AI News Archive: September 2, 2026 — Part 14

Sourced from 500+ daily AI sources, scored by relevance.

• Transfer Safety Awareness for Cross-Modal Safety Drift in Multimodal Large Language Models

  Visual modality enhances the capabilities of multimodal large language models (MLLMs) but also introduces a safety concern: a benign textual query may convey harmful intent when grounded in a visual image. We term this cross-modal safety drift and our pilot studies show that the safety response rate...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02082v1
• Implicit Manipulation for Skill Selection in LLM Agents with Semantic Matching

  Skill selection is a key stage in LLM-agent workflows, determining which installed skill should handle a user request. Existing attacks on this stage primarily rely on explicit prompt injection or instruction-level steering, which can expose recognizable manipulation signals. In this work, we identi...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02035v1
• Overcoming the Randomness-Utility Trade-off in Answering Differentially Private Linear Queries

  We study the question of answering linear queries with differential privacy using few (expected) random bits. We provide a randomness-efficient analog of the $\| \cdot \|_K$-norm mechanism of Hardt and Talwar [HT10]. For the $\ell_\infty$-error, our algorithm can answer $d$ linear queries with $O(d ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02880v1
• When Does Authorization End? Effect Closure at Provider Boundaries

  Revocation completion, clean state, or operation success can leave authorized work able to cause an effect the application rejects while the provider stays within its contract. We call the absence of all such paths policy-relative effect closure, or effect closure for short. Thus, a grant is closed ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02866v1
• SPADE: SPaT Attack Detection from the Connected Vehicle's Perspective

  Signal Phase and Timing (SPaT) messages are a cornerstone of connected vehicle (CV) safety, enabling CVs to perceive and respond to intersection state through Vehicle-to-Infrastructure (V2I) and Vehicle-to-Vehicle (V2V) communication. The integrity of these messages is threatened by a range of appli...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02741v1
• Card-Based Computation in the Virtual Player Simulation Model

  Player simulation has recently emerged as a new direction in card-based cryptography, with protocols developed for simulating virtual players in physical card games such as Old Maid, UNO, and President. Unlike conventional card-based secure computation, player simulation imposes additional constrain...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02716v1
• PrimSynth: An Agentic Approach to Discover, Validate, and Synthesize Exploit Primitives for Linux Kernel Vulnerabilities

  Linux kernel vulnerabilities are critical to downstream systems. Despite extensive research on automated kernel exploitation, a fundamental challenge remains the conceptual gap between abstract exploit strategies and concrete technical operations. To fill this gap, this paper introduces a systematic...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02647v1
• SpiderSapien: Client-Centric Web Crawler and Security Scanner

  Black-box web application crawling and scanning play an important role for security testing of web applications. Yet state-of-the-art scanners fall short of addressing key characteristics of a modern web application: its extreme dynamism and interactivity on the client side. This paper identifies im...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02532v1
• Can Risk-Based Alerting Mitigate Cybersecurity Alert Fatigue?

  Security operations centers (SOCs) face large numbers of false alerts, making detection of cyberattacks difficult under typical resource constraints. Risk-based alerting (RBA) has been proposed as a means to reduce false alerts and has reportedly succeeded in doing so in various enterprise deploymen...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02465v1
• CAPTCHAs in the Agentic Era: Solvers That Learn from Every Encounter

  Vision-language models (VLMs) can solve visual CAPTCHAs without task-specific training, but the agents built on them approach every challenge from scratch. For such an agent, the hundredth instance of a familiar puzzle costs as much time and compute as the first. Specialized detectors invert the tra...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02393v1
• Removing Speech, Keeping Activities: A Privacy Firewall for Acoustic Sensing in Assisted Living

  Acoustic sensing offers a promising non-intrusive approach for monitoring daily activities of older adults, yet speech privacy concerns remain a critical barrier to real-world deployment. We present a privacy firewall pipeline based on a U-Net encoder-decoder, trained entirely on synthetic data, tha...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02376v1
• Poisoning Attacks on the PGM-index

  The PGM-index (Ferragina and Vinciguerra, VLDB'20) is one of the most practical learned indexes, owing to its theoretical elegance and consistently strong empirical performance. It is built on optimal piecewise linear approximations (PLAs) that minimize the number of segments. In this paper, we ask ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02328v1
• SEAL: Reinforcing Global Safety in Mixture-of-Experts through Shared Expert ALignment

  Mixture-of-Experts (MoE) is a scaling architecture for large language models that activates only a small subset of expert modules per token, enabling massive parameter growth with nearly constant computation. Recent Hybrid MoE architecture adds \textit{shared experts} to capture consistently useful ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02293v1
• Agentic Settlement Protocol: An Application Profile for Refundable, Delayed-Fulfilment Agent Commerce on Stablecoin Rails

  Autonomous agents can already pay per request: HTTP-native protocols such as x402 let an agent sign a stablecoin authorization and receive a resource in the same round trip. That model is atomic and final, which suits metered access and fails commerce: a purchase made on a person's behalf -- a servi...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02208v1
• Stored Is Not Supported: Typed Provenance and Assertion Guardrails for Persistent AI Agents

  Persistent AI agents construct autobiographical state through reflection, retrieval, and consolidation. Persistence changes availability, not epistemic standing: stored or retrieved material is not thereby supported. Untrusted inputs, prompt injections, and model inferences can therefore enter persi...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02127v1
• Type-Directed, Secure-by-Construction Enclave Partitioning for LLVM

  Trusted Execution Environments (TEEs) provide hardware-supported isolation through enclaves that protect code and data independently of software abstractions. However, TEEs alone cannot enforce information-flow security. This problem is further aggravated in LLVM-like low-level languages that allow ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02048v1
• C$^2$T-OpenMax: A Novel Open-Set WiFi RF Fingerprinting Method via Center Constrained Learning and Confidence-Guided Tail Modeling

