Truswell Alec, Abraham Rebecca, O'Dea Mark, Lee Zheng Zhou, Lee Terence, Laird Tanya, Blinco John, Kaplan Shai, Turnidge John, Trott Darren J, Jordan David, Abraham Sam
Antimicrobial Resistance and Infectious Diseases Laboratory, College of Science, Health, Engineering and Education, Murdoch University, Murdoch, Australia.
SciRobotics Ltd, Kefar Sava, Israel.
J Antimicrob Chemother. 2021 Jun 18;76(7):1800-1807. doi: 10.1093/jac/dkab107.
Surveillance of antimicrobial resistance (AMR) is critical to reducing its wide-reaching impact. Its reliance on sample size invites solutions to longstanding constraints regarding scalability. A robotic platform (RASP) was developed for high-throughput AMR surveillance in accordance with internationally recognized standards (CLSI and ISO 20776-1:2019) and validated through a series of experiments.
Experiment A compared RASP's ability to achieve consistent MICs with that of a human technician across eight replicates for four Escherichia coli isolates. Experiment B assessed RASP's agreement with human-performed MICs across 91 E. coli isolates with a diverse range of AMR profiles. Additionally, to demonstrate its real-world applicability, the RASP workflow was then applied to five faecal samples where a minimum of 47 E. coli per animal (239 total) were evaluated using an AMR indexing framework.
For each drug-rater-isolate combination in Experiment A, there was a clear consensus of the MIC and deviation from the consensus remained within one doubling dilution (the exception being gentamicin at two dilutions). Experiment B revealed a concordance correlation coefficient of 0.9670 (95% CI: 0.9670-0.9670) between the robot- and human-performed MICs. RASP's application to the five faecal samples highlighted the intra-animal diversity of gut commensal E. coli, identifying between five and nine unique isolate AMR phenotypes per sample.
While adhering to internationally accepted guidelines, RASP was superior in throughput, cost and data resolution when compared with an experienced human technician. Integration of robotics platforms in the microbiology laboratory is a necessary advancement for future One Health AMR endeavours.
对抗菌药物耐药性（AMR）的监测对于降低其广泛影响至关重要。由于其对样本量的依赖，需要解决长期以来在可扩展性方面的限制。开发了一个机器人平台（RASP），用于按照国际认可的标准（CLSI和ISO 20776-1:2019）进行高通量AMR监测，并通过一系列实验进行了验证。
实验A比较了RASP与人类技术人员在对四种大肠杆菌分离株进行八次重复实验时获得一致最低抑菌浓度（MIC）的能力。实验B评估了RASP与人类进行的MIC在91株具有不同AMR谱的大肠杆菌分离株中的一致性。此外，为了证明其在实际中的适用性，随后将RASP工作流程应用于五个粪便样本，使用AMR索引框架对每只动物至少47株大肠杆菌（共239株）进行了评估。
在实验A中，对于每种药物-评分者-分离株组合，MIC有明确的共识，与共识的偏差保持在一个稀释倍数内（庆大霉素在两种稀释度时除外）。实验B显示机器人和人类进行的MIC之间的一致性相关系数为0.9670（95%置信区间：0.9670-0.9670）。RASP应用于五个粪便样本突出了肠道共生大肠杆菌在动物体内的多样性，每个样本识别出五到九种独特的分离株AMR表型。
在遵循国际认可指南的同时，与经验丰富的人类技术人员相比，RASP在通量、成本和数据分辨率方面更具优势。将机器人平台整合到微生物实验室是未来“同一个健康”AMR努力的必要进步。
