文献速递临床微生物质量管理与危急值

BD尿液培养人工智能辅助判读应用评估

Evaluation of the BD urine culture application for artificial intelligence-assisted urine culture interpretation.
PubMed 文献 · J Clin Microbiol · 2026-06-22

本研究在集中化微生物实验室评估BD Kiestra尿液培养应用(UCA)对1304份中段尿和留置导尿管尿标本的判读性能,以人工盲法读图为参照。在<10^6 CFU/L分诊阈值下总体一致率为95.9%,阳性与阴性符合率分别为89.9%和98.9%,30.5%培养可潜在自动放行或批量审核。UCA对大肠埃希菌推定鉴定PPA达98.9%,但肠球菌属PPV仅25.4%,提示需人工复核。

要点

  • UCA在<10^6 CFU/L分诊阈值下与人工读图一致率达95.9%。
  • 30.5%培养达到该阈值,可潜在自动放行或批量审核。
  • 肠球菌属推定鉴定PPV低(25.4%),需警惕假阳性。

核心结论

  • UCA在临床相关阈值下表现良好,可实质性减少人工读片负担。
  • 专家规则分类差异主要与纯度输出和推定鉴定有关,尤其是淡蓝色生长被误判为肠球菌属。

研究设计与样本

  • 在集中化微生物实验室连续纳入中段尿和留置导尿管尿标本,接种于BD BBL CHROMagar Orientation。
  • 共1304份培养,其中1251份可评估生长评分比较,1239份可评估专家规则。
  • 以盲法技术员图像读图作为 comparator。

主要发现

  • 四个生长类别(<10^5、10^5、10^6、≥10^7 CFU/L)的精确一致率为86.2%。
  • 在≥10^7 CFU/L纯培养或优势培养中,UCA对大肠埃希菌推定鉴定PPA为98.9%。
  • 肠球菌属推定鉴定PPA为89.5%,但PPV低(25.4%),反映频繁假阳性推定调用。

关键数值与推荐

  • 使用<10^6 CFU/L分诊阈值时,总体一致率95.9%,阳性符合率89.9%,阴性符合率98.9%。
  • 30.5%的培养达到该阈值,可潜在自动放行或批量审核。

临床怎么用

  • UCA可用于低生长或无生长培养的安全分诊,帮助技术员聚焦可能需进一步处理与临床干预的标本。
  • 对肠球菌属等推定鉴定结果需人工复核,尤其淡蓝色或模糊生长模式。

局限性

  • 多数差异反映总菌落计数输出与基于形态型的实验室定量惯例之间的差异,以及遗漏细微生长。
  • 专家规则类别差异主要与纯度输出和推定鉴定差异相关,淡蓝色生长被误判为肠球菌属是优先改进领域。
  • 研究强调本地验证和人工监督仍然必要。

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)。

原文

Artificial intelligence-based culture reading tools can potentially accelerate reading, improving reporting consistency in image-based interpretation, which is relevant for high-volume urine cultures. This study evaluated the BD Kiestra Urine Culture Application (UCA) in a centralized microbiology laboratory. Consecutive midstream and indwelling catheter urine specimens inoculated onto BD BBL CHROMagar Orientation were analyzed by the UCA. Growth score output at different reporting thresholds was evaluated, along with expert rule-based categories constructed using the app's growth score, purity score, and presumptive identification parameters, with blinded technologist image reads as the comparator. Of 1,304 cultures, 1,251 were evaluable for growth score comparison, and 1,239 for expert-rule assessment. The exact agreement across four growth categories (<105, 105, 106, ≥107 CFU/L) was 86.2%. Using a <106 CFU/L triage threshold, the overall agreement was 95.9%, with positive and negative percent agreement of 89.9% and 98.9%, respectively. In this dataset, 30.5% of cultures met this threshold for potential auto-release or batch-review. For presumptive identification in ≥107 CFU/L pure or predominant cultures, UCA performed best for E. coli (PPA 98.9%). For Enterococcus spp., PPA was 89.5%, but PPV was low (25.4%), reflecting frequent false-positive presumptive calls. UCA performed well at clinically relevant thresholds and could meaningfully reduce the burden of manual reading. Most discrepancies reflected differences between total colony count outputs and morphotype-based laboratory quantitation conventions, along with missed subtle growth. Expert-rule category discrepancies were primarily related to differences in purity output and presumptive identification, particularly fine light-blue growth being called presumptive Enterococcus spp., highlighting a priority area for improvement.IMPORTANCEUrine cultures are among the most frequently ordered culture tests in clinical microbiology. Timely, consistent plate interpretation impacts both patient care and laboratory capacity. Digital imaging of culture plates facilitates the application of artificial intelligence to urine culture specimens, but laboratories need clear evidence of where these tools perform reliably and where their outputs require caution. This study shows how an automated urine culture reading application can support safe triage of cultures with little or no growth, helping technologists focus on specimens that likely require workup and clinical action. At the same time, it highlights an important limitation: organism labeling can be vulnerable to look-alike growth patterns. In our dataset, fine or hazy, light-blue growth that was not typical of Enterococcus spp. was classified by the application as presumptive Enterococcus spp., a mismatch that could affect downstream decisions if not recognized. These findings reinforce that local validation and human oversight remain essential, even for common uropathogens, such as Escherichia coli.

来源:J Clin Microbiol,查看原文。版权归原作者所有。

引用

  1. van der Walt P, Naidu P, Avoledo S, et al. Evaluation of the BD urine culture application for artificial intelligence-assisted urine culture interpretation. J Clin Microbiol. 2026. PMID: 42324609. doi:10.1128/jcm.00270-26

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。