  Radio frequency fingerprinting (RFF) enables device authentication from transmitter-specific hardware imperfections, but practical deployment requires cross-environment open-set recognition. Data augmentation improves environmental generalization, yet may yield dispersed, low-confidence known-class ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02007v1
• ShikumiMiner: Mining Recurring Implementation Patterns in AI Codebases

  Large language models are paving the way towards innovation by understanding, analyzing, summarizing and generating content in the modern world. Currently there are thousands of LLM projects developed by engineers in open-source repositories. However, whether these LLM projects have underlying patte...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02789v1
• From Prompting to Engineering: A Research Agenda for Prompt Engineering in Software Engineering

  Prompt engineering is increasingly used across Software Engineering (SE) activities, including requirements analysis, coding, testing, documentation, repository analysis, and planning. Yet prompts and related instruction artifacts are often created and evolved through task-specific and informal prac...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02248v1
• ARFT: A Synchronized Multimodal RF-Acoustic Dataset for Positioning in Distributed Environments

  This paper documents the acoustic-radio fusion in Techtile (ARFT) dataset, a synchronized measurement campaign for distributed wireless sensing and positioning in the Techtile testbed. Ultrasonic and radio frequency (RF) signals are simultaneously transmitted and captured at multiple positions in a ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02657v1
• VibeVoice-ASR-Streaming Technical Report

  Traditional speaker-attributed ASR systems treated ASR and speaker diarization as two separate tasks. Recently, end-to-end models such as VibeVoice-ASR have unified the two tasks within a single model. However, existing unified models still mainly support offline recognition, making it difficult to ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02812v1
• Beyond Modality Harmony: Orthogonal Purification and Topology-Guided MoE for Conflict-Aware Multimodal Recommendation

  Multimodal Recommender Systems (MRSs) typically rely on a flawed "modality harmony" assumption, presuming that multimodal features are inherently beneficial and strictly aligned with users' collaborative interaction patterns. However, modality-topology conflicts are ubiquitous in real-world scenario...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02152v1
• Seed-Anchored Budget-Bounded Graph Rendering for Question Answering on Industry-Standard Power-Grid Information and Exchange Models

  Large language model question answering over power-grid models must respect a fixed context budget. We introduce seed-anchored graph rendering, a deterministic method that prioritizes query-local graph evidence without adding method-specific tuned or learned parameters beyond the shared hop bound an...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02011v1
• Recommender System as Slow and Fast Thinkers

  Sequential recommendation models are foundational to modern personalized services, yet their effectiveness varies substantially across heterogeneous user environments. In particular, static one-pass recommenders often perform well on common behavior patterns but degrade on operationally challenging ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02671v1
• Adaptive Test-Time Inference for Text2Cypher with Trace Budgeting and Selective Refinement

  Large language models have enabled natural language interfaces for structured databases, but generated queries may still contain syntactic errors, violate database schemas, or fail during execution. Test-time inference strategies improve generation reliability without additional training, but existi...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02324v1
• Group-Aware Adaptive Retrieval for Evidence Navigation

  Reasoning-intensive retrieval addresses queries whose relevance cannot be identified by surface-level matching, thereby requiring multi-step reasoning. Because relevant documents rarely appear in the initial candidate set, retrieval systems suffer from the bounded recall problem. Existing methods it...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02188v1
• GenCAR: Generative Counterfactual Alignment with Risk-Controlled Selection for Out-of-Distribution Recommendation

  Serving useful recommendations under distribution shift is crucial for balancing utility and risk in out-of-distribution (OOD) recommendation. However, most existing OOD methods improve ranking or construct counterfactual candidates without controlling the proxy-label false discovery rate (FDR) of t...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02162v1
• A Power Law in Logarithm's Clothing: On the Scalability of Graph-Based Vector Search

  Most vector databases rely on graph-based indexes, notably HNSW and Vamana, for approximate nearest neighbor search. With embedding models widely adopted, the datasets these databases store grow rapidly. At a fixed accuracy, how does search cost scale with dataset size? The prevailing answer is poly...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02143v1
• Beyond Context Windows: Persistent Discovery Context for Data-Centric Agents

  Data-centric agents repeatedly perform a discovery step before planning or execution: identifying the data objects relevant to a task. Yet successful discovery outcomes are typically discarded rather than reused. We introduce persistent discovery context, a lightweight memory layer that stores prior...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02129v1
• SPAR: Enhancing Industrial-Scale Generative POI Recommendation via Real-World Spatial Perception

  Generative Point-of-Interest (POI) recommendation, autoregressively generating a target POI's semantic ID (SID), holds great promise for Location-Based Services, where a recommendation helps only if the user can reach it. Yet, existing methods operate within an interest space defined by behavior seq...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02062v1
• GeoStore: Finding Small Storefronts in Large Scenes -- A Fine-Grained POI Localization Benchmark with Global-to-Local Asymmetric Matching

  Point-of-interest (POI) localization -- matching a user's close-up storefront photograph against large-scale geo-tagged street-view imagery -- underpins map construction, POI verification, and location-based services. Its closest existing paradigm, visual place recognition (VPR), assumes symmetric, ...

  📄 ResearchSep 2, 2026http://arxiv.org/abs/2609.02012v1
• Benchmarking ten frontier large language models on 1,477 board style multiple choice questions in hematology

  Large Language Models (LLMs) are increasingly used by clinicians and patients for medical queries, yet their accuracy and safety at the specialist level in hematology remain insufficiently characterised. We benchmarked ten frontier proprietary and open-weight LLMs across two generations on 1,477 board-style hematology multiple-choice questions (MCQs) derived from five educational datasets spanning nine disease areas and six clinical skill domains, including text-only and multimodal case vignettes. Claude Opus 5 had the highest mean accuracy (92.7% text, 76.9% multimodal), followed closely by Gemini-3.1 Pro (91.4% and 78.7%), Gemini-3.6 Flash (91.0% and 74.8%) and GPT-5.6 Sol (89.9% and 76.7%). Accuracy significantly correlated with model size both for text-only and multimodal MCQs. Between model generations, the largest improvements in accuracy were seen for open-weight models whereas proprietary models showed only marginal gains. In error analysis, top-performing models exhibited highly concordant failure patterns, suggesting shared limitations on challenging cases. Frontier LLMs exhibit substantial specialist hematology knowledge across diverse subspecialist domains and clinical skill sets. Yet, despite high accuracy on board-style questions in hematology, continuous expert-on-the-loop output monitoring is paramount.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.09.01.26361881v1?rss=1
• Are Frontier Large Language Models Safer Than Government-Backed Symptom Checkers for Clinical Self-Triage? A Standardised Vignette Evaluation

