Search Results (136)

Machine learning is reshaping gel-based additive manufacturing by enabling accelerated material design and predictive process optimization. This review provides a comprehensive overview of recent progress in applying machine learning across gel formulation development, printability prediction, and real-time process control. The integration of algorithms such as neural networks, random forests, and support vector machines allows accurate modeling of gel properties, including rheology, elasticity, swelling, and viscoelasticity, from compositional and processing data. Advances in data-driven formulation and closed-loop robotics are moving gel printing from trial and error toward autonomous and efficient material discovery. Despite these advances, challenges remain regarding data sparsity, model robustness, and integration with commercial printing systems. The review results highlight the value of open-source datasets, standardized protocols, and robust validation practices to ensure reproducibility and reliability in both research and clinical environments. Looking ahead, combining multimodal sensing, generative design, and automated experimentation will further accelerate discoveries and enable new possibilities in tissue engineering, biomedical devices, soft robotics, and sustainable materials manufacturing. Full article

Introduction/Objective: Peritonitis remains a serious complication in patients undergoing automated peritoneal dialysis (APD), requiring prompt and effective antibiotic administration. This study evaluated whether delivering antibiotics directly through APD bags is as effective as administering them via an additional manual daytime exchange. Methods: We conducted a randomized, single-blind, non-inferiority clinical trial involving patients diagnosed with peritonitis. Participants were randomly assigned to receive Ceftazidime and Vancomycin, either via APD bags or through a combined approach of continuous ambulatory peritoneal dialysis (CAPD) plus APD. A total of 64 patients (32 per group) were enrolled, with comparable baseline demographic and clinical profiles, including laboratory markers of infection severity and dialysis history. Results: Peritonitis resolved in 90.6% of the patients treated via APD bags and in 81.3% of those receiving antibiotics through manual exchange plus APD. Although this difference did not reach statistical significance ($p$ = 0.281), the observed absolute difference of 9.3% was well within the predefined non-inferiority margin of 30%, supporting the clinical non-inferiority of the APD-only method. The mean time to resolution was similar between groups ($p$ = 0.593). Post hoc power analyses indicated limited statistical power (18.5% for the resolution rate and 9.2% for time to resolution), suggesting that modest differences may not have been detectable given the sample size. Nevertheless, the high resolution rates observed in both groups reflect valid and encouraging clinical outcomes. Conclusion: Antibiotic administration via APD bags demonstrated comparable clinical effectiveness to the combined manual exchange plus APD method for treating peritonitis. Given its operational simplicity and favorable results, the APD-only strategy may offer a pragmatic alternative in routine care. Further studies with larger sample sizes are recommended to confirm these findings and optimize treatment protocols. Trial registration: NCT04077996. Funding source: None to declare. Full article

Purpose: Standard automated perimetry (SAP) remains the gold standard functional test in glaucoma, used primarily for evaluating peripheral vision loss. Central contrast sensitivity (CCS) has emerged as a potential early functional marker of glaucomatous damage. This systematic review aimed to describe the different methods used to measure CCS in randomized controlled trials (RCT) involving glaucoma patients. Methods: We searched the MEDLINE, Embase, CINAHL, Cochrane Central Register of Controlled Trials, Epistemonikos, and ClinicalTrials.gov databases on 25 January 2023, and updated the search on 12 February 2025. Eligible studies comprised RCTs that reported CCS as an outcome in patients with glaucoma, suspected glaucoma, or ocular hypertension. No restrictions were placed on age, sex, ethnicity, geography, intervention, or publication year. Abstracts and full texts were screened independently by two reviewers. Descriptive statistics were used. No formal risk of bias assessment was performed, due to the descriptive nature of the review. Results: Of 1066 records screened, 31 studies met the eligibility criteria. The study sample size ranged from 7 to 207 (median: 23), with most studies involving primary open-angle glaucoma. Interventions were diverse, mainly involving topical medications, with timolol being the most frequent. Eleven CCS test methods were identified. Five studies did not report the method used. The CSV-1000 was the most commonly used test, being applied in 11 studies. Conclusions: CCS has been measured using a wide range of methods in glaucoma RCTs, with limited standardization. Most of the included studies were small, variably reported, and conducted over 10 years ago, suggesting a decreasing interest in CCS as an outcome measure in glaucoma RCTs. Funding: This review was funded by Oslo University Hospital and the Research Council of Norway. Registration: This review was registered on the OSF. Full article

