Danish Ministry of Defence

A stability-indicatingHPLC-UV method enabling the simultaneousseparation and monitoring of perampanel (PER) and all degradationproducts was developed and applied, providing adequate resolutionand peak purity. Forced degradation studies were conducted to definethe degradation profile of PER under acidic, alkaline, oxidative,photolytic, and metal-ion stress conditions. Five degradation products(DP-1 to DP-5) were identified by HPLC–UV and structurallycharacterized by HRMS/MS, and 1D/2D NMR spectroscopy(1H, 13C, COSY, and HSQC). Acidic hydrolysisproduced a benzamide derivative (DP-1) and a benzoic acid derivative(DP-2), oxidative stress generated a pyridine N-oxidederivative (DP-3), and alkaline conditions yielded two previouslyunreported products, a ring-cleavage derivative of perampanel (DP-4)and a hydroxylated pyridone derivative of perampanel (DP-5). DP-1was also detected during accelerated stability studies of PER tabletsformulated with acidic excipients, demonstrating excipient–drugincompatibility under stressed storage conditions (40 °C/75%RH). The toxicity risk of the five degradation products was assessedusing complementary in silico approaches (statistical and expert-rulemethods) in accordance with ICH M7 guidance. None of the degradationproducts showed mutagenic concern and were therefore classified asICH M7 Class 4 or Class 5 impurities. These results expand the structuralknowledge of PER degradation products, clarify its main degradationpathways, and provide experimental and toxicological support for formulationdevelopment and impurity control.

Despite decades of efforts to optimize procedure-specific pain management, up to 60% of surgical patients still experience moderate to severe pain and opioid-related adverse effects. Current guidelines primarily rely on standardised dosing protocols rather than individual patient characteristics, offering limited support for tailoring opioid therapy to patient-specific needs. This systematic review aimed to identify studies on decision support tools for adjusting perioperative systemic opioid dosing as individualised postoperative pain management and evaluate their effects on postoperative outcomes. We searched EMBASE, MEDLINE and Cochrane Library up to August 2025for studies including adult surgical patients developing, validating or testing decision support tools (prediction models or nociception monitoring systems). Prediction model studies had to include at least two predictors and target the first 24 postoperative hours. Intervention studies had to report pain, opioid use, or opioid-related adverse events within 24 h. Studies on chronic pain, intraoperative-only agents (e.g., remifentanil), or unpublished literature were excluded. Risk of bias was assessed using the Prediction model Risk of Bias Assessment Tool in prediction model studies and Cochrane Risk of Bias and ROBINS-I tool in intervention studies. We conducted meta-analyses for intervention studies (RCTs only) when outcomes were comparable. We included 19 studies on decision support tools for perioperative opioid management: seven development and one impact study on personalised prediction models and 11 intervention studies on nociception monitoring systems. All prediction model studies predicted postoperative opioid requirements, had a high risk of bias, limited external validation and poor overall performance (R2 range: 0.100-0.313). None were clinically tested or provided advice on optimal opioid dosing. For intervention trials evaluating nociception monitoring systems, meta-analyses showed no significant effect on pain, opioid use, or opioid-related adverse events within 24 h after surgery. No studies tested risk stratified/individualised opioid treatment. No prediction model study provided concrete guidance for individualising systemic opioid treatment. Nociception monitor-guided opioid doses did not improve acute postoperative pain, opioid use or opioid-related adverse events. This systematic review assessed if available tested prediction models for individual guiding of opioid analgesic dosing, including monitors designed to assess nociceptive signal influence on clinical status, were of benefit in clinical practice. Findings, for outcomes analgesic effect, opioid dosing, or adverse events related to opioids, none of the models or systems have been demonstrated to have cliear clinical utility for optimizing opioid dosing.

The growing use of injectable cosmetic fillers has been accompanied by an increase in reports of acute vision loss, typically caused by inadvertent intravascular injection and subsequent embolization of the ophthalmic artery, resulting in ocular ischemia. Effective management of this complication remains challenging, as clear clinical guidelines for retrobulbar hyaluronidase injection are lacking. This ultrasound-based observational anatomical guideline provides practical recommendations for retrobulbar administration of hyaluronidase. Real-time ultrasonography allows for direct visualisation of orbital anatomy, enabling the injector to identify the optic nerve sheath complex and surrounding vascular structures. The central retinal artery’s entry point into the optic canal is a key anatomical landmark for directing treatment. Based on ultrasound observations, the inferolateral quadrant of the orbit is identified as the safest and most accessible entry site, reducing the risk of damage to extraocular muscles, cranial nerves, and major vessels. The use of ultrasonography enhances both precision and safety, supporting timely and effective management of filler-related ocular complications. Ultrasound observation further confirms that the inferolateral orbital approach offers the most reliable and safe access for retrobulbar hyaluronidase injection. The superior orbital rim and close proximity of the globe restrict needle trajectory from above, increasing the risk of ocular injury. In contrast, the inferior approach provides a wider anatomical corridor, reduces the likelihood of direct contact with the eyeball, and allows safer navigation around critical neurovascular structures. Incorporating real-time ultrasound enhances precision, making the inferior approach the preferred route for managing filler-related ocular complications.