  Background: Consumer use of AI chatbots for health advice is rising, yet triage safety relative to established services remains unclear. Australia's Healthdirect, a government-backed symptom checker with 2.4 million uses in FY2024-25, remains unevaluated against frontier large language models (LLMs), and whether premium subscriptions improve triage safety remains unexplored. This study compared the triage accuracy and safety of Healthdirect against six LLM configurations across ChatGPT, Claude, and Gemini, assessed whether paid subscriptions improve triage safety, and characterised each system's error patterns. Methods: Forty-five clinical vignettes from the Semigran et al. benchmark spanning emergency, non-emergent, and self-care categories (15 each) were evaluated across seven systems. Healthdirect was tested following a seven-rule interaction protocol. LLMs were evaluated using first-person patient-language prompts under free-tier and paid-tier conditions. Outcomes were triage accuracy, emergency sensitivity, under-triage, and critical misses, analysed using Cochran's Q, Bonferroni-corrected McNemar tests, Cohen's kappa, and Wilson intervals. Findings: Triage accuracy differed significantly (Cochran's Q = 36.79, p < 0.001). Healthdirect achieved 48.9% accuracy (95% CI 35.0% to 63.0%; kappa = 0.233) versus 73.3% to 86.7% for LLMs (kappa = 0.600 to 0.800). Healthdirect operated under conservative interactive defaults while LLMs received complete information in a single prompt, which may have disadvantaged Healthdirect. Emergency sensitivity was 46.7% versus 80.0% to 86.7% for LLMs. Healthdirect produced two critical misses; no LLM produced any across 270 evaluations (95% CI 0% to 1.4%). When LLMs undertriaged, they recommended GP care rather than self-care. No tier differences were significant (all p > 0.05), and most systems over-triaged self-care cases. Interpretation: Frontier LLMs demonstrated higher triage accuracy and safer error profiles than Healthdirect. All LLMs avoided critical misses; Healthdirect did not. Premium subscriptions did not significantly improve triage safety. These findings support clinical governance decisions about whether LLMs warrant formal evaluation alongside government-backed symptom checkers.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.09.01.26361908v1?rss=1
• ECG-based longitudinal risk prediction across diseases and organ systems

  Artificial intelligence applied to routine electrocardiograms (ECGs) has largely focused on detecting existing disease or predicting individual cardiovascular outcomes. Whether ECGs can support prediction of multiple future diseases across organ systems remains unclear. We developed ECG-RISK, a multitask survival model for 67 incident three-character ICD-10 endpoints using ECG waveforms, demographic characteristics and routinely collected laboratory data from 86,673 MIMIC-IV patients. Discrimination was highest for heart, brain, kidney and lung endpoints, with organ-level C-indices ranging from 0.796 to 0.825, whereas liver and pancreatic endpoints showed lower discrimination. The ECG-only model achieved strong discrimination across most endpoints, whereas the incremental improvement gained by incorporating ECG and laboratory inputs beyond demographic information varied substantially across endpoints. Across the nine exploratory aggregated outcomes, Kaplan Meier curves showed clear separation among model-score tertiles. Discrimination was highest for dementia (C-index, 0.891) and heart failure (C-index, 0.857). These findings support the feasibility of ECG-based longitudinal risk prediction across multiple diseases. External validation and competing-risk analyses are required to assess generalisability and clinical utility.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.29.26361697v1?rss=1
• Certified large language model-based diagnostic decision support in rheumatology: the ALLIANCE multicentre randomised controlled trial

  Objectives To evaluate whether access to a certified large language model (LLM)-based clinical decision support system improves physician diagnostic performance in rheumatology compared with conventional diagnostic resources alone. Methods In this multicentre, open-label, randomised controlled trial, 82 physicians from seven hospitals in two countries were randomised 1:1 to conventional diagnostic resources plus Prof. Valmed or conventional resources alone. Participants assessed three rheumatology vignettes before and after assistance. The primary outcome was top-1 diagnostic accuracy. Secondary outcomes included top-3 accuracy, diagnostic reasoning, confidence, case-processing time and perceived support quality. Results Top-1 accuracy increased from 22.2% to 33.3% in the intervention group and from 23.3% to 35.0% in the control group, with no between-group difference in improvement (adjusted OR 0.99, 95% CI 0.45 to 2.19; p=0.979). Differences in top-3 accuracy, diagnostic reasoning and confidence were also not significant. Assisted case-processing time was substantially shorter with LLM support (94 vs 206 s; adjusted mean difference -112 s, 95% CI -141 to -83; p<0.001). Information timeliness and perceived diagnostic support quality were rated significantly higher in the intervention group. Exploratory analyses showed persistent overconfidence and substantial AI over-reliance. Conclusions Certified LLM-based diagnostic support did not improve diagnostic accuracy compared with conventional resources, but substantially reduced case-processing time and improved perceived support quality. These findings suggest potential workflow benefits while highlighting overconfidence and over-reliance as important safety considerations.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.29.26361715v1?rss=1
• Left ventricular hypertrophy, brain atrophy and cognitive decline in type 2 diabetes mellitus: Diabetes & Dementia (D2) cohort study