This study presents a unified low-parameter approach to multi-class classification of microorganisms (micrococci, diplococci, streptococci, and bacilli) based on automated machine learning. The method is designed to produce interpretable taxonomic descriptors through analysis of the external geometric characteristics of microorganisms, including cell shape, colony organization, and dynamic behavior in unfixed microscopic scenes. A key advantage of the proposed approach is its lightweight nature: the resulting models have significantly fewer parameters than deep learning-based alternatives, enabling fast inference even on standard CPU hardware. An annotated dataset containing images of four bacterial types obtained under conditions simulating real clinical trials has been developed and published to validate the method. The results (Precision = 0.910, Recall = 0.901, and F1-score = 0.905) confirm the effectiveness of the proposed method for biomedical diagnostic tasks, especially in settings with limited computational resources and a need for feature interpretability. Our approach demonstrates performance comparable to state-of-the-art methods while offering superior efficiency and lightweight design due to its significantly reduced number of parameters. Full article

RNA-based cancer vaccines have emerged as transformative immunotherapeutic platforms, leveraging advances in mRNA technology and personalized medicine approaches. Recent clinical breakthroughs, particularly the success of mRNA-4157 combined with pembrolizumab in melanoma patients, have demonstrated significant improvements in efficacy, with a 44% reduction in recurrence risk compared to checkpoint inhibitor monotherapy. Breakthrough results from pancreatic cancer vaccines and novel glioblastoma treatments using layered nanoparticle delivery systems mark 2024–2025 as a pivotal period for RNA cancer vaccine development. Current RNA vaccine platforms include conventional mRNA, self-amplifying RNA, trans-amplifying RNA, and emerging circular RNA technologies, with over 120 clinical trials currently underway across various malignancies. Critical advances in delivery optimization include next-generation lipid nanoparticles with tissue-specific targeting and novel nanoengineered systems achieving rapid immune system reprogramming. Manufacturing innovations focus on automated platforms, reducing production timelines from nine weeks to under four weeks for personalized vaccines, while costs remain challenging at over $ 100,000 per patient. Artificial intelligence integration is revolutionizing neoantigen selection through advanced algorithms and CRISPR-enhanced platforms, while regulatory frameworks are evolving with new FDA guidance for therapeutic cancer vaccines. Non-coding RNA applications, including microRNA and long non-coding RNA therapeutics, represent emerging frontiers with potential for enhanced immune modulation. With over 60 candidates in clinical development and the first commercial approvals anticipated by 2029, RNA cancer vaccines are positioned to become cornerstone therapeutics in personalized oncology, offering transformative hope for cancer patients worldwide. Full article

Background/Objectives: Subperiosteal implants (SPIs) were first used in the 1940s, but due to their complications and the rise of dental implants, they were discontinued. Thanks to new technologies and new materials, nowadays they are being used again and studied as a treatment for severe bone defects. This review analyzes the clinical results—survival rates and complications—of SPIs used to support full arch rehabilitations of severely resorbed maxillae and mandibles, comparing the outcomes resulting from implant placement conducted in one or two surgical interventions. Methods: An automated search was conducted in four databases (Medline/Pubmed, Scopus, Web of Science, and Cochrane Library), as well as a manual search for relevant clinical articles published before 28 February 2025. The review included human studies with at least four patients, in which SPIs were placed to restore full-arch edentulous maxillae and mandibles. Quality of evidence was evaluated using the Newcastle–Ottawa Quality Assessment Scale and the Joanna Briggs Institute Critical Appraisal tool. Results: A total of 14 studies met the inclusion criteria and were included for analysis, including 958 patients and 973 SPIs. The survival rate was 100% when one surgical intervention was performed and 85% when two interventions were performed after 4–38 months and 3–22 years follow-up, respectively. Conclusions: SPIs would appear to offer a good alternative for patients with severe bone atrophies, especially SPIs fabricated using digital techniques in a single step, presenting promising survival rates and a low complication rate, although more randomized clinical trials with long-term follow-up are needed. Full article

