Background: Instant-messaging platforms like WhatsApp may be a useful and scalable adjunct to undergraduate medical education, particularly when it comes to image-rich disciplines like radiology. We developed a concise, repeatable case-discussion framework for large-scale instant-messaging group chats (IIIDDS: Indication, Imaging, Interpretation, Differentials, Discussion, and Safety-net) to support a student-led, faculty-moderated case-discussion group model. Objective: To explore learner perceptions of a student-led, faculty-moderated radiology WhatsApp group and its structured IIIDDS case-discussion approach between preclinical and clinical undergraduate medical students. Methods: A Checklist for Reporting Results of Internet E-Surveys (CHERRIES) informed exploratory cross-sectional survey was distributed to the Radiology Club WhatsApp members via email and WhatsApp. The survey explored perceived usability, educational value, learning environment, and challenges of the platform and IIIDDS model. Descriptive analyses and subgroup comparisons were performed between medical students with and without clinical rotation experience. Results: In total, 128 responses were evaluated in the analysis. Overall comfort with WhatsApp was high (mean 4.50/5). Participants commonly valued shared radiology cases (82.0%), annotated images (76.6%), and faculty guidance (70.3%). Clinical students reported being more comfortable using WhatsApp (mean 4.75 vs. 4.32; p = 0.006) and a stronger perception of a safe learning environment (mean 3.75 vs. 3.33; p = 0.031) compared to preclinical students. Reported challenges included overwhelming message volume (68.0%) and difficulty keeping up with discussions (61.7%). Discussion: A structured, student-led, faculty-moderated instant-messaging-based case discussion model was perceived as an accessible and credible adjunct for undergraduate radiology education. Providing a structured case presentation framework and a platform for accessible multimedia sharing was perceived to support reliable and constructive case discussions by combining structured discussion with faculty moderation. Findings suggest practical design priorities for similar implementations, including faculty moderation, structure, and strategies to reduce message-overload.

Introduction: Large language models (LLMs) are an emerging form of generative artificial intelligence (AI) with promising applications in medical education, and their ability to provide automated feedback may enhance medical imaging education for trainees. This review aims to systematically examine and synthesize the published literature on the use of LLMs in providing automated feedback in medical imaging education. Methods: We conducted this systematic review in accordance with the PRISMA 2020 guidelines. A comprehensive search of the PubMed, Scopus, and Embase databases was conducted, covering studies published through January 2026. Our search strategy included keywords related to “feedback, generative artificial intelligence, large language models, radiology, and medical imaging.” Studies were eligible if they examined the use of LLMs to generate automated feedback for medical trainees within medical imaging education. Extracted data were synthesized using descriptive synthesis, with quality appraisal assessed using ROBINS-I and GRADE. Results: Of 1,003 identified records, 7 met the inclusion criteria. All studies examined the applications of automated LLM feedback in the medical education of radiology residents, with one study also including fellows. Reported educational outcomes included enhanced report quality, improved diagnostic accuracy, and increased efficiency in discrepancy detection. LLM feedback was generally well-received among trainees, with learners expressing satisfaction with the LLM feedback and preferring a hybrid human-AI feedback model. Additionally, fine-tuned models generally showed stronger performance than general-purpose LLMs and demonstrated variable agreement with expert-human consensus. Conclusion: LLMs show a potentially promising role as supportive tools for providing automated feedback in medical imaging education, alongside human feedback. This includes reported gains in accuracy, efficiency, and learner satisfaction. However, the current published evidence is preliminary and limited. Larger multicenter studies with standardized methods are necessary before widespread adoption can be justified. Our systematic review emphasizes that human expert oversight remains essential, as the current evidence supports preliminary technical feasibility, but not yet definitive educational effectiveness. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251081394 , Identifier CRD420251081394

Background Large Language Models (LLMs) represent an ever-emerging and rapidly evolving generative artificial intelligence (AI) modality with promising developments in the field of medical education. LLMs can provide automated feedback services to medical trainees (i.e. medical students, residents, fellows, etc.) and possibly serve a role in medical imaging education. Aim This systematic review aims to comprehensively explore the current applications and educational outcomes of LLMs in providing automated feedback on medical imaging reports. Methods This study employs a comprehensive systematic review strategy, involving an extensive search of the literature (Pubmed, Scopus, Embase, and Cochrane), data extraction, and synthesis of the data. Conclusion This systematic review will highlight the best practices of LLM use in automated feedback of medical imaging reports and guide further development of these models.