  BACKGROUND People with type 2 diabetes mellitus (T2DM) are at higher risk of cerebral small vessel disease and left ventricular hypertrophy (LVH), potentially contributing to cognitive decline and dementia. We aimed to describe brain volume and cognitive trajectories over 2 years in a cohort of people with T2DM and to determine whether LVH causes increased brain atrophy and cognitive decline. METHODS Diabetes and Dementia (D2) study is a multicentre observational cohort study in Melbourne, Australia. Participants aged >50 years were recruited via 2 hospital outpatient clinics, 3 private clinics, and study advertisements. Participants with pre-existing cognitive impairment, life-limiting medical illness, and severe chronic renal impairment were excluded. Participants attended study visits for brain MRI, transthoracic echocardiography (TTE), and cognitive testing at baseline and 2 years. The exposure was LVH determined on baseline TTE. Pre-specified outcomes were total brain volume (TBV) change and cognitive decline (z-score change?-1 in any cognitive domain) over 2 years. Regression analyses examined associations between baseline variables and outcomes. A causal inference approach was utilized using inverse probability of treatment weighting to standardize for confounding covariates, excluding participants for non-positivity on age and baseline TBV. RESULTS Participants were recruited 20May2016 to 20March2020: 2378 screened, 702 eligible, 196 consented, 150 baseline and 123 2-year assessments with complete MRI, TTE, and cognitive data (17.4% attrition). At baseline, LVH was associated with female sex, older age, lower educational attainment, lower mood, hypertension, obesity, beta-blocker use, and smaller TBV. Participants with baseline cognitive impairment exhibited greater brain atrophy. Lower educational attainment, hypertension, and lower baseline cognitive scores were associated with cognitive decline. Causal inference analysis included 62 participants with no LVH (20(32%) women; mean [SD]=66.9[5.9] years), and 31 with LVH (17(55%) women, 67.4[5.4] years). LVH caused lower TBV change: standardized mean difference (95% CI) 6.3 (0.1, 12.5) cm3, P=.048. LVH had no effect on cognitive decline. CONCLUSIONS Brain atrophy and cognitive decline were associated with baseline cognitive impairment. LVH caused less brain atrophy and cognitive decline in people with T2DM. We conclude that guideline-directed LVH therapies such as beta-blockers have both cardioprotective (remodelling) and neuroprotective effects. TRIAL REGISTRATION ACTRN12616000546459 UTN: U1111-1181-6659

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361868v1?rss=1
• Hiatal Hernia Size and De Novo Gastroesophageal Reflux Disease After Sleeve Gastrectomy: A Single-Center Retrospective Study

  Background: Sleeve gastrectomy is the most frequently performed bariatric procedure worldwide but is associated with the development of de novo gastroesophageal reflux disease (GERD). Hiatal hernia has been identified as a relevant anatomical factor in postoperative reflux, although most studies evaluate it dichotomously without analyzing whether its size influences GERD risk. The aim was to evaluate the association between preoperative hiatal hernia size and de novo GERD after sleeve gastrectomy. Methods: Retrospective, single - center, observational study of patients undergoing sleeve gastrectomy at Hospital Central Norte de Petroleos Mexicanos (2018 - 2025). Demographic and clinical characteristics, endoscopic classification of hiatal hernia size (small <2 cm, medium 2.1 - 4 cm, large >4 cm), and evidence of de novo GERD were analyzed using descriptive statistics, Fisher's exact test, odds ratio (=R) estimation with 95% confidence intervals (CI), and binary logistic regression. Statistical significance was set at p<0.05. Results: Fiftysix patients were included (mean age 48.3 {+/-} 8.1 years; 67.9% male). Hiatal hernia classification was conclusive in 46 patients (82.1%): 63.0% no hernia, 4.3% small, 30.4% medium, and 2.2% large. De novo GERD occurred in 14.0% of patients without preexisting GERD (6/43). No significant association was found between hiatal hernia size and de novo GERD (Fisher p=0.515). In the reduced logistic model, neither hiatal hernia (medium/large vs. absent/small; OR 3.47; 95% CI 0.50 - 29.43; p=0.207) nor age (OR 1.02; 95% CI 0.90 - 1.13; p=0.754) was significantly associated. No evaluated factor (sex, smoking, alcohol, age) reached significance. Conclusions: In this cohort, no statistically significant association was demonstrated between preoperative hiatal hernia size and de novo GERD after sleeve gastrectomy; however, the low number of events limits the ability to exclude a clinically relevant association. These findings are compatible with a multifactorial mechanism rather than with the isolated presence of this finding. Prospective studies with larger sample sizes and standardized reflux assessment instruments are required to confirm these results.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361833v1?rss=1
• Non-Inferiority Margins in Randomized Controlled Trials in Abdominal Surgery- a Systematic Review

  Importance Non-inferiority trials are becoming increasingly popular in abdominal surgery. The non- inferiority margin is critical in the interpretation and conclusion of these trials. Objective This systematic review aims to assess the methodological and reporting quality of non- inferiority randomized controlled trials in abdominal surgery. Evidence Review Non-inferiority trials were systematically identified by searching Ovid Medline, Embase and the CENTRAL databases from 2006 until December 2025. Randomized controlled trials in adult patients with any type of abdominal surgical intervention in at least one trial arm and a sample size greater than or equal to 100 were eligible for inclusion. The primary outcome was the definition of the non- inferiority margin. Secondary outcomes were the reporting of the non-inferiority margin, the robustness of its estimation, the uncertainty of the point estimate and the adequacy of conclusions. Findings A total of 11 045 trials were identified, of which 101 were eligible, enrolling 44 370 patients. Most trials provided a rationale for the non-inferiority design, while six (5.9%) trials did not. Previous literature was commonly used (n=56; 55.4%), but the non-inferiority margin was most often based on a clinical fixed margin or on historical comparison of the treatment and the active comparator. Based on the margin, investigators tolerated substantially worse outcomes of the treatment compared to the comparator. Conclusions were appropriate based on the confidence interval and the predefined non- inferiority margin in 88 (87.1%) of trials. The clinical judgement of the conclusion was overall adequate. Confidence interval estimations were reported in 16 (15.8%) of trials. Simulation studies were limited by the reporting quality. Conclusions and Relevance Clinical fixed margins are commonly used in abdominal surgery non-inferiority randomized controlled trials, however, substantial shortcomings in reporting limit the interpretability and reproduction of study findings. Based on the findings of this study, guidance on surgical- specific non-inferiority margin definitions is needed.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.29.26361719v1?rss=1
• Senotherapeutic role of pemafibrate through autophagy/mitophagy regulation in chronic obstructive pulmonary disease

