Citations

Background: Large language models (LLMs) are increasingly integrated into clinical documentation, decision support, and patient-facing applications across healthcare, including plastic and reconstructive surgery. Yet, their evaluation remains bottlenecked by costly, time-consuming human review. This has given rise to LLM-as-a-judge, in which LLMs are used to evaluate the outputs of other AI systems. Methods: This review examines LLM-as-a-judge in healthcare with particular attention to judging architectures, validation strategies, and emerging applications. A narrative review of the literature was conducted, synthesizing LLM judge methodologies as well as judging paradigms, including those applied to clinical documentation, medical question-answering systems, and clinical conversation assessment. Results: Across tasks, LLM judges align most closely with clinicians on objective criteria (e.g., factuality, grammaticality, internal consistency), benefit from structured evaluation and chain-of-thought prompting, and can approach or exceed inter-clinician agreement, but remain limited for subjective or affective judgments and by dataset quality and task specificity. Conclusions: The literature indicates that LLM judges can enable efficient, standardized evaluation in controlled settings; however, their appropriate role remains supportive rather than substitutive, and their performance may not generalize to complex plastic surgery environments. Their safe use depends on rigorous human oversight and explicit governance structures.

Summary In adults with classic Hodgkin lymphoma (cHL), metabolic tumour volume (MTV) is valuable for risk stratification. This systematic review explored the role of volumetric parameters in children, adolescents and young adults (CAYA) with cHL. This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) method. Five databases were searched on 16 October 2024. Eligible studies were peer‐reviewed manuscripts, written in English, published since database inception that included patients ≤21 years of age with a diagnosis of cHL treated with chemotherapy ± radiotherapy with a baseline 18 F‐Fluorodeoxyglucose (F‐FDG) PET/CT scan for MTV quantification. The search strategy identified 3669 studies, of which 35 were eligible for inclusion. Fifteen studies (43%) evaluated a cohort exclusively ≤21 years of age. All eight studies that explored the association of baseline PET parameters with known risk factors identified a significant correlation. Nine of 12 papers (75%) demonstrated a statistically significant correlation with disease response to therapy on PET/CT. Fifteen of 20 papers (75%) demonstrated an independent statistically significant correlation with clinical outcomes. In the CAYAHL population, volumetric PET parameters are valuable for risk stratification despite heterogeneity in results. Future harmonization of the segmentation approach will facilitate incorporation of MTV into clinical practice.

Abstract Background The Joint Propagation Model (JPM)‐based CenTauR scale was recently introduced to harmonize tau‐PET quantification across different radiotracers. This study examined how CenTauR harmonization enhances the comparability of tau‐PET quantification across matched cohorts using different tracers. The evaluation focused on three key aspects: (1) comparing tau‐PET positivity rates as defined by CenTauR, (2) evaluating its diagnostic accuracy for symptomatic AD, and (3) analyzing longitudinal changes in tau‐PET rates over time. Method The JPM, developed by the CPAD‐led Tau PET Harmonization Working Group (Leuzy et al., Alzheimers Dement. 2024; Figure 1A), models relationships between anchor point subjects and head‐to‐head tau‐PET SUVR data onto the CenTauR scale, providing conversion equations for multiple tracers. JPM equations were applied to [ 18 F]flortaucipir SUVR (Meta‐temporal ROI) data from 561 cognitively impaired participants in the ADNI and OASIS‐3 studies. Using ROC analysis, we determined the CenTauR cut‐off for positivity (T+) that maximized the Youden index for discriminating between FDA‐approved positive/negative visual reads. Three separate cohorts, scanned using [ 18 F]flortaucipir (ADNI), [ 18 F]MK‐6240 (CPAS), and [ 18 F]RO‐948 (BioFINDER‐2), were matched 1:1 by age, MMSE, amyloid‐β, and clinical diagnosis. CenTauR‐based metrics were compared across these cohorts. Result A cut‐off of 18 CenTauRs was found to optimally classify [ 18 F]flortaucipir PET visual reads (Figure 1B). The matching procedure identified 1089 participants (363 per cohort). The frequency of T+ using the 18 CenTauRs cut‐off for the Meta‐Temporal ROI was highly comparable across tracers across groups, except for the Aβ‐positive cognitively unimpaired group, where variability was more pronounced due to a smaller sample size (Figure 2A). Similarly, tau‐PET discriminative accuracy for symptomatic AD vs controls remained similar across [ 18 F]flortaucipir and [ 18 F]MK‐6240 (Figure 2B). In a separate sample of 212 matched participants from ADNI ([ 18 F]flortaucipir) and AIBL ([ 18 F]MK‐6240) with baseline and 1‐year follow‐up tau‐PET scans, 1‐year change in CenTauRs was comparable (Figure 2C). An exploratory voxelwise transformation, utilizing the Meta‐Temporal conversion equation in a representative case scanned with both [ 18 F]flortaucipir and [ 18 F]MK‐6240, demonstrates the potential for voxelwise CenTauR harmonization (Figure 3). Conclusion These analyses suggest that CenTauR harmonization increases the comparability of tau‐PET data acquired with different tracers. Additional validation analyses with larger cohorts including other radiotracers ([ 18 F]PI‐2620) are underway.

