颈动脉斑块RADS超声应用的观察者一致性及学习曲线
本研究首次评估斑块RADS分类系统在颈动脉超声中的适用性,纳入302例受检者共604张图像。三位不同年资超声医师独立分类并3个月后重复评估,观察者间κ为0.670至0.804,观察者内κ为0.804至0.901。结果显示该系统经培训可被不同年资医师有效使用,适用于斑块标准化评估与卒中风险分层。
要点
- 斑块RADS在超声中观察者间及观察者内一致性均较高,可用于标准化报告。
- 低年资医师与高年资医师一致性随病例经验增加而稳步提升,需重视系统培训。
- 该分类可用于颈动脉斑块相关卒中风险分层,报告时建议注明RADS分级。
核心结论
- 斑块RADS分类系统在颈动脉超声中表现出高度适用性,经培训后可被不同年资医师有效使用。
- 该系统为颈动脉斑块标准化评估和卒中风险分层提供了可靠工具。
研究设计与样本
- 单中心研究,连续性纳入302例受检者共604张颈动脉超声图像。
- 三位不同年资超声医师独立依据斑块RADS系统分类以评估观察者间一致性。
- 3个月后进行第二次评估用于观察者内一致性分析,并按连续病例组评估学习曲线。
关键影像表现/征象
- 观察者间一致性方面,专家与副主任医师间近乎完美一致,κ=0.804(95% CI: 0.765-0.842)。
- 专家与中级医师间κ=0.754(95% CI: 0.711-0.793),副主任医师与中级医师间κ=0.670(95% CI: 0.620-0.713)。
- 三位医师两次评估的观察者内κ分别为0.901、0.843和0.804。
诊断标准与关键数值
- 专家与副主任医师观察者间κ=0.804(95% CI: 0.765-0.842),为近乎完美一致。
- 专家与中级医师观察者间κ=0.754(95% CI: 0.711-0.793)。
- 副主任医师与中级医师观察者间κ=0.670(95% CI: 0.620-0.713)。
- 观察者内κ:专家0.901(95% CI: 0.872-0.930),副主任医师0.843(95% CI: 0.808-0.878),中级医师0.804(95% CI: 0.766-0.843)。
临床应用与阅片要点
- 不同年资医师使用斑块RADS分类颈动脉斑块时具有较高可重复性,适合推广到日常超声报告。
- 低年资医生初期一致性低于高年资医生,提示开展该分类时应配合系统培训。
- 学习曲线显示增加病例经验可逐步提高与专家的一致性,报告质量会随培训改善。
局限性
- 研究为单中心设计,结果能否推广到其他机构尚需验证。
- 三位观察者均来自同一单位,观察者间高度一致可能受共同阅片习惯影响。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVES: This study represents the first evaluation of the Plaque-RADS (Reporting and Data System) classification system's applicability in carotid ultrasound examinations. It focuses on analyzing inter-observer and intra-observer agreement among physicians of varying experience levels while investigating its learning curve to validate the system's suitability and stability within the ultrasound field.
METHODS: This single-center study consecutively enrolled 604 carotid ultrasound images from 302 examinees. Three sonographers with varying levels of experience (expert, associate senior physician, and intermediate physician) independently classified all images using the Plaque-RADS system to assess inter-observer agreement. A second assessment was conducted 3 months after the initial evaluation to examine intra-observer agreement. Moreover, learning curves for each physician level were evaluated using 5 consecutive case groups.
RESULTS: Regarding inter-observer agreement, near-perfect agreement was observed between the expert and associate senior physician (κ = 0.804, 95% CI: 0.765-0.842). Agreement between the expert and intermediate physician (κ = 0.754, 95% CI: 0.711-0.793) and between the associate senior and intermediate physician (κ = 0.670, 95% CI: 0.620-0.713) was also statistically significant. Intra-observer agreement demonstrated good stability across 2 assessments for all 3 physicians, with κ values of 0.901 (95% CI: 0.872-0.930) (expert), 0.843 (95% CI: 0.808-0.878) (associate senior physician), and 0.804 (95% CI: 0.766-0.843) (intermediate physician). Learning curve analysis revealed that both associate senior physician and intermediate physician steadily improved their agreement with the expert as case experience increased.
CONCLUSION: Research findings indicate that the Plaque-RADS classification system demonstrates high applicability in the field of ultrasound. Through training, this system can be effectively utilized by physicians of varying skill levels, providing a reliable tool for standardized assessment of carotid plaque and stroke risk stratification.
引用
- Dai, Mei, Wang, et al. Application of the Carotid Plaque-RADS Classification System in Ultrasound: Inter- and Intra-Observer Agreement Analysis and Learning Curve Analysis. J Ultrasound Med. 2026. PMID: 42247322. doi:10.1002/jum.70311
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。