  Inadequate autophagy promotes smoking-induced cellular senescence involved in chronic obstructive pulmonary disease (COPD) pathogenesis. Transcription factor EB (TFEB) is a master regulator of the autophagy-lysosome axis. For the first time, we investigated the therapeutic potential of pemafibrate, a putative TFEB inducer. COPD lung epithelial cells showed reduced TFEB expression. Pemafibrate enhanced autophagy/mitophagy flux and restored lysosomal acidification observed during cigarette smoke (CS) extract exposure in human bronchial epithelial cells, resulting in reduced cellular senescence. TFEB knockdown demonstrated involvement of pemafibrate-induced TFEB in these effects. Pemafibrate induced TFEB expression, mitigated alveolar enlargement and airflow obstruction, and attenuated the CS-induced increase in static lung compliance in a long-term CS-exposed mouse model. It reduced the CS exposure-induced cellular senescence, possibly through autophagy/mitophagy, as suggested by bulk RNA sequencing of mouse lungs. A retrospective cohort study showed that patients given pemafibrate displayed attenuated FEV1.0 decline compared with those given bezafibrate or fenofibrate. In conclusion, pemafibrate is a promising therapeutic agent for COPD, potentially exerting its effects through the regulation of the TFEB-autophagy/mitophagy-lysosome axis.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361865v1?rss=1
• Thermal variability and the geography of optimal temperature for child survival: childhood respiratory-infection mortality in 171 countries: a systematic analysis of the Global Burden of Disease Study 2023 and the C-LSAT high-resolution climate dataset

  Background In adults, cold-attributable mortality exceeds heat-attributable mortality roughly 17-fold. Child-specific evidence has begun to emerge only recently - a nationwide Brazilian case-crossover study located the minimum mortality temperature (MMT) for under-five deaths, and a 56-country survey-based analysis linked monthly temperature anomalies to under-five mortality - but no multi-country, climate-zone-resolved estimate of the childhood respiratory-infection MMT exists, and whether temperature variability is independently associated with childhood respiratory mortality at the global scale is unknown. We quantified both. Methods We combined Global Burden of Disease 2023 mortality estimates, lower respiratory infection (LRI) deaths at ages 0-19 years and asthma deaths at ages 0-24 years, 171 countries, 1990-2023 - with 0.5 deg monthly land temperature and diurnal temperature range (DTR) fields from C-LSAT/C-LDTR (1901-2023). Four exposure dimensions (annual mean, DTR, seasonal amplitude, interannual variability) entered two-way fixed-effects models with Driscoll-Kraay standard errors. A quadratic term in mean temperature located the MMT, with percentile confidence intervals from a 300-replication country-cluster bootstrap. Future-exposure leads, country-level detrending, and permutation tests assessed contemporaneous causality, applied to both the linear coefficients and the quadratic term generating the MMT; national pneumococcal conjugate vaccine (PCV3) coverage and ambient PM2.5 exposure series were added as time-varying mechanistic covariates. Results The childhood LRI MMT was 17.1 C (95% CI 14.7-19.8), the 36th percentile of the annual-temperature distribution; zone estimates were 24.7 C in tropical and 15.8 C in subtropical countries, with weak temperate and no subarctic identification. The quadratic term underpinning the MMT, however, failed both falsification checks - future temperatures reproduced the U-shape and country-level detrending erased it - so these MMT values describe a trend-level geographic pattern of the annual construct rather than a contemporaneous dose-response. Interannual temperature variability was positively associated with LRI (+0.278, 95% CI 0.102-0.454; p = 0.002) and asthma mortality (+0.836, 95% CI 0.447-1.226; p = 2.6 x 10^-5) per 1 C, but future-exposure models returned nearly identical significant coefficients and detrending erased significance, supporting only a trend-level association; adjustment for national PCV3 coverage and PM2.5 exposure left these estimates essentially unchanged. Annual mean temperature was likewise inversely associated with both outcomes at the trend level; DTR and seasonal amplitude showed no independent within-country effects. Conclusions This study provides the first multi-country, climate-zone-resolved geography of the optimal temperature for childhood respiratory survival, spanning 171 countries; because the underlying quadratic association is trend-level, the estimates are directional. The observed variability-mortality associations are trend-level signals rather than contemporaneous causal evidence; daily-scale, child-specific designs are required to determine whether short-term thermal variability affects paediatric respiratory mortality.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361864v1?rss=1
• A single-session randomised crossover fNIRS study comparing three upper-limb mirror therapy task paradigms in healthy adults: a study protocol

