Citations

Large language models (LLMs) have the potential to transform healthcare by assisting in documentation, diagnosis, patient communication, and medical education. However, their integration into clinical practice remains a challenge. This perspective explores the barriers to implementation by synthesizing recent evidence across five challenge domains: workflow misalignment and diagnostic safety, bias and equity, regulatory and legal governance, technical vulnerabilities such as hallucinations or data poisoning, and the preservation of patient trust and human connection. While the perspective focuses on barriers, LLM capabilities and mitigation strategies are advancing rapidly, raising the likelihood of near-term clinical impact. Drawing on recent empirical studies, we propose a framework for understanding the key technical, ethical, and practical challenges associated with deploying LLMs in clinical environments and provide directions for future research, governance, and responsible deployment.

Introduction/Objectives: Little is known about the prevalence of patient-reported social risk factors and the use of the HOUSES Index, a simple, reliable method of assessing socioeconomic status (SES) based on publicly available housing data, in a predominantly rural, primary care population. Methods: We conducted a cross-sectional analysis of adult patients paneled to family medicine clinicians in a US Midwest health system as of December 31, 2022. Patients’ listed address determined HOUSES Index as quartile rank (Q1 lowest SES) and rural/urban status. Social risk data including housing, food, transportation, finances, and violence were collected from health record questionnaires. A mixed effect model was used to assess associations between social risk, HOUSES Index, and rurality. Results: Of the 352 355 patients included, rural patients were more likely than urban patients to report all social risk factors and had lower SES as measured by HOUSES quartiles. In the mixed effects analysis, HOUSES quartile was independently predictive of reporting an at-risk social risk factor (Q1 vs Q4 OR = 2.27, 95% CI = 2.19-2.37), but rurality was not (OR = 1.02, 95% CI = 0.97-1.07) after adjusting for HOUSES. Conclusions: The increased prevalence of social risk factors among rural residents is largely explained by individual SES measured by HOUSES Index.

ABSTRACT Objective Sacrifice of pterygopalatine fossa (PPF) neurovascular structures during endoscopic endonasal transpterygoid approach (EETPA) may impact a patient's comorbidity. We present anatomical and surgical techniques for maximizing PPF transposition while preserving its neurovascular structures through orbito‐pterygo‐sphenoidal (OPS) ligament release and descending palatine canal (DPC) decompression. Methods The EETPA was performed on six specimens. Two measurements were obtained to assess PPF transposition: (1) Inferior transposition (distance between the superior margins of the base of the pterygoid process and PPF); (2) lateral transposition (distance from Eustachian tube lateral margin to PPF medial margin). Results After incising the OPS ligament, a mean gain of PPF inferior transposition of 7 mm (4–11 mm, p = 0.03) was observed, with a total inferior transposition of 12 mm (6–15 mm). Subsequently, the posterior half of the inferior turbinate was removed and the DPC was decompressed, with a mean gain of PPF lateral transposition of 12 mm (range 8–15 mm, p = 0.01). Conclusion The OPS ligament release offers a significant advantage for the PPF inferior transposition, allowing access to the inferolateral recess of the sphenoid sinus, cavernous sinus, and paramedian middle cranial fossa, while the DPC decompression provides a significant advantage for the PPF lateral transposition, granting access to the Eustachian tube and ITF. Level of Evidence N/A.

Purpose of review To synthesize the current literature regarding the evaluation and management of genitourinary fistula in women. Recent findings Genitourinary fistula are aberrant communications between the urinary tract and genital tract that present with urinary leakage per the vagina. Initial management often involves conservative measures, such as urethral catheter or ureteral stent placement, progressing to surgical repair when needed. Key surgical principles include a tension-free, watertight closure with well-vascularized tissue including tissue interposition as appropriate, and postoperative urinary drainage. When surgical management of vesicovaginal fistula is necessary, a transvaginal repair is the most common. Other minimally invasive approaches are increasing and result in similarly high success rates with lower patient morbidity compared to open abdominal surgeries. The initial management of ureterovaginal fistula commonly includes ureteral stent placement. When not feasible or in persistent fistula, laparoscopic and robotic surgical repair with ureteral reconstruction offers high success rates with lower morbidity than an open approach. Summary Successful management of genitourinary fistula ranges from conservative urinary tract drainage to surgical interventions based on etiology, location, and complexity. Approaches to repair are shifting toward less invasive procedures. With optimal technique and surgical planning, high success rates can be achieved, particularly in primary repairs.

