Bryan Lim

Background/Objectives: Large language models (LLMs) are increasingly used in healthcare, but their utility in plastic surgery and skin cancer management remains in contention. Furthermore, the comparative performance of LLMs in the setting of skin cancer surgery is unclear. This study assessed four widely used LLMs and comparatively evaluated their relevance, accuracy, and clinical representativeness. Methods: This prospective study evaluated six standardised skin cancer prompts submitted to ChatGPT, Gemini, Claude, and Copilot. Specialist plastic surgeons assessed textual responses for relevance and accuracy, and generated images for clinical representativeness and accuracy. Readability and DISCERN scores were also assessed. Results: All four LLMs produced highly relevant textual responses, with relevance scores ranging from 4 to 5, while accuracy scores ranged from 3 to 5. ChatGPT and Copilot achieved the highest individual accuracy scores. Only ChatGPT and Gemini generated images. ChatGPT demonstrated greater clinical representativeness and accuracy across all image prompts. ChatGPT produced the most linguistically complex responses, while Copilot produced the most readable outputs. All models generated text requiring at least a high-school level of education. Claude achieved the highest mean DISCERN score (43.67), followed by ChatGPT (40.33), Copilot (36.00), and Gemini (35.67). Conclusions: In this exploratory comparison, contemporary LLM-based platforms produced relevant and largely accurate skin cancer surgery information, but differed in image generation, readability, and information quality. The generated images depicted predominantly lighter skin tones, which warrants greater input from more diverse patient populations, and outputs should be used alongside clinician review. Larger studies using patient-specific scenarios and diverse skin tones are needed.

Machine learning (ML), a subset of artificial intelligence (AI), is revolutionizing health care. Given the intricate anatomy and functional complexity of the hand, AI applications in hand pathology are gaining prominence, particularly in the diagnosis of fractures, nerve injuries, and musculoskeletal disorders. This review summarizes ML advancements in hand pathology diagnosis, enhancing clinical history analysis, examination findings interpretation, and imaging assessment. A systematic search of PubMed, Google Scholar, EMBASE, Cochrane CENTRAL, Scopus, and Web of Science from 1901 to March 2025 was conducted, focusing on ML applications in diagnosing hand fractures, tendon injuries, and nerve compressions. The inclusion criteria encompassed original research, systematic reviews, meta-analyses, and observational studies that demonstrated the use of AI in diagnostic imaging, preoperative assessment, or postoperative management. Studies that did not involve AI-driven diagnostic applications or lacked sufficient methodological transparency were excluded. Machine learning models, particularly convolutional neural networks (CNNs), have demonstrated high diagnostic accuracy in detecting fractures. Studies have validated CNNs in the detection of distal radius and ulna fractures, outperforming trainees and urgent care physicians. In addition, AI has enhanced clinical examination by integrating wearable motion sensors for objective joint assessment. However, challenges such as external validation, interpretability, and integration into clinical workflows remain, limiting broader clinical adoption. Machine learning shows promise in hand pathology diagnosis, but widespread adoption requires improved external validation, standardized methods, and interdisciplinary collaboration. Future research should prioritize multicentric studies and explainable AI to enhance trust and integration.

Background: Painful digital neuromas are disorganized nerve proliferations that develop following traumatic or iatrogenic digital nerve injury, often resulting in debilitating allodynia, hypersensitivity, and impaired hand function. While conservative therapies such as desensitization, splinting, and pharmacological agents may offer temporary relief, over 50% of patients ultimately require surgical intervention. However, surgical strategies remain highly heterogeneous, and no consensus exists regarding the most effective operative technique. Methods: A comprehensive search of 4 databases through May 2025 identified 30 studies encompassing over 600 adult patients who underwent surgical treatment for digital neuromas. Techniques assessed included neuroma excision with nerve stump transposition, direct end-to-end neurorrhaphy, nerve capping, regenerative peripheral nerve interface (RPNI) surgery, and targeted muscle reinnervation (TMR) surgery. Results: Excision with transposition relieved pain in up to 80% of cases but demonstrated inconsistent sensory recovery. Neurorrhaphy was associated with superior functional outcomes and a reduced recurrence rate. Nerve capping and polyglycolic acid-collagen conduits significantly improved pain scores and protective sensation. RPNI and TMR surgery, though reported in small series, provided over 80% pain relief with minimal complications. Meta-analysis was precluded by outcome heterogeneity. Conclusion: Surgical treatment of painful digital neuromas provides meaningful pain relief and functional improvement, particularly when techniques restore physiologic targets for regenerating axons rather than relying on excision alone. Continuity-restoring and reinnervation-based approaches, including neurorrhaphy, nerve capping, RPNI, and TMR, demonstrate favourable outcomes with lower recurrence rates, although comparative evidence remains limited. Standardised outcome measures and higher-quality comparative studies are required to guide optimal, patient-specific surgical decision-making.