  Background Mirror therapy (MT)-specifically paradigms using mirror visual feedback (MVF)-is widely used in neurorehabilitation; however, mechanistic implementations vary substantially in movement content, rhythmicity and attentional demands. This protocol describes an acute mechanistic, within-participant fNIRS screening study designed to compare three prespecified upper-limb mirror-therapy task paradigms and to quantify associated subjective experience after each condition in healthy adults during a single visit. Methods and analysis This is a single-centre, within-participant, randomised crossover study conducted at Wuhan Wuchang Hospital (Wuhan, China). Healthy adults aged 18-35 years will complete three task conditions once each in a counterbalanced order using a 3*3 Latin-square scheme: UMT1 (task-oriented rhythmic functional movement), UMT2 (open-ended free movement with auditory control), and UMT3 (non-functional rhythmic movement). fNIRS will be acquired using the NirSmart-6000A system during a standardised block design. The primary outcome is ROI-level HbO activation quantified as GLM-derived {beta} estimates within the prespecified primary ROIs (bilateral SM1/M1 and bilateral PMC). Secondary outcomes include ROI-level windowed {Delta}HbO (5-20 s post-onset relative to the immediately preceding rest; descriptive only), ROI-level {Delta}HbR, and post-condition subjective ratings (illusion, immersion, confusion and fatigue; 1-7 Likert). Condition effects will be analysed using linear mixed-effects models with fixed effects for condition and period and prespecified multiplicity-adjusted pairwise contrasts. Ethics and dissemination Ethics approval was obtained from the Ethics Committee of Wuchang Hospital Affiliated to Wuhan University of Science and Technology (Approval No.: 2025-112-01; approved on 2025-08-21). The study is expected to be minimal risk. Findings will be disseminated through publication of this protocol manuscript and subsequent results manuscripts and conference presentations. Trial registration number Chinese Clinical Trial Registry (ChiCTR2600116634). This study is conducted as a prespecified mechanistic sub-study under the overarching registered project.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.28.26361691v1?rss=1
• Medial Plantar Nerve Shear Wave Elastography and Viscosity Imaging for Differentiating Mild from Moderate Diabetic Peripheral Neuropathy

  Objective: To examine how medial plantar nerve shear wave speed (Cs) and viscosity coefficient (Vi) are associated with the severity of diabetic peripheral neuropathy (DPN), and to assess their ability to differentiate adjacent severity categories. Materials and Methods: Based on TCSS, the 113 patients with type 2 diabetes mellitus were assigned to the non-DPN (n = 33), mild DPN (n = 46), and moderate DPN (n = 34) groups. Medial plantar nerve Cs and Vi were measured using shear wave elastography and viscosity imaging. Receiver operating characteristic analysis evaluated Cs, Vi, and their logistic regression-based combination; areas under the curves (AUCs) were compared using DeLong tests. Results: Cs and Vi increased progressively across the three groups (both P < 0.001). For non-DPN versus mild DPN, the AUCs of Cs, Vi, and the combined model were 0.688 (95% CI, 0.604-0.772), 0.741 (0.660-0.822), and 0.745 (0.665-0.826), respectively, without significant pairwise differences. For mild versus moderate DPN, the corresponding AUCs were 0.707 (0.625-0.789), 0.794 (0.724-0.865), and 0.799 (0.731-0.867). The combined model outperformed Cs (P = 0.045), whereas Cs versus Vi and Vi versus the combined model did not differ significantly (P = 0.162 and 1.000, respectively). Conclusion: Medial plantar nerve Cs and Vi increased with DPN severity. Their combination improved discrimination between mild and moderate DPN compared with Cs alone but not with Vi alone. Quantitative medial plantar nerve viscoelastic assessment may complement clinical severity grading.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.28.26361645v1?rss=1
• Antimicrobial resistance genomics across Africa: critical determinants, repository bias and regional coordination

  Public genomic repositories can support antimicrobial resistance (AMR) surveillance, but unequal sampling can bias interpretation. We characterised AMR determinants, multicountry genomic cluster overlap and surveillance gaps across Africa using an NCBI Pathogen Detection snapshot retrieved on 24 August 2026 for 55 African Union member states. Records were validated and deduplicated by BioSample, and complete AMRFinderPlus calls were summarised across five United Nations M49 subregions and eight overlapping regional economic communities (RECs). Country-pair cluster overlap was assessed using the Jaccard index, while project-based and composition-standardised sensitivity analyses evaluated repository bias. The dataset contained 86,829 unique BioSamples from 51 states; South Africa, Malawi and Kenya contributed 55.8%. Complete extended-spectrum {beta}-lactamase calls were detected in 21,513 isolates and carbapenemase calls in 4,642. blaCTX-M-15 dominated the ESBL profile, while NDM and OXA types predominated. Seventy clusters contained carbapenemase-positive isolates from at least two countries. A shared REC covered all participating countries in 38 clusters, while 32 crossed REC boundaries. Normalised country-pair overlap was low, with a maximum Jaccard index of 9.5%. Project balancing reduced the Northern African carbapenemase estimate from 32.3% to 17.9% and the Eastern African ESBL estimate from 36.9% to 12.5%. Public repositories identify determinants and clusters for investigation but do not estimate prevalence or transmission. AMR surveillance should combine national confirmation, regional institution-led investigation where countries share an REC, and continent-wide coordination through Africa CDC for cross-REC signals, supported by representative One Health sampling, standardised metadata and sustained African sequencing capacity.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361859v1?rss=1
• From Housing to Hotspots: Integrating a Housing-Based Measure of Individual Socioeconomic Status with Geospatial Analysis to Target Colorectal Cancer Screening in Rural Communities

  Purpose: This study assesses the association between colorectal cancer (CRC) screening and a validated, housing-based measure of individual-level socioeconomic status (SES, called HOUSES hereafter) within rural communities and determines whether HOUSES-integrated geospatial analysis can be used to tailor interventions. Methods: We used CRC screening data from a subset of Mayo Clinic Midwest patients living in cities without ready access to routine care in the Mayo Clinic Health System in 2019 to represent rural communities. At the individual level, we assessed the association between CRC screening rates and the HOUSES index, adjusting for age, sex, race/ethnicity, comorbidity, distance from home address to clinic, and area deprivation index, using a multilevel mixed-effects logistic regression model. Additionally, we conducted geospatial analysis to examine the correlation between hotspots of 1) lower CRC screening rates and 2) lower SES of the subject population (HOUSES quartile 1). Findings: Among 34,489 individuals (median age 64.0 years, 52.4% female), those with the lowest SES (HOUSES Q1) had 37% lower odds of being CRC screening adherent than those with the highest SES (HOUSES Q4) (adj. OR [95% CI]: 0.63 [0.58-0.69]). In the 14 identified HOUSES Q1 hotspots, there was a significant correlation in counts of HOUSES Q1 and low CRC screening (correlation coefficient=0.81). Conclusion: Lower SES was significantly associated with lower CRC screening among rural populations. HOUSES-enabled geospatial analysis identified geographic hotspots with lower CRC screening rates for targeted interventions to address disparities in CRC screening in rural communities. HOUSES may be a useful digital tool for cancer preventive care and research.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.28.26361444v1?rss=1
• A Randomized Controlled Trial Evaluating a Community-Based, Family Network Heart Health Intervention - the SERVE OC Trial: Design, Rationale and Baseline Findings