IntroductionIn academy football players, absolute adductor squeeze strength increases during maturation. Cross-sectional studies suggest that in adult football players, previous groin pain may negatively affect adductor squeeze strength. However, it is uncertain if groin injuries affect the natural development of adductor strength in youth elite football players. This study examines longitudinal development of adductor squeeze strength influenced by age, pain during testing, and time-loss groin injury.Materials and MethodsLong lever adductor squeeze strength (handheld dynamometry) was collected weekly in 69 male West African academy football players (age 15.6±1.2 years at baseline). Repeated-measures adductor squeeze strength was monitored in six-month intervals for 2.5 years. A multilevel modeling approach with hierarchical random effects was created to assess adductor squeeze strength (N, N/kg) based on age, pain during the squeeze test, and previous time-loss groin injury days.ResultsAcross 250 total observations in 69 players, absolute adductor squeeze strength (N) increased 13.9 N with every one-year increase in age (p<0.0001) and decreased 3.5 N for every increment of squeeze pain (p=0.007). Relative adductor squeeze strength (N/kg) did not increase with age and decreased 0.07 N/kg with every increment of squeeze pain (p=0.004). Neither model was influenced by previous time-loss groin injury.ConclusionAbsolute adductor squeeze strength increases during adolescence in West African academy football players, but relative adductor squeeze strength is not influenced by age. Higher pain intensity during the squeeze test led to poorer adductor squeeze strength, but previous time-loss groin injury did not affect strength development.

To view the video, click on the link to the right.

We introduce the concept of Procedural Content Generation via Knowledge Transformation (PCG-KT), a new lens and framework for characterizing PCG methods and approaches in which content generation is enabled by the process of knowledge transformation—transforming knowledge derived from one domain in order to apply it in another. Our work is motivated by a substantial number of recent PCG works that focus on generating novel content via repurposing derived knowledge. Such works have involved, for example, performing transfer learning on models trained on one game's content to adapt to another game's content, as well as recombining different generative distributions to blend the content of two or more games. Such approaches arose in part due to limitations in PCG via Machine Learning (PCGML) such as producing generative models for games lacking training data and generating content for entirely new games. In this paper, we categorize such approaches under this new lens of PCG-KT by offering a definition and framework for describing such methods and surveying existing works using this framework. Finally, we conclude by highlighting open problems and directions for future research in this area.

Time consuming Speech Intelligibility (SI) listening tests with human subjects can be replaced by algorithmic SI predictors. In recent years, data-driven SI predictors have been showing promising results. A major limiting factor in the advancement of data-driven SI prediction is that there is a scarcity of SI listening test data available to train the data-driven methods. In this article we propose a data-driven SI predictor that does not require access to an underlying noise-free reference signal, i.e., <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">non-intrusive</i> , and which does not require listening test data for training. Instead, the proposed method exploits a hypothesized link between SI and Speech Presence Probability (SPP). We show that a neural network can be trained on easily obtainable speech in additive noise data to estimate SPP, and that a simple post-processing stage can be applied in order to map the estimated SPP to SI predictions with high accuracy. The proposed method is evaluated and compared to other state-of-the art non-intrusive SI predictors, and achieves the highest performance even in the presence of processed noisy speech, which the SPP estimator has not been trained on.

The recruitment process can be arduous and time-consuming for both recruiters and job seekers. To address this issue, we propose the Beetle platform that leverages artificial intelligence (AI) and Blockchain technology to provide seamless and efficient solutions for the recruitment journey. Beetle platform includes a ChatGPT styled prompt interface, a Tinder-styled matching interface, an AI-powered video based digital twin bot for interview preparation, as well as a Blockchain-driven vetting system. The searching/shortlisting phase utilizes continuous machine learning algorithms to enhance the matching of job seekers with relevant recruiters and/or positions. The digital twin bot integrates Natural Language Processing to prepare candidates for interviews and provide feedback. The vetting phase uses Blockchain technology to store information to improve Beetle's reliability and security. Information stored on it is immutable, transparent, and easily verifiable, thus making the vetting process faster and more reliable.