Abstract Introduction To investigate the effect of empagliflozin on right ventricular (RV) function in patients with heart failure with reduced ejection fraction (HFrEF). Sodium-glucose cotransporter-2 (SGLT2) inhibitors improve outcomes and reverse left ventricular (LV) remodelling in HFrEF. The impact on RV function remains uncertain. Methods The Empire HF trial was an investigator-initiated, double-blind, randomized, placebo-controlled trial of 190 participants with a left ventricular ejection fraction (LVEF) of 40% or lower, with New York Heart Association (NYHA) Class I−III symptoms. Participants were randomized to receive either empagliflozin (at a dose of 10 mg once daily) or placebo, on top of recommended therapy for 12 weeks. The primary endpoint of this exploratory substudy was changes in RV free wall strain (RVFWS) across the whole cohort. RVFWS was also stratified into tertiles based on baseline RVFWS. Secondary endpoints included changes in tricuspid annular plane systolic excursion (TAPSE) and RV S’ velocity. Results Between June 2017 and September 2019, a total of 190 patients were enrolled, of whom 160 were included in this substudy. Baseline characteristics were balanced between the groups (mean age of 64 ± 11 years, 86% male, mean LVEF: 29 ± 8%, 79% in NYHA Class II, 83 patients (52%) had an implanted cardiac device, and tricuspid valve regurgitation was absent/trace in 136 (85%) and mild in 24 (15%)). A high proportion of the population were on optimal medical treatment for HFrEF, and device therapy remained unchanged during follow-up. The overall mean RVFWS was −16.7 ± 6.1%, with the empagliflozin group at −16.4 ± 6.2% and the placebo group at −16.9 ± 5.9%. No differences were observed in RVFWS between the groups. When stratified by baseline RVFWS into tertiles, patients in the lowest tertile demonstrated a significant improvement with empagliflozin (treatment effect: −2.9% [95% CI: −5.0 to −0.3]; P = .027). This finding was independent of LVEF, plasma volume, and weight loss, and remained unchanged after additional adjustment for LV remodelling and the presence of resynchronization therapy. No significant treatment effect was observed in the middle or highest tertiles of RVFWS, nor in the overall or lowest tertiles of TAPSE or RV S’. Conclusion This exploratory substudy of the Empire HF, treatment with empagliflozin exerted no overall effect on RV function in stable HFrEF patients, but significantly improved RVFWS in patients in the lowest tertile of RVFWS after 12 weeks of treatment. Trial Registration ClinicalTrials.gov, Unique Identifier: NCT03198585