  Background: Cardiovascular disease (CVD) disproportionately burdens underserved communities, where social determinants of health (SDOH) perpetuate persistent disparities. Family-based interventions leveraging social support represent a promising yet understudied approach. We describe the rationale, design, and methods of the Skills-based Educational strategies for the Reduction of Vascular Events in Orange County (SERVE OC) RCT and present baseline characteristics of enrolled families. Methods: SERVE OC is a 2-arm RCT of 190 Latino and Vietnamese families (486 individuals) randomized to the family-based intervention or individual self-management. The intervention was grounded in social network theory while employing community engaged strategies. Primary outcomes include achieving ideal cardiovascular health (CVH) defined by AHA Life's Essential 8 (LE8) and systolic blood pressure reduction at 12, 24, and 36 months. Baseline assessments include demographics, LE8, psychosocial factors, food security, and SDOH. Descriptive statistics and regression analyses examined cohort characteristics and associations between SDOH, food security, and LE8. Results: Over 83% of participants had suboptimal LE8 scores. Average adult total LE8 scores were 66.61 {plus minus}11.96, with physical activity as the weakest domain, compared to an average of 76.52{plus minus}10.15 in children. Greater SDOH burden and food security were associated with significantly lower odds of ideal CVH and lower LE8 scores respectively. Conclusions: SERVE OC demonstrates the feasibility of enrolling families in community-engaged RCT targeting CVD disparities in underserved population. Baseline findings confirm substantial CVD risk and SDOH burden underscoring the need for multi-level, culturally tailored interventions. Trials results will inform scalable, family-focused strategies for CVD prevention across the life course. Clinical Trial Registration: URL: https://www.clinicaltrials.gov/; Unique Identifier: NCT05641519.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361871v1?rss=1
• ClinSeg: Robust Brain Segmentation for Clinically Acquired Pediatric MRI

  Clinical brain MRIs from pediatric health systems represent a viable resource for modeling early neurodevelopmental trajectories and studying neurodevelopmental risk in real-world populations. However, a limitation to date has been the performance of existing segmentation tools for measuring various brain phenotypes in clinical scans. In particular, many tools underperform in infant scans due to morphological and physical changes such as rapid myelination. Here, we introduce ClinSeg: a robust segmentation approach tailored to early-life clinical MRIs with variable orientation, resolution, and contrast. We leverage existing registration and synthetic data generation tools to construct a training corpus for a 3d U-Net spanning anatomical and contrast diversity, including scans with morphological abnormalities from a pediatric hospital. Validated against manual segmentations, ClinSeg outperforms existing models in infancy while matching them in childhood and adolescence. Finally, ClinSeg enables the construction of reference brain growth trajectories in 11,699 individuals from 0-21 years of age, leading to the detection of more nuanced age-related findings in clinical groups.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.28.26361643v1?rss=1
• Warming, thermal variability, and the 96% decline in childhood respiratory-infection mortality in China: a national time-series analysis of the Global Burden of Disease Study 2021 and the C-LSAT high-resolution climate dataset

  Background Childhood respiratory mortality in China has fallen by over 90% in three decades alongside sustained national warming, yet national long-run evidence on temperature and child respiratory mortality is lacking. Methods We linked Global Burden of Disease (GBD) 2021 mortality estimates for China - lower respiratory infections (LRI), ages 0-19, and asthma, ages 0-24, 1990-2021 - with C-LSAT 0.5 deg gridded temperature data (1990-2019), aggregated nationally and to five climate zones. Four annual indicators (mean temperature, diurnal temperature range, seasonal amplitude, interannual variability) entered regressions of log mortality rates with Newey-West standard errors. A bootstrapped (500 resamples) quadratic model probed the minimum mortality temperature (MMT), with PM2.5-adjusted analyses and future-exposure, permutation, and detrended falsification tests. Results LRI deaths fell by 96.3% (330,194 in 1990 to 12,098 in 2021; 95% uncertainty interval 9,669-14,891) and asthma deaths by 94.9% (3,287 to 167), while mean temperature rose 0.364 deg C per decade and diurnal temperature range narrowed 0.092 deg C per decade. Baseline coefficients were large (mean temperature -1.696, SE 0.174; diurnal temperature range +2.408, SE 0.336; seasonal amplitude -0.162, SE 0.082; interannual variability +2.924, SE 1.514, per 1 deg C in log rate), but the future-exposure test failed and detrending nullified every coefficient: the associations are trend-level, and short-cycle causal effects are not identifiable. Nor was the national MMT identifiable - observed temperature support spans only 6.66-8.13 deg C, and the nominal turning point of 35.84 deg C is an extrapolation artifact (quadratic term p = 0.963). Within the observed range, warming and declining mortality moved in the same direction. Conclusions The 96% decline in childhood respiratory mortality cannot be attributed to warming. China sits on the low-temperature side of the optimum, and the marginal direction of future warming requires stronger designs to establish. The falsification framework offers a discipline for climate-health inference in China.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361879v1?rss=1
• Immune Checkpoint Blockade Modifies Drug-Associated Toxicity Across Phenotypes and Time

