文献速递神经技术与安全

近期神经放射学研究中的证据质量与临床影响:对美国神经放射学杂志文章的系统评估

Quality of Evidence and Clinical Impact in Recent Neuroradiology Research: A Systematic Assessment of American Journal of Neuroradiology Articles.
PubMed 文献 · AJNR Am J Neuroradiol · 2026-07-01

该研究系统评估了2018-2019年和2021-2022年发表在AJNR上的1085篇原创文章,按证据水平(1-5级)和临床效能(1-6级)分类。结果显示平均证据水平为2.67,超过一半为EL-3,大多数研究集中于诊断准确性(CE-2),患者结局和社会影响研究较少。

要点

  • AJNR文章平均证据水平为2.67,54.5%为EL-3,整体证据水平中等。
  • 诊断准确性研究证据水平强于临床效用研究,且AI/软件类诊断研究证据水平更强。
  • 62%的研究为CE-2(评估诊断准确性),患者结局研究仅占11.4%。

核心结论

  • 近期AJNR文章证据水平中等,临床效能多集中于较低层级,主要关注诊断准确性。
  • 高证据水平设计及评估对管理、结局或成本效益影响的研究仍然有限。
  • 研究结果强调需要方法学严谨且具有临床影响力的神经放射学研究,以更贴近循证实践和效能/价值分析。

研究设计与样本

  • 系统评估,纳入2018-2019年和2021-2022年AJNR所有原创文章,共1085篇(2018-2019年592篇,2021-2022年493篇)。
  • 按证据水平(1-5级)和临床效能(1-6级)分类,分析研究类型、设计、队列选择及引用次数与EL和CE的关联,并比较两个时间段的趋势。

诊断标准与关键数值

  • 证据水平采用1-5级量表,临床效能采用1-6级量表。
  • 2018-2019年诊断准确性/监测研究EL为2.10±1.0,临床效用研究为2.91±0.7(P < .001)。
  • 2021-2022年诊断监测/准确性研究EL为1.73±0.8。
  • AI或软件类诊断研究EL为2.02±0.8,新序列或技术类为2.72±0.9(P < .001),改进现有框架类为2.52±1.0(P = .002)。
  • 平均EL从2018-2019年的2.75±0.81改善至2021-2022年的2.58±0.87(P < .001)。

临床应用与阅片要点

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

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

BACKGROUND AND PURPOSE: Evidence-based medicine frameworks in radiology emphasize rigorous study design and clinical applicability, yet the extent to which neuroradiology research aligns with these standards remains unclear. This study aimed to evaluate the methodologic quality and clinical relevance of recent American Journal of Neuroradiology (AJNR) publications by assessing their evidence level (EL) and clinical efficacy (CE).

MATERIALS AND METHODS: We systematically reviewed all original articles published in the AJNR in 2018-2019 and 2021-2022. Articles were classified by EL (scale 1-5) and by CE (scale 1-6). Study type (diagnostic, prognostic, therapeutic, economic), design (prospective versus retrospective), cohort selection (consecutive versus nonconsecutive), and citation count were analyzed for associations with EL and CE and trends were assessed by comparing articles between 2018-2019 and 2021-2022.

RESULTS: A total of 1085 articles (592 between 2018-2019 and 493 between 2021-2022) were included. The average EL was 2.67 (SD 0.85); more than one-half were EL-3 (54.5%). Between 2018-2019, most of articles were diagnostic (84%) and among them accuracy/monitoring studies had markedly stronger EL (2.10 ± 1.0) than clinical-utility studies (2.91 ± 0.7; P < .001). Stronger EL correlated with greater citation frequency (P < .001). Then in 2021-2022, diagnostic monitoring/accuracy studies again had the strongest evidence (1.73 ± 0.8). In addition, AI- or software-based diagnostic studies demonstrated significantly stronger EL (2.02 ± 0.8) than those introducing new sequences or techniques (2.72 ± 0.9; P < .001) or refining current frameworks (2.52 ± 1.0; P = .002). Average EL improved from 2.75 ± 0.81 to 2.58 ± 0.87 (P < .001), driven by more diagnostic-accuracy and AI/software-based papers, categories linked to stronger evidence. According to CE, the average score was 2.64 (SD: 1.17); 62% (670/1085) of studies were CE-2, evaluating diagnostic accuracy. Few studies addressed patient outcomes (11.4%), or societal impact (<1%), and the proportion remained stable during the study period.

CONCLUSIONS: Most recent AJNR publications demonstrate moderate evidence levels and are concentrated at lower clinical efficacy tiers, primarily focused on diagnostic accuracy. High-level study designs and research evaluating impact on management, outcomes, or cost-effectiveness remain limited. These findings underscore the need for methodologically rigorous and clinically impactful neuroradiology research that aligns more closely with evidence-based practice and effectiveness/value analysis.

来源:AJNR Am J Neuroradiol,查看原文。版权归原作者所有。

引用

  1. Ghozy S, Futela D, Singh A, et al. Quality of Evidence and Clinical Impact in Recent Neuroradiology Research: A Systematic Assessment of American Journal of Neuroradiology Articles. AJNR Am J Neuroradiol. 2026. PMID: 41962955. doi:10.3174/ajnr.A9193

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。