Background: Successful trapeziometacarpal (TM) joint arthroplasty depends heavily on bone quality, particularly of the trapezium, where cup loosening remains a leading cause of implant failure. While dual-energy x-ray absorptiometry measures bone mineral density (BMD), it lacks regional specificity. Computed tomography (CT)-derived Hounsfield units (HU) offer quantitative, site-specific bone density estimation and may serve as a surrogate for BMD in preoperative planning. This study investigated the relationship between CT-derived HU values and tactilely perceived bone hardness of the trapezium to evaluate HU as a reliable predictor of bone quality. Methods: Ten embalmed Body Donor hands were scanned using a Philips Spectral CT 7500 scanner. Bone segmentation and density analyses were conducted in 3D Slicer software, focusing on trabecular regions of the trapezium (proximal, ridge, and submetacarpal) and the first metacarpal (proximal, medial, and distal). Hounsfield unit values were converted to density (g/cm ³ ) using validated equations. An experienced anatomist assessed trapezium hardness (“soft” or “hard”) using bone forceps. Correlations between HU values, bone volume, and perceived hardness were analyzed using Pearson correlation coefficient. Results: The mean trapezium HU was 354 (682.88 g/cm ³ ), with regional trabecular variation: proximal 160.7 HU, ridge 286.5 HU, submetacarpal 241.7 HU. The first metacarpal demonstrated broader density variation (−285.8 to 228.8 HU). Hounsfield unit values correlated significantly with perceived hardness ( r = .71, P = .02), whereas bone volume showed no correlation ( r = .08, P = .82). Male trapezia displayed higher HU and density (472.7 ± 89.8 HU; 740.9 ± 44.0 g/cm ³ ) than females (436.1 ± 87.2 HU; 723.0 ± 42.7 g/cm ³ ). Conclusions: Computed tomography–derived HU values strongly correlate with perceived bone hardness and provide an objective surrogate for assessing bone quality in TM joint arthroplasty. Incorporating preoperative HU assessment may guide implant selection and fixation strategies, reducing trapezium cup loosening and implant failure.

Introduction: This research compares the patterns of hand injuries, treatments and outcomes between male and female Australian rules football players. Methods: This retrospective cohort study reviewed Australian rules football players treated for surgically operable hand injuries at Frankston Hospital in Melbourne, Australia, between January 2021 and August 2024. Data collected included demographics, injury mechanisms, types of fractures or dislocations, treatment methods and postoperative complications. Results: Among 46 Australian rules football players (36 males, 10 females) in this cohort study, notable gender differences in injury patterns were identified. Males predominantly sustained axial loading fractures (30.6%), whereas females more often experienced hyperextension injuries (30.0%) and trampling injuries (20.0%). Both genders primarily underwent open reduction and internal fixation (55.6% males, 60.0% females). Complication rates were low (5.6% males, 10.0% females), but recovery times were generally longer for females. Conclusions: The study highlights distinct injury patterns and outcomes by gender, suggesting physiological and anatomical differences may influence injury susceptibility and recovery in Australian rules football players. These findings underscore the importance of gender-specific prevention and rehabilitation strategies to enhance player safety, recovery and performance.