Background: Infliximab (IFX) is a chimeric monoclonal antibody against tumor necrosis factor-alpha (TNF-α) that is widely used for induction and maintenance therapy in pediatric inflammatory bowel disease (IBD), yet its effects on the developing intestinal microbiome and treatment response remain unclear. Objective: To systematically synthesize evidence on IFX-associated microbiome changes in pediatric IBD and evaluate microbiome features linked to treatment response. Methods: A systematic review was conducted following the PRISMA 2020 guidelines. PubMed, Scopus, Embase, and CENTRAL were searched from database inception until March 2026, and our results were synthesized narratively. Results: Of the 945 records identified, 13 studies met the inclusion criteria, comprising 242 pediatric patients. Findings for alpha diversity were variable across studies. In contrast, beta-diversity patterns, taxonomic profiles, and functional analyses more consistently showed positive IFX-associated microbial changes. Responders to treatment more frequently showed enrichment of beneficial taxa, including Faecalibacterium, Subdoligranulum, and Bifidobacterium , while non-responders exhibited a tendency towards dysbiotic microbial signatures, evidenced by increased levels of Gammaproteobacteria and Candida . Functional and metabolomic studies suggested beneficial shifts in bile-acid metabolism, short-chain fatty acid-related pathways, and inflammatory signaling post IFX treatment. Clinical and biochemical improvements with IFX were consistently reported; however, these improvements were not always accompanied by uniform recovery of overall microbial diversity. Conclusion: In pediatric IBD, IFX is more consistently associated with shifts in microbial composition and function than with broad increases in overall microbial diversity. Further large, standardized studies are warranted to refine the role of microbiome features in biologic treatment stratification. Systematic Review Registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251133625 , identifier CRD420251133625.

Background Somatic genetic alterations across different types of tumors are key to understand pan-cancer etiology. Although, pan-cancer genomics analyses have been conducted extensively, the contribution of mutation burden and hotspots is less characterized. Results In this study, we analyzed the somatic variant burden in 10,008 tumor samples spanning 27 cancer types, leveraging data from whole-exome and whole-genome sequencing datasets. Our gene-level tumor variant burden analysis identified 1930 pan-cancer driver genes that harbor high somatic variant burden (top 10th percentile) across at least 10 malignancies. Furthermore, from the low-variant burden group, we extracted 6 candidate pan-cancer genes ( ECRG4, GCOM1, H1-0, H1-6, H2BC15, H2AX, and SEPTIN11 ) with significant variant burden in at least five malignancies. Moreover, to dissect hypermutated gene pathways associated with malignant cell types, single-cell transcriptomic analysis of over 400,000 tumor cells was performed across the five most prevalent cancers. We found that the top 20 pathways were consistently downregulated in immune cell types and upregulated in endothelial cell types. Genome-wide somatic mutation hotspot analysis revealed 960 pan-cancer mutation hotspots across 27 tumors, including 578 candidate pan-cancer hotspot regions not cataloged in the Catalogue of Somatic Mutations in Cancer (COSMIC) database, suggesting unexplored potential drivers of tumorigenesis. DOCK3 , TEAD2 and MSH3 genes were found to be enriched with new somatic variant hotspots in at least five different cancer types. Conclusion This study establishes the pan-cancer landscape of mutation burden. We highlighted pan-cancer-specific genome-wide known and candidate genes/hotspots that may contribute to understanding tumor etiology and progression, and to exploring potential novel therapeutic interventions.