEES frequently reveals clinically relevant discrepancies between bedside assessment and actual swallowing function and substantially influences swallowing management. Integrating radiologic risk stratification with FEES-guided decision-making may improve safety, nutrition, and rehabilitation planning in this high-risk population.
Structured reporting in knee MRI represents a transformative advancement in musculoskeletal radiology, promising enhanced clarity and consistency in evaluating the knee’s complex anatomy—menisci, ligaments, cartilage, and bone. This review article explores the foundations, evolution, and clinical applications of structured reporting, with a special focus on its role in knee MRI. Furthermore, we examine different studies highlighting the benefits of structured reporting, like improved clinician communication and support for emerging artificial intelligence (AI) tools, while also addressing challenges like balancing structure and a certain amount of narrative flexibility. Frameworks such as anatomical versus tissue-based approaches are explored, and insights into possible best practices are offered. The outlook is promising with AI-driven automation, natural language processing, and patient-centric innovations, driven by global standardization efforts from radiological societies. Structured reporting in knee MRI invites further exploration to unlock its full potential, ultimately leading to improved patient care.
Effective pain management is essential in trauma care, as it reduces suffering and prevents long-term complications. However, prehospital analgesia is often inadequate, particularly in challenging environments such as alpine rescue settings, where establishing intravenous access can be difficult. Oral transmucosal fentanyl citrate (OTFC) offers a fast-acting, needle-free alternative, but has rarely been studied in civilian prehospital settings, particularly among paediatric and elderly populations. This study aimed to evaluate the feasibility, effectiveness and safety of OTFC for prehospital analgesia in a civilian mountain rescue context, covering all age and sex groups. This retrospective observational cohort study included all trauma patients treated with oral transmucosal fentanyl citrate (OTFC). Patients experiencing severe pain (NRS ≥ 4) and lacking intravenous access were eligible. OTFC was administered by trained emergency medical services providers. Data on demographics, pain scores and adverse events were prospectively recorded using a standardised digital protocol. The primary outcome was pain reduction, as measured by the Numeric Rating Scale (NRS); safety was assessed by recording adverse events. Statistical analysis included paired and unpaired tests, ANOVA and non-parametric equivalents, as appropriate. Significance was set at p < 0.05. A total of 365 patients (60.5
Background:Large language models (LLMs) can accurately classify biomedical documents, but strong benchmark performance does not establish that predictions are grounded in the supplied text. In biomedical literature tasks, titles, abstracts, digital object identifiers (DOIs), journal metadata, and trial identifiers may have been seen during pretraining and can trigger parametric knowledge or learned associations. Objective:This study aimed to test whether oncology randomized trial success classification is driven by abstract evidence or by identifier-conditioned prior knowledge, and assess whether models follow counterfactual outcome evidence when it conflicts with original trial identifiers. Methods:We evaluated 250 two-arm oncology randomized controlled trials from 7 major journals published between 2005 and 2023, each with a single primary endpoint and previously adjudicated positive or negative ground-truth label. The corpus included 58.4% (146/250) positive and 41.6% (104/250) negative trials. GPT-5.2, Gemini 3 Flash, and Claude Opus 4.5 were queried via vendor APIs under default settings using a single-token output instruction. For each trial, we created 5 deterministic input conditions: title+abstract, title only, DOI only, counterfactual title+abstract in which the primary endpoint outcome statement was minimally flipped, and the same counterfactual input paired with the original DOI to create an identifier-text conflict. Performance was assessed using valid format rate, accuracy, sensitivity, specificity, and F1-score. Results:The models showed high format adherence, with valid prediction rates of 97.2% to 100%. In the title+abstract condition, all models achieved high and balanced performance (accuracy and F1-score=0.96-0.97; sensitivity=0.96-0.97; specificity=0.96-0.98). Removing evidence reduced performance stepwise: title-only accuracy and F1-score fell to 0.79 to 0.88, and DOI-only performance fell to 0.63-0.67, exceeding the 58.4% majority class baseline but indicating limited identifier-driven signal. Counterfactual edits were concentrated in outcome-bearing text, with the Results and Conclusions sections modified for all trials, whereas the titles and Methods sections required edits in only 5.2% (13/250) and 1.6% (4/250) of trials. Against inverted labels, models followed counterfactual evidence with near-ceiling performance (accuracy and F1-score=0.96-0.99). Reintroducing the original DOI caused little change for GPT-5.2 (accuracy and F1-score=0.99) but modestly reduced F1-scores for Gemini (0.97) and Claude (0.95), mainly through lower sensitivity. Conclusions:The evaluated LLMs robustly followed explicit end point statements in abstracts, including when those statements contradicted original trial outcomes. However, above-chance title-only and DOI-only performance, together with small decrements under counterfactual DOI conflicts, showed that identifiers can carry predictive signal and occasionally compete with textual evidence. Progressive content removal combined with counterfactual identifier-text conflicts offers a practical, reproducible audit for grounding in biomedical LLM evaluations.
Tenecteplase (TNK) and Alteplase (rt-PA) are intravenous thrombolytics used in acute ischemic stroke (AIS). While Alteplase has long been the standard of care, TNK is now also approved for AIS by both the EMA and FDA. Its single-bolus administration and favorable pharmacokinetics offer practical advantages. Despite growing guideline support, real-world data comparing both agents remain limited. This study aimed to compare clinical characteristics, imaging findings, treatment outcomes, and complication rates of AIS patients treated with TNK versus Alteplase at a single stroke center. We conducted a retrospective monocentric analysis of 101 AIS patients treated with TNK (n = 39) or Alteplase (n = 62). We compared demographics, vascular risk factors, imaging findings, reperfusion therapies, and clinical outcomes using independent samples t-tests, Chi-square, and Mann–Whitney U-tests as appropriate. Baseline characteristics were comparable between groups, including age (TNK: 72.7 ± 14.4 vs. Alteplase: 70.1 ± 14.3), sex, and pre-stroke mRS. Initial NIHSS scores were similar (TNK: 6.0 ± 6.3 vs. Alteplase: 7.2 ± 8.2, p = 0.83). No significant differences were found in LVO rates (15.4