Background/Objectives: During the COVID-19 pandemic, the importance of disinfection and quarantine significantly increased, particularly in situations of staff shortages. Automated disinfection methods, such as hydrogen peroxide vaporization (HPV), are increasingly considered as alternatives to traditional manual disinfection. This study aimed to evaluate the efficacy of HPV compared to standard disinfection practices. Methods: Experiments were conducted at the Infectious Disease Clinical Research Simulation Center of Soonchunhyang University Hospital using Geobacillus stearothermophilus spores as biological indicators. The spores were inoculated on various hospital surfaces and allowed to dry for 120 min. Three disinfection methods were tested: (1) scrubbing with a disposable towel soaked in sodium hypochlorite; (2) placing sodium hypochlorite-soaked towels on the surface for one minute; and (3) HPV alone. Samples were collected post-disinfection and incubated at 55–60 °C. Bacterial cultures were assessed after 24, 48, and 168 h. Results: After 24 h of incubation, sterilization rates were 0% for the scrubbing method, 27% for sodium hypochlorite towels, 68% for HPV alone, and 95% for the combination of sodium hypochlorite and HPV. HPV alone demonstrated statistically greater efficacy compared to standard disinfection practices (p = 0.03). Conclusions: HPV alone may serve as a viable disinfection method in clinical environments, particularly during pandemics when staffing limitations hinder thorough manual cleaning. Further clinical trials are warranted to validate these findings and improve disinfection methods for challenging materials such as fabrics. Full article

Clinical documentation, particularly the hospital discharge report (HDR), is essential for ensuring continuity of care, yet its preparation is time-consuming and places a considerable clinical and administrative burden on healthcare professionals. Recent advancements in Generative Artificial Intelligence (GenAI) and the use of prompt engineering in large language models (LLMs) offer opportunities to automate parts of this process, improving efficiency and documentation quality while reducing administrative workload. This study aims to design a digital system based on LLMs capable of automatically generating HDRs using information from clinical course notes and emergency care reports. The system was developed through iterative cycles, integrating various instruction flows and evaluating five different LLMs combined with prompt engineering strategies and agent-based architectures. Throughout the development, more than 60 discharge reports were generated and assessed, leading to continuous system refinement. In the production phase, 40 pneumology discharge reports were produced, receiving positive feedback from physicians, with an average score of 2.9 out of 4, indicating the system’s usefulness, with only minor edits needed in most cases. The ongoing expansion of the system to additional services and its integration within a hospital electronic system highlights the potential of LLMs, when combined with effective prompt engineering and agent-based architectures, to generate high-quality medical content and provide meaningful support to healthcare professionals. Hospital discharge reports (HDRs) are pivotal for continuity of care but consume substantial clinician time. Generative AI systems based on large language models (LLMs) could streamline this process, provided they deliver accurate, multilingual, and workflow-compatible outputs. We pursued a three-stage, design-science approach. Proof-of-concept: five state-of-the-art LLMs were benchmarked with multi-agent prompting to produce sample HDRs and define the optimal agent structure. Prototype: 60 HDRs spanning six specialties were generated and compared with clinician originals using ROUGE with average scores compatible with specialized news summarizing models in Spanish and Catalan (lower scores). A qualitative audit of 27 HDR pairs showed recurrent divergences in medication dose (56%) and social context (52%). Pilot deployment: The AI-HDR service was embedded in the hospital’s electronic health record. In the pilot, 47 HDRs were autogenerated in real-world settings and reviewed by attending physicians. Missing information and factual errors were flagged in 53% and 47% of drafts, respectively, while written assessments diminished the importance of these errors. An LLM-driven, agent-orchestrated pipeline can safely draft real-world HDRs, cutting administrative overhead while achieving clinician-acceptable quality, not without errors that require human supervision. Future work should refine specialty-specific prompts to curb omissions, add temporal consistency checks to prevent outdated data propagation, and validate time savings and clinical impact in multi-center trials. Full article