  Importance: Immune checkpoint inhibitors (ICIs) produce diverse immune toxicities, but whether checkpoint blockade also modifies associations between other drugs and adverse events is poorly understood. Objective: To define ICI-associated toxicity organization and determine whether drug-associated adverse events and onset vary with ICI exposure and checkpoint pathway. Design and Setting: Cross-sectional analysis of deduplicated FAERS reports from 2016 through 2025; analyses performed in 2026. Participants: Among 13,701,106 deduplicated reports, 2,365,269 were cancer associated and 256,940 contained an ICI. Median age among cancer reports with observed age was 66 years (IQR, 56-75 years); 1,031,999 (43.6%) were female and 1,003,154 (42.4%) were male. Exposures: ICI exposure in any reported drug role, individual primary-suspect drugs, and checkpoint-pathway exposure. Main Outcomes and Measures: Reporting odds ratios (ORs), cross-organ adverse-event communities, adjusted primary-suspect drug x ICI interaction ORs for Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN), drug reaction with eosinophilia and systemic symptoms (DRESS), acute generalized exanthematous pustulosis (AGEP), interstitial nephritis, drug-induced liver injury (DILI), and vomiting (VOM), and accelerated failure-time model time ratios for documented onset. Results: Of 3001 eligible Preferred Terms in cancer-associated reports, 2091 differed at a false discovery rate (FDR) less than .05. Four cross-organ toxicity communities were identified. Of 138 eligible drug-phenotype pairs, 65 had FDR-significant interactions, including moxifloxacin-SJS/TEN amplification (interaction OR, 101.72; 95% CI, 39.11-264.55), enfortumab vedotin-SJS/TEN attenuation (interaction OR, 0.17; 95% CI, 0.13-0.23), and omeprazole-interstitial nephritis amplification (interaction OR, 10.35; 95% CI, 7.62-14.05). Among 60,324 reports contributing to temporal analyses, ICI exposure was associated with longer adjusted documented time to onset for 5 of 6 phenotypes (time ratios, 1.37-1.59) but not AGEP (time ratio, 0.99; 95% CI, 0.67-1.46). Temporal associations also differed across checkpoint pathways. Conclusions and Relevance: ICIs were associated with a structured cross-organ toxicity landscape, phenotype-specific modification of drug-associated adverse events, and distinct temporal patterns across checkpoint pathways. These findings support checkpoint blockade as a modifier of drug-associated toxicity and motivate longitudinal and mechanistic validation.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361880v1?rss=1
• Can Dental AI Really Beat Dentists? DentalPair-Cert for Rigorous AI-Dentist Inference

  A dental AI system and a dentist reading the same radiographs form a paired comparison. Published comparative studies often report the two arms separately against a reference standard, leaving the joint pattern of correctness between them unavailable for secondary paired inference. We show what that omission costs. The accuracy difference remains exactly identified; its sampling variance does not, so the report contains the estimate and not its uncertainty. On a study of 282 units, two published accuracies are consistent with 38 distinct joint tables whose confidence intervals differ in width by a factor of 2.5. The consequence is a three-zone decision map rather than a single threshold: differences at or below 1.06 points are non-significant under every compatible table, differences at or above 6.03 points are significant under every compatible table, and in between the published numbers cannot decide. We then show the omission is repairable at negligible cost. One additional integer, the number of units both arms classify correctly, identifies the joint table exactly and restores standard paired inference. For a panel of readers the pairwise dependences must arise from one joint distribution, a constraint that binds once three readers are present; publishing each reader's joint-correct count against a single reference reader cannot widen and may tighten every pairwise bound, and in a 7-arm experiment reduced them by a median of 37% even for pairs excluding that reference. Where the integer was never published we give DentalPair-Cert, an interval with finite-sample coverage uniformly over every admissible within-unit AI-dentist dependence under the independent-sampling-unit model, certified in both the nuisance maximization and the inversion. Across 4,200,000 simulated comparisons an independence analysis falls to 74.5% coverage with 12.2% type-I error; in a purposive sample of 9 recent comparative studies, 1 reported a paired test on discordant units.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.09.01.26361874v1?rss=1
• Impact of stepwise dual antiplatelet therapy de-escalation in patients with multivessel disease undergoing drug-coated balloon angioplasty: insights from the REC-CAGEFREE II trial

  BACKGROUND The optimal antiplatelet regimen in patients with acute coronary syndrome (ACS) and multivessel disease undergoing drug-coated balloon (DCB) angioplasty remains unclear. METHODS This was a prespecified subgroup analysis of the REC-CAGEFREE II trial, which was conducted at 41 sites in China and randomized 1948 exclusively DCB-treated participants with ACS to stepwise dual antiplatelet therapy (DAPT) de-escalation or standard DAPT. The primary endpoint was net adverse clinical events (NACE; including all-cause death, stroke, myocardial infarction, revascularization, and BARC type 3 or 5 bleeding) at 12 months. Participants were stratified into multivessel and single-vessel subgroups according to angiographic characteristics. RESULTS Overall, 720/1948 (37.0%) patients had multivessel disease. The multivessel subgroup was associated with a significantly higher risk of NACE compared with the single-vessel subgroup (12.5% versus 6.7%, HR IPTW:1.84, 95%CI:1.35-2.51, P<0.001). No significant interaction was observed between vessel status (multivessel or single-vessel) and treatment allocation with respect to NACE (Pinteraction=0.542). In the multivessel subgroup, NACE occurred in 44/368 (12.1%) and 45/352 (12.9%) in the stepwise de-escalation and standard DAPT groups (HR IPTW:0.95, 95%CI:0.62-1.75, P=0.818), respectively. In the single-vessel subgroup, NACE occurred in 43/607 (7.1%) and 39/621 (6.3%) in the stepwise de-escalation and standard groups (HR IPTW:1.12, 95%CI:0.72-1.70, P=0.611), respectively. For the prespecified hierarchical secondary endpoint, win ratio analyses yielded more wins for stepwise de-escalation in both subgroups. CONCLUSIONS Among patients with ACS undergoing DCB-only angioplasty, those with multivessel disease were associated with a higher risk of NACE than those with single-vessel disease. Stepwise DAPT de-escalation and standard DAPT exhibited similar risk-benefit profiles in both subgroups.

  📄 ResearchSep 2, 2026https://www.medrxiv.org/content/10.64898/2026.08.31.26361869v1?rss=1