📄 论文动态
本期论文以AI/LLM在医疗场景的应用为主(9篇),涵盖ICU死亡率预测、内镜诊断、超声报告生成、病历术语识别、谵妄筛查等,同时涉及心力衰竭药物(AC01)、RAAS系统靶向、白血病药物敏感性及免疫分析技术等基础与临床转化研究。
### 🏭 行业动态
行业动态聚焦FDA对AI医疗设备的监管新规,包括简化审批流程、风险-获益报告要求及AHA反馈意见;同时多份市场报告显示AI医疗市场高速增长(CAGR约37%-38%),预计2035年规模超700亿美元。
### 📊 高频期刊 TOP10
| 期刊 | IF | 篇数 |
|:---|---:|---:|
| J Med Internet Res | 5.8 | 4 |
| NPJ Digit Med | 12.4 | 2 |
| J Ethnopharmacol | 4.8 | 2 |
| JMIR Med Inform | 3.1 | 2 |
| Comput Biol Med | 6.5 | 1 |
| Value Health | 4.0 | 1 |
| Lancet | 98.4 | 1 |
| Biotechnol Adv | 12.1 | 1 |
| Circ Res | 20.1 | 1 |
| Clin Cancer Res | 10.0 | 1 |
## 二、研究主题聚类
### LLM与AI在临床辅助决策与影像中的应用(9篇)
**趋势:** LLM从概念验证向临床部署加速,多模态、医学知识增强成为提升可靠性的关键方向
### 心血管与肿瘤药物研发(3篇)
**趋势:** 心力衰竭口服ghrelin受体激动剂进入早期临床,BH3模拟物在白血病亚型中显示差异化敏感性
### FDA监管与市场趋势(5篇)
**趋势:** 监管框架趋于简化与透明,AI医疗器械认证数量快速上升,市场资本持续涌入
## 三、重点推荐论文
### Safety, pharmacokinetics, and exploratory efficacy of the oral ghrelin receptor agonist AC01 in heart failure with reduced ejection fraction (GOAL-HF1): a randomised, double-blind, placebo-controlled, phase 1b/2a study.
**Lancet** (IF:98.4) | [PubMed ⚠️](https://pubmed.ncbi.nlm.nih.gov/42341796/) | [DOI](https://doi.org/10.1016/S0140-6736(26)00904-9)
**推荐理由:** 发表于Lancet(IF 98.4),首个口服ghrelin受体激动剂治疗HFrEF的1b/2a期试验,具有转化意义
### Bootstrapping multimodal large language model with medical knowledge for automatic esophagogastroduodenoscopy diagnosis and reporting.
**Nat Commun** (IF:14.7) | [PubMed ⚠️](https://pubmed.ncbi.nlm.nih.gov/42469206/) | [DOI](https://doi.org/10.1038/s41467-026-75377-y)
**推荐理由:** 发表于Nat Commun(IF 14.7),展示多模态LLM在消化内镜诊断与报告中的自动化潜力,方法创新性强
### Enhancing Large Language Models for Identifying and Prioritizing Important Medical Jargons From Electronic Health Record Notes Using Data Augmentation: Comparative Study.
**JMIR AI** (IF:5.8) | [PubMed ⚠️](https://pubmed.ncbi.nlm.nih.gov/42467970/) | [DOI](https://doi.org/10.2196/75561)
**推荐理由:** 系统评估数据增强对LLM识别医疗术语的作用,对电子病历信息提取有直接实用价值
### Development and Clinical Evaluation of a Large Language Model-Based System for Generating Patient-Friendly Echocardiography Reports: Two-Stage Retrospective Validation and Prospective Survey Study.
**J Med Internet Res** (IF:5.8) | [PubMed ⚠️](https://pubmed.ncbi.nlm.nih.gov/42462074/) | [DOI](https://doi.org/10.2196/97136)
**推荐理由:** 结合回顾验证与前瞻调查,证明LLM生成易于理解的超声报告可行,临床应用前景明确
## 四、行业动态
### Benefit-Risk Reporting for FDA-Cleared Artificial Intelligence−Enabled Medical Devices
## and machine learning (ML) are increasingly available for diagnosis and management in clinical areas including cancer, cardiology, and neurology.1,2 Both the US Food and Drug Administration (FDA) a