文献速递泌尿生殖腹部技术与安全CT

基于机器学习的尿路结石检测诊断效能:系统综述与Meta分析

The diagnostic performance of machine learning based detection of urinary tract stones: a systematic review and meta-analysis.
PubMed 文献 · Eur J Radiol · 2026-04-04

该系统综述与双变量Meta分析纳入45项研究,其中24项(49,277张测试图像)进入定量合并。NCCT(10项研究)合并敏感度96%、特异度98%;外部验证NCCT队列(4项研究,1,056张图像)合并敏感度95%、特异度96%。ML模型在NCCT上检测结石具有高图像级诊断效能,但异质性和外部验证不足限制其转化。

要点

  • NCCT上ML检测结石合并敏感度96%、特异度98%,外部验证队列敏感度95%、特异度96%。
  • 结石部位是异质性来源之一(p=0.014),其他调节因素不显著。
  • 外部验证研究仅4项、1,056张图像,转化证据仍有限。

核心结论

  • ML模型在NCCT上检测尿路结石具有高图像级诊断效能,可作为放射科医生的决策支持工具。
  • 异质性和外部验证有限制约其临床转化,未来研究应超越单纯检测任务,转向对放射科及泌尿科、肾内科临床可操作的输出。

研究设计与样本

  • 系统综述与双变量Meta分析,前瞻性注册于PROSPERO(CRD42024542409),采用QUADAS-2评估偏倚风险。
  • 定性纳入45项研究;24项研究(49,277张测试图像)提供可提取的2×2数据进入Meta分析。
  • 亚组分析按模型类型、分类任务、结石部位、数据集来源和CT方位进行;双变量Meta回归探索异质性;Deeks检验评估发表偏倚。

诊断标准与关键数值

  • NCCT(10项研究)ML检测尿路结石合并敏感度96%(95%CI 92-98%)、合并特异度98%(95%CI 97-99%)。
  • 外部验证NCCT队列(4项研究,1,056张图像)合并敏感度95%(95%CI 92-97%)、合并特异度96%(95%CI 70-100%)。
  • Meta回归显示结石部位对异质性有贡献(p=0.014),其他调节因素不显著。
  • Deeks检验未见小样本效应(p=0.571)。

临床应用与阅片要点

  • ML在NCCT上检测结石的高敏感度和特异度支持其作为阅片决策支持工具,但解读外部验证结果时需注意特异度置信区间宽(70-100%)。
  • 结石部位可影响模型效能,阅片时对不同部位结石的ML提示应结合人工判读复核。

局限性

  • 外部验证研究数量有限(仅4项,1,056张图像),限制结论的泛化性。
  • 研究间异质性持续存在,Meta回归仅结石部位显著,其他调节因素未能解释全部变异。

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: Urolithiasis is a prevalent urological condition, and Non-Contrast Computed Tomography (NCCT) is the gold standard for diagnosis. In recent years, there has been growing interest in investigating machine learning (ML)- based detection of urolithiasis and the wider potential of AI in urology.

PURPOSE: To synthesise the diagnostic accuracy of ML-based UTS detection on NCCT and in externally validated cohorts.

MATERIALS AND METHODS: We performed a systematic review and bivariate meta-analysis of studies evaluating ML for detecting urinary stones. We used QUADAS-2 to assess the risk of bias. Subgroup analyses examined performance by model type, classification task, stone site, dataset source, and CT orientation. Bivariate meta-regression was performed to further explore heterogeneity. Publication bias was assessed using Deeks' test. The study was prospectively registered in Prospero (CRD42024542409).

RESULTS: Forty-five studies were included qualitatively. 24 studies (49,277 test images) provided extractable 2 × 2 data for meta-analysis. For NCCT (10 studies), pooled sensitivity was 96% (95% CI 92-98%) and pooled specificity was 98% (95% CI 97-99%). In externally validated NCCT cohorts (4 studies; 1,056 images), pooled sensitivity was 95% (95% CI 92-97%) and pooled specificity was 96% (95% CI 70-100%). Subgroup performance remained high, but heterogeneity persisted; meta-regression found stone site contributed to variability (p = 0.014), while other moderators were not significant. Deeks' test showed no small-study effects (p = 0.571).

CONCLUSION: ML models show high image-level diagnostic performance for stone detection on NCCT and may support radiologists as decision support tools. Translation is limited by heterogeneity and limited external validation. Future studies should move beyond detection-alone tasks towards clinically meaningful outputs that are actionable for radiologists and downstream clinicians, including urologists and nephrologists.

来源:Eur J Radiol,查看原文。版权归原作者所有。

引用

  1. Hajalamin M, Msherghi A, Alhammadi I, et al. The diagnostic performance of machine learning based detection of urinary tract stones: a systematic review and meta-analysis. Eur J Radiol. 2026. PMID: 41950864. doi:10.1016/j.ejrad.2026.112836

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。