Abstract Aims Type 2 myocardial infarction due to myocardial oxygen supply–demand imbalance is associated with poor outcomes. There are no guidelines to inform care for these patients. The consensus on the assessment and management of type 2 myocardial infarction is gained. Methods and results An international e-Delphi study including experts in type 2 myocardial infarction identified through systematic review was conducted. Participants were asked to describe their approach to (i) definition and diagnosis, (ii) risk stratification, (iii) assessment of coronary artery disease and cardiac function, (iv) specialty management, (v) treatment and secondary prevention, and (vi) communication and rehabilitation. Statements generated in round one were circulated, with consensus defined a priori as ≥70% agreement on a 5-point Likert scale. Where no consensus was reached, statements were amended and recirculated for a final round. The response rate was 56% (38/68), 54% (37/68), and 72% (49/68) in the first, second, and third rounds, respectively. Following the first round, 67 unique statements were generated across six domains. Overall, consensus was achieved on 64% (43/67) of statements. Consensus was achieved for 42% (5/12) of statements on the diagnosis of type 2 myocardial infarction, 75% (3/4) on risk stratification, 50% (9/18) on the assessment of coronary artery disease and cardiac function, 60% (6/10), on specialty management, 100% (9/9) on treatment and secondary prevention, and 79% (11/15) on communication and rehabilitation. Conclusion Consensus was obtained across a number of domains for the assessment and management of patients with type 2 myocardial infarction. However, there was limited agreement amongst experts on the diagnostic criteria, which may benefit from refinement.

Abstract Background Physical activity (PA) is an established protective factor for colorectal cancer (CRC), but it is unclear if genetic variants modify this effect. To investigate this possibility, we conducted a genome-wide gene–PA interaction analysis. Methods Using logistic regression and two-step and joint tests, we analyzed interactions between common genetic variants across the genome and PA in relation to CRC risk. Self-reported PA levels were categorized as active (≥ 8.75 MET-h/wk) vs. inactive (< 8.75 MET-h/wk) and as study- and sex-specific quartiles of activity. Results PA had an overall protective effect on CRC (OR [active vs. inactive] = 0.85; 95%CI = 0.81–0.90). The two-step GxE method identified an interaction between rs4779584, an intergenic variant near the GREM1 and SCG5 genes, and PA for CRC risk (p-interaction = 2.6×10 − 8 ). Stratification by genotype at this locus showed a significant reduction in CRC risk by 20% in active vs. inactive participants with the CC genotype (OR = 0.80; 95%CI = 0.75–0.85), but no significant PA–CRC association among CT or TT carriers. When PA was modeled as quartiles, the 1-d.f. GxE test identified that rs56906466, an intergenic variant near the KCNG1 gene, modified the association between PA and CRC (p-interaction = 3.5×10 − 8 ). Stratification at this locus showed that increase in PA (highest vs. lowest quartile) was associated with a lower CRC risk solely among TT carriers (OR = 0.77; 95%CI = 0.72–0.82). Conclusions In summary, we identified two genetic variants that modified the association between PA and CRC risk. One of them, related to GREM1 and SCG5 , suggests that the bone morphogenetic protein (BMP)-related, inflammatory, and/or insulin signaling pathways may be associated with the protective influence of PA on colorectal carcinogenesis.

Abstract Objective This study explores and documents the patient experience during the prodrome phase of migraine. Background Migraine attacks can be divided into four stages: the prodrome (or premonitory phase), the aura, the headache phase, and postdrome. Qualitative data on the range of symptoms during the prodrome and their timing relative to headache onset are sparse. Prodromal symptoms may predict the onset of migraine headache pain and provide a clinically useful benchmark for initiation of treatment early in an attack before pain begins. Design/Methods Eligible participants with a clinician‐confirmed diagnosis of migraine and at least one prodromal symptom were consented, screened, and then participated in 60‐min interviews. The interview guide included open‐ended questions to elicit spontaneous reports and specific probes about prodromal symptoms based on a clinician‐established list. A theory approach was used to analyze the qualitative data collected in interviews to identify key themes and gather insights; data were analyzed using ATLAS.ti. Results Data collection for this study occurred from March 8, 2022, to May 16, 2022. Twenty interviews were conducted, and analyses demonstrated that concept saturation was achieved. Participants reported 36 unique prodromal symptoms, and each participant experienced a mean of 13 symptoms (standard deviation [SD] = 6.6) and a median of 11 symptoms (interquartile range [IQR] = 7.8–17.0) during the prodrome phase. The most commonly reported prodromal symptoms were nausea ( n = 17/20, 85%), fatigue/tiredness ( n = 16/20, 80%), sensitivity to light ( n = 13/20, 65%), neck pain/stiffness ( n = 12/20, 60%), and dizziness/vertigo/light‐headedness ( n = 10/20, 50%). Of the symptoms reported by at least four participants ( n ≥ 4/20, 20%), neck pain/stiffness was rated most bothersome (8.9 out of 10) and sensitivity to light was rated most severe (8.5 out of 10). Almost 40% of all symptoms reported occurred less than 2 h before the start of migraine headache. Of the commonly reported symptoms, nausea (0.8h), sensitivity to light (1.0 h), and dizziness/vertigo/lightheadedness (2.0 h) began closest to headache onset; fatigue/tiredness (4.0 h) and neck pain/stiffness (4.8 h) were the most remote from headache onset. Conclusion This study identified frequently reported prodromal symptoms, with participants commonly reporting that a migraine headache would follow their experience of prodrome within a 1‐ to 6‐h window. Characterizing the prodrome experience may improve measurement strategies for the burden of migraine and create opportunities to treat during the prodromal phase to prevent the onset of moderate or severe headaches.