Large-language models (LLMs) show promise for automating evidence synthesis, yet head-to-head evaluations remain scarce. We benchmarked five state-of-the-art LLMs—openai/o1-mini, x-ai/grok-2-1212, meta-llama/Llama-3.3-70B-Instruct, google/Gemini-Flash-1.5-8B, and deepseek/DeepSeek-R1-70B-Distill—on extracting protocol details from transcranial direct-current stimulation (tDCS) trials enrolling older adults. A multi-LLM ensemble pipeline ingested ClinicalTrials.gov records, applied a structured JSON schema, and generated comparable outputs from unstructured text. The pipeline retrieved 83 aging-related tDCS trials—roughly double the yield of a conventional keyword search. Across models, agreement was almost perfect for the binary field brain stimulation used (Fleiss κ ≈ 0.92) and substantial for the categorical primary target (κ ≈ 0.71). Numeric parameters such as stimulation intensity and session duration showed excellent consistency when explicitly reported (ICC 0.95–0.96); secondary targets and free-text duration phrases remained challenging (κ ≈ 0.61; ICC ≈ 0.35). An ensemble consensus (majority vote or averaging) resolved most disagreements and delivered near-perfect reliability on core stimulation attributes (κ = 0.94). These results demonstrate that multi-LLM ensembles can markedly expand trial coverage and reach expert-level accuracy on well-defined fields while still requiring human oversight for nuanced or sparsely reported details. The benchmark and open-source workflow set a solid baseline for future advances in prompt engineering, model specialization, and ensemble strategies aimed at fully automated evidence synthesis in neurostimulation research involving aging populations. Overall, the five-model multi-LLM ensemble doubled the number of eligible aging-related tDCS trials retrieved versus keyword searching and achieved near-perfect agreement on core stimulation parameters (κ ≈ 0.94), demonstrating expert-level extraction accuracy. Full article

Osteoarthritis (OA) is a prevalent and disabling chronic disease, with knee OA being the most common form, affecting approximately 73% of individuals over 55 years. Traditional clinical assessments often fail to predict knee structural progression accurately, highlighting the need for improved prognostic methods. This perspective explores the complexity of stratifying knee OA patients based on rapid structural progression. It underscores the importance of such early identification to enable timely and personalized intervention and optimize disease-modifying OA drug clinical trial design, as many trial participants show minimal progression, complicating the assessment of treatment efficacy. We highlight the potential of machine learning (ML) and deep learning (DL) in overcoming this prognostic challenge, as these methodologies enhance classification/stratification capabilities by leveraging multidimensional data and capturing the intricate relationships between diverse features. These include panels of biochemical markers and imaging markers, such as those from magnetic resonance imaging (MRI), as integrating MRI data into ML/DL prognostic models enhances such prediction performance. These automated ML/DL models will offer a transformative approach to stratifying knee OA patients and represent a paradigm shift in disease management. Ultimately, ML/DL applications will not only improve patient outcomes but will also promote innovation in OA research, clinical practice, and therapeutics. Full article

Background/Objectives: Atrial fibrillation (AF) remains a major risk factor for stroke. It is often asymptomatic and paroxysmal, making it difficult to detect with conventional electrocardiography (ECG). While photoplethysmography (PPG)-based devices like smartwatches have demonstrated efficacy in detecting AF, they are rarely integrated into hospital infrastructure. The study aimed to establish a seamless system for real-time AF screening in hospitalized high-risk patients using a wrist-worn PPG device integrated into a hospital’s data infrastructure. Methods: In this investigator-initiated prospective clinical trial conducted at the University Hospital Basel, patients with a CHA₂DS₂-VASc score ≥ 2 and no history of AF received a wristband equipped with a PPG sensor for continuous monitoring during their hospital stay. The PPG data were automatically transmitted, analyzed, stored, and visualized. Upon detection of an absolute arrhythmia (AA) in the PPG signal, a Holter ECG was administered. Results: The analysis encompassed 346 patients (mean age 72 ± 10 years, 175 females (50.6%), mean CHA₂DS₂-VASc score 3.5 ± 1.3)). The mean monitoring duration was 4.3 ± 4.4 days. AA in the PPG signal was detected in twelve patients (3.5%, CI: 1.5–5.4%), with most cases identified within 24 h (p = 0.004). There was a 1.3 times higher AA burden during the nighttime compared to daytime (p = 0.03). Compliance was high (304/346, 87.9%). No instances of AF were confirmed in the nine patients undergoing Holter ECG. Conclusions: This study successfully pioneered an automated infrastructure for AF screening in hospitalized patients through the use of wrist-worn PPG devices. This implementation allowed for real-time data visualization and intervention in the form of a Holter ECG. The high compliance and early AA detection achieved in this study underscore the potential and relevance of this novel infrastructure in clinical practice. Full article