Abstract Background The Joint Propagation Model (JPM)‐based CenTauR scale was recently introduced to harmonize tau‐PET quantification across different radiotracers. This study examined how CenTauR harmonization enhances the comparability of tau‐PET quantification across matched cohorts using different tracers. The evaluation focused on three key aspects: (1) comparing tau‐PET positivity rates as defined by CenTauR, (2) evaluating its diagnostic accuracy for symptomatic AD, and (3) analyzing longitudinal changes in tau‐PET rates over time. Method The JPM, developed by the CPAD‐led Tau PET Harmonization Working Group (Leuzy et al., Alzheimers Dement. 2024; Figure 1A), models relationships between anchor point subjects and head‐to‐head tau‐PET SUVR data onto the CenTauR scale, providing conversion equations for multiple tracers. JPM equations were applied to [¹⁸F]flortaucipir SUVR (Meta‐temporal ROI) data from 561 cognitively impaired participants in the ADNI and OASIS‐3 studies. Using ROC analysis, we determined the CenTauR cut‐off for positivity (T+) that maximized the Youden index for discriminating between FDA‐approved positive/negative visual reads. Three separate cohorts, scanned using [¹⁸F]flortaucipir (ADNI), [¹⁸F]MK‐6240 (CPAS), and [¹⁸F]RO‐948 (BioFINDER‐2), were matched 1:1 by age, MMSE, amyloid‐β, and clinical diagnosis. CenTauR‐based metrics were compared across these cohorts. Result A cut‐off of 18 CenTauRs was found to optimally classify [ 18 F]flortaucipir PET visual reads (Figure 1B). The matching procedure identified 1089 participants (363 per cohort). The frequency of T+ using the 18 CenTauRs cut‐off for the Meta‐Temporal ROI was highly comparable across tracers across groups, except for the Aβ‐positive cognitively unimpaired group, where variability was more pronounced due to a smaller sample size (Figure 2A). Similarly, tau‐PET discriminative accuracy for symptomatic AD vs controls remained similar across [ 18 F]flortaucipir and [ 18 F]MK‐6240 (Figure 2B). In a separate sample of 212 matched participants from ADNI ([¹⁸F]flortaucipir) and AIBL ([¹⁸F]MK‐6240) with baseline and 1‐year follow‐up tau‐PET scans, 1‐year change in CenTauRs was comparable (Figure 2C). An exploratory voxelwise transformation, utilizing the Meta‐Temporal conversion equation in a representative case scanned with both [¹⁸F]flortaucipir and [¹⁸F]MK‐6240, demonstrates the potential for voxelwise CenTauR harmonization (Figure 3). Conclusion These analyses suggest that CenTauR harmonization increases the comparability of tau‐PET data acquired with different tracers. Additional validation analyses with larger cohorts including other radiotracers ([¹⁸F]PI‐2620) are underway.

ABSTRACT Telemedicine is a broad term that encompasses all aspects of utilizing information and communication technologies to provide health services at distant sites. The primary goal of telementoring is the education and training of gastrointestinal (GI) endoscopists particularly in areas where there is a shortage of endoscopy trainers. Telementoring in GI endoscopy can help to provide teaching and skills training both in simulator‐based learning, and endoscopic skills training in the context of patient care. This position statement aims to outline the role of telementoring within GI endoscopy, offering a general model for the introduction and evaluation of a telementoring program that incorporates both simulators for junior trainees and skills training for practicing endoscopists. We hope to highlight its utility and implementation in remote locations. WEO recognizes that telementoring is valuable to expand access to high‐quality endoscopy training especially in parts of the world where there is a shortage of endoscopy trainers and this document can serve as a guide in the planning and implementation of telementoring programs.