BACKGROUND MRI-guided targeted fascicular biopsy has been a valuable technique for diagnosing idiopathic neuropathies. It has an 80% yield and a minor permanent morbidity rate of 5%. OBSERVATIONS A 48-year-old woman with a history of localized amyloidosis to the orbital muscle 2 years earlier presented with new neurological symptoms. Extensive evaluations including skin, muscle, and distal cutaneous nerve biopsies were negative, and empirical treatment trials did not improve her condition. An MRI-guided targeted biopsy of the right ulnar nerve in the distal arm was planned. During surgery, a medial triceps branch arising from the ulnar nerve was identified and was biopsied along with a neighboring fascicle of the ulnar nerve to the flexor carpi ulnaris. The pathology of the ulnar nerve fascicle but not the contiguous triceps branch showed amyloid (AL [kappa] subtype on mass spectroscopy). The patient was started on treatment. LESSONS The positive biopsy of the ulnar nerve fascicle allowed targeted treatment of the neural amyloidosis. While the negative biopsy of the triceps branch could be attributed to sampling error, it could also be due to the fact that this branch found within the ulnar nerve is thought to originate from the radial nerve. https://thejns.org/doi/10.3171/CASE25484

Purpose of review This paper reviews the current understanding of coronary vascular dysfunction (CVDys) in patients with angina and no obstructive coronary artery disease (ANOCA), based on recent findings from a comprehensive, large-scale study. It also discusses potential future directions for research and clinical practice. Recent findings CVDys involves enhanced vasoconstriction or impaired vasodilation caused by endothelium-dependent and/or -independent dysfunction in the epicardial or microvascular coronary arteries. It is classified into four subtypes based on anatomical location and underlying mechanisms, and can be assessed using coronary reactivity testing (CRT). Previous studies have evaluated their prognostic implications separately or in limited populations. A recent study assessed all subtypes within a single ANOCA cohort and confirmed that endothelium-dependent epicardial dysfunction and both types of microvascular dysfunction are associated with worse prognosis. Moreover, it demonstrated that microvascular function – both endothelium-dependent and -independent – serves as an independent prognostic factor for major adverse cardiovascular events. These findings emphasize the value of CRT-based comprehensive evaluation for risk stratification. Summary While evidence regarding the prognostic impact of CVDys has become increasingly robust, effective treatment strategies remain undefined. Future efforts should focus on developing CRT-guided interventions targeting coronary dysfunction, with the goal of improving clinical outcomes.

Abstract Deconvolution models are a powerful tool for extracting cell-type-specific information from bulk gene expression profiles. Current methods leverage advanced machine learning models and high-resolution sequencing, like single-cell RNA-sequencing, showing promising results across diverse tissues and conditions. However, they still present important limitations: First, many depend on selecting a robust reference, which can strongly affect the deconvolution. Second, pseudobulk data used for training and real bulk RNA-seq samples often exhibit strong distribution shifts, which are currently unaccounted for. Finally, most deconvolution approaches behave as black boxes, which can compromise the reliability of the results. Here, we present Sweetwater, an adaptive and interpretable autoencoder that efficiently deconvolves bulk samples leveraging multiple classes of reference data. Moreover, we propose an improved way of generating training data from a mixture of FACS-sorted FASTQ files, reducing platform-specific biases and outperforming current single-cell-based references. Furthermore, we introduce a gold standard dataset to facilitate fair and accurate evaluation of deconvolution approaches. Finally, we demonstrate that Sweetwater adapts effectively to deconvolved samples during training, uncovering biologically meaningful patterns and enhancing result’s reliability. Sweetwater is available at https://doi.org/10.6084/m9.figshare.29609180, and we anticipate it will expedite the accurate examination of high-throughput clinical data across diverse applications.