Introduction: The feasibility of enrolling and retaining diverse kidney transplant (KT) seekers in remote studies is sparsely reported. Aims: This study examined the use of a mobile communication strategy to enroll and retain participants within a clinical trial of an automated digital intervention to promote self-learning for kidney transplant access. Materials and Methods: Adult KT-seekers were identified from an administrative database at a transplant center and recruited by email or text supplemented by verbal prompts. Multivariable logistic regression was used to explore participant- and study-level characteristics associated with enrollment and response rates. Results: Between April 2022 and June 2023, 743 patients were invited to participate, and 422 were enrolled. Enrollers were more likely to be younger (aOR 1.02; p < 0.001). Early enrollment was associated with text message invitation (OR 2.69, p ≤ 0.014). Survey completion at 1 month was similar across patient sociodemographic, clinical, and study characteristics; however, participants self-reporting Black race were underrepresented at month 6 (OR 0.55, p = 0.015) and month 12 (aOR 0.37, p = 0008), and males were underrepresented at month 12 (aOR 0.45, p = 0.028). Conclusion: Mobile communication methods are viable for enrolling diverse KT-seeking patients and collecting survey data remotely. More work is needed to learn how to best recruit older people and retain diverse groups long-term. Full article

This work introduces TrialSieve, a novel framework for biomedical information extraction that enhances clinical meta-analysis and drug repurposing. By extending traditional PICO (Patient, Intervention, Comparison, Outcome) methodologies, TrialSieve incorporates hierarchical, treatment group-based graphs, enabling more comprehensive and quantitative comparisons of clinical outcomes. TrialSieve was used to annotate 1609 PubMed abstracts, 170,557 annotations, and 52,638 final spans, incorporating 20 unique annotation categories that capture a diverse range of biomedical entities relevant to systematic reviews and meta-analyses. The performance (accuracy, precision, recall, F1-score) of four natural-language processing (NLP) models (BioLinkBERT, BioBERT, KRISSBERT, PubMedBERT) and the large language model (LLM), GPT-4o, was evaluated using the human-annotated TrialSieve dataset. BioLinkBERT had the best accuracy (0.875) and recall (0.679) for biomedical entity labeling, whereas PubMedBERT had the best precision (0.614) and F1-score (0.639). Error analysis showed that NLP models trained on noisy, human-annotated data can match or, in most cases, surpass human performance. This finding highlights the feasibility of fully automating biomedical information extraction, even when relying on imperfectly annotated datasets. An annotator user study (n = 39) revealed significant (p < 0.05) gains in efficiency and human annotation accuracy with the unique TrialSieve tree-based annotation approach. In summary, TrialSieve provides a foundation to improve automated biomedical information extraction for frontend clinical research. Full article

Recent advancements in electronic medical records (EMRs) have transformed clinical trials and healthcare systems by improving data accuracy, regulatory compliance, and integration with decision support tools. These innovations enhance trial efficiency, streamline patient recruitment, and enable large-scale data analysis while bridging clinical practice with research. Despite these benefits, challenges such as data standardization, privacy concerns, and usability issues persist. Overcoming these barriers through policy reforms, technological innovations, and robust methodologies is essential to maximizing the potential of EMRs. We examine current developments, challenges, and future directions for optimizing EMRs in clinical trials and healthcare delivery. Full article

Improved image quality, better scanners, innovative software technologies, enhanced computational power, superior network connectivity, and the ease of virtual image reproduction and distribution are driving the potential use of digital pathology for diagnosis and education. Although relatively common in clinical oncology, its application in liver pathology is under development. Digital pathology and improving subjective histologic scoring systems could be essential in managing obesity-associated steatotic liver disease. The increasing use of digital pathology in analyzing liver specimens is particularly intriguing as it may offer a more detailed view of liver biology and eliminate the incomplete measurement of treatment responses in clinical trials. The objective and automated quantification of histological results may help establish standardized diagnosis, treatment, and assessment protocols, providing a foundation for personalized patient care. Our experience with artificial intelligence (AI)-based software enhances reproducibility and accuracy, enabling continuous scoring and detecting subtle changes that indicate disease progression or regression. Ongoing validation highlights the need for collaboration between pathologists and AI developers. Concurrently, automated image analysis can address issues related to the historical failure of clinical trials stemming from challenges in histologic assessment. We discuss how these novel tools can be incorporated into liver research and complement post-diagnosis scenarios where quantification is necessary, thus clarifying the evolving role of digital pathology in the field. Full article