Background The association between lifestyle factors and Alzheimer's disease (AD) pathophysiology remains incompletely understood. Objective The aim of this study was to assess the association of alcohol consumption, smoking behavior, sleep quality and physical, cognitive, and social activity with cerebral amyloid pathology. Methods For this cross-sectional study, we selected participants from the Amyloid Biomarker Study data pooling initiative. We used generalized estimating equations to assess associations of dichotomized lifestyle measures with amyloid pathology. Results We included 9171 participants with normal cognition (NC) and 2555 participants with mild cognitive impairment (MCI) from the Amyloid Biomarker Study. Of participants with NC, 58% were women, 34% were APOE ε4 carrier, and 27% had amyloid pathology. Of participants with MCI, 48% were women, 47% were APOE ε4 carrier, and 57% had amyloid pathology. In NC, cognitively active participants were less likely to have amyloid pathology (OR = 0.77, 95%CI 0.66–0.89, p < 0.001). In MCI, participants who had ever smoked or had sleep problems were less likely to have amyloid pathology (OR = 0.85, 95%CI 0.73–0.99, p = 0.029; OR = 0.62, 95%CI 0.45–0.86, p = 0.004). Conclusions In NC, cognitive activity was associated with a lower frequency of amyloid pathology. In MCI, favorable lifestyle behaviors were not associated with a lower frequency of amyloid pathology. The results of the current study contribute to the broader evidence base on lifestyle and AD by further characterizing the role of lifestyle behaviors in AD pathology across different clinical stages.

ABSTRACT IPTA created the SMARTER Initiative to address persistent gaps in pediatric transplant education worldwide. Through webinars, podcasts, case discussions, and interactive resources, SMARTER provides trainees and allied professionals with accessible, peer‐reviewed content across all major organ systems. The program is designed for global reach, with downloadable modules, mobile compatibility, and recorded sessions to accommodate different regions and time zones. Since its launch, SMARTER has shown strong engagement, with fellows reporting that the platform helps them prepare for board exams, improve clinical decision‐making, and connect learning directly to patient care. Survey feedback highlights high satisfaction and identifies opportunities for further growth, such as expanded case‐based modules, procedural videos, and simulation exercises. By reducing disparities in transplant training and supporting collaboration, SMARTER is positioned to strengthen existing curricula and promote a better‐prepared, more globally connected transplant workforce.

Introduction/Background: CHIP is a risk factor for cardiovascular (CV) disease, cancer, and all-cause mortality. Previous work has shown that CHIP, and particularly larger CHIP clones, are associated with adverse CV outcomes, yet the molecular pathways through which CHIP impacts CV risk are poorly defined. Hypothesis/Research Question: We hypothesize that the integration of whole blood transcriptomics and methylomics will provide novel insights into the pathophysiology of CHIP. Methods/Approach: Whole blood DNA methylation profiling, transcriptomics, and whole exome sequencing with CHIP calling for variant allele frequencies (VAF) of ≥2% (CHIP) and ≥10% (large CHIP) were performed for 507 ISCHEMIA and ISCHEMIA-CKD participants with moderate-severe ischemia. We identified transcriptomic and methylomic differences between participants with CHIP and large CHIP vs no CHIP using DESeq2 and limma, adjusted for age, sex, and race/ethnicity. Gene set enrichment analysis (GSEA) and probe set enrichment analysis (PSEA) were performed to identify pathway-level alterations in transcription and methylation, respectively. Results/Data: Clinical characteristics of study participants are described in Fig 1A . Compared to no CHIP (n=391), transcriptomics identified 6 differentially expressed genes (DEGs) in CHIP (n=116) and 27 DEGs in large CHIP (n=35) (p-adj<0.05; abs(logFC)>0.25) (Fig 1B) . Compared to no CHIP, methylation identified no differentially methylated probes in CHIP and 6 in large CHIP (padj<0.20, abs(logFC)>0.03). GSEA identified 137 pathways significantly different in both CHIP and large CHIP vs. no CHIP (padj<0.05), while PSEA identified 724 and 2356 pathways (padj<0.20), respectively. Given its stronger relationship with methylation and transcription, downstream analyses focused on large CHIP. Integrating these data, we found 58 pathways to be both hypomethylated and transcriptionally upregulated in large CHIP, including azurophil granule-related pathways implicated in neutrophil degranulation (Fig 1C) . Further investigation into the gene-probe pairs driving the azurophil granule pathway enrichment in large CHIP revealed hypomethylation and increased transcription of genes implicated in neutrophil extracellular trap formation, including ELANE, PTRN3, AZU1, and CTSG (Fig 1D) . Conclusions: Integration of the methylome and transcriptome suggests large CHIP is linked to neutrophil-mediated immune pathways in patients with stable coronary artery disease.