Background: Large language models (LLMs) such as ChatGPT have evolved rapidly, with notable improvements in coherence, factual accuracy, and contextual relevance. However, their academic and clinical applicability remains under scrutiny. This study evaluates the temporal performance evolution of LLMs by comparing earlier model outputs (GPT-3.5 and GPT-4.0) with ChatGPT-4.5 across three domains: aesthetic surgery counseling, an academic discussion base of thumb arthritis, and a systematic literature review. Methods: We replicated the methodologies of three previously published studies using identical prompts in ChatGPT-4.5. Each output was assessed against its predecessor using a nine-domain Likert-based rubric measuring factual accuracy, completeness, reference quality, clarity, clinical insight, scientific reasoning, bias avoidance, utility, and interactivity. Expert reviewers in plastic and reconstructive surgery independently scored and compared model outputs across versions. Results: ChatGPT-4.5 outperformed earlier versions across all domains. Reference quality improved most significantly (a score increase of +4.5), followed by factual accuracy (+2.5), scientific reasoning (+2.5), and utility (+2.5). In aesthetic surgery counseling, GPT-3.5 produced generic responses lacking clinical detail, whereas ChatGPT-4.5 offered tailored, structured, and psychologically sensitive advice. In academic writing, ChatGPT-4.5 eliminated reference hallucination, correctly applied evidence hierarchies, and demonstrated advanced reasoning. In the literature review, recall remained suboptimal, but precision, citation accuracy, and contextual depth improved substantially. Conclusion: ChatGPT-4.5 represents a major step forward in LLM capability, particularly in generating trustworthy academic and clinical content. While not yet suitable as a standalone decision-making tool, its outputs now support research planning and early-stage manuscript preparation. Persistent limitations include information recall and interpretive flexibility. Continued validation is essential to ensure ethical, effective use in scientific workflows.

Background: Mycobacterium ulcerans causes Buruli ulcer (BU), a necrotizing skin disease endemic in over 30 countries. Its toxin, mycolactone, drives tissue destruction, and the infection is transmitted via environmental reservoirs or vectors. Disease patterns vary globally, and an improved understanding of their pathogenesis may enhance current antimicrobial and surgical treatments. Methods: A comprehensive literature search from 1901 to 2025 was conducted across major databases to explore antimicrobial and reconstructive surgical strategies for Mycobacterium ulcerans. Search terms included BU, key antibiotics, and surgical interventions. Relevant English-language studies on treatment outcomes were reviewed to summarize evolving management trends and emerging therapeutic approaches. Results and Discussion: This review highlights the importance of early diagnosis and timely antimicrobial therapy in preventing disease progression and limb loss. It reviews WHO-recommended antibiotic regimens and discusses the theoretical risk of drug resistance, although clinical resistance remains rare and unreported in Australia. Surgical interventions in select cases are crucial, with timing being a significant factor in functional outcomes. The review also covers pediatric-specific challenges, including growth preservation and psychosocial support for young patients. Reconstructive options focus on limb salvage and staged reconstructions, with multidisciplinary care essential for optimal outcomes. The paper advocates for RCTs to refine treatment protocols, surgical guidelines, and explore emerging antibiotic therapies such as telacebec. Conclusions: BU remains a global health challenge, requiring early diagnosis, timely antimicrobial therapy, and surgery in selected cases. Future research will refine treatment and reduce long-term impacts.

Background/Objectives: Artificial intelligence (AI) holds great potential to reshape the academic paradigm. They can process large volumes of information, assist in academic literature reviews, and augment the overall quality of scientific discourse. This narrative review examines the application of various AI tools in surgical research, its present capabilities, future directions, and potential challenges. Methods: A search was performed by two independent authors for relevant studies on PubMed, Cochrane Library, Web of Science, and EMBASE databases from January 1901 until March 2025. Studies were included if they were written in English and discussed the use of AI tools in surgical research. They were excluded if they were not in English and discussed the use of AI tools in medical research. Results: Forty-two articles were included in this review. The findings underscore a range of AI tools such as writing enhancers, LLMs, search engine optimizers, image interpreters and generators, content organization and search systems, and audio analysis tools, along with their influence on medical research. Despite the multitude of benefits presented by AI tools, risks such as data security, inherent biases, accuracy, and ethical dilemmas are of concern and warrant attention. Conclusions: AI could offer significant contributions to medical research in the form of superior data analysis, predictive abilities, personalized treatment strategies, enhanced diagnostic accuracy, amplified research, educational, and publication processes. However, to unlock the full potential of AI in surgical research, we must institute robust frameworks and ethical guidelines.