Rapid culture-free diagnosis of clinical pathogens via integrated microfluidic-Raman micro-spectroscopy

Q: What diagnostic performance did the culture-free platform achieve in clinical testing?

In a 305-patient clinical study using primary urine and other samples, it showed 95.4% agreement with traditional culture methods and 98.5% sensitivity for diagnosing infections.

Sep 30, 2026

Uploaded by Lucas Martin

Antimicrobial resistance (AMR) requires rapid, accurate diagnostics to guide timely antimicrobial therapy, but conventional culture-based testing is slow and struggles with low pathogen loads. This work introduces a culture-free platform combining microfluidics, Raman micro-spectroscopy, and deep learning to deliver sample-to-report results within 20 minutes. Microfluidic dialysis-dielectrophoresis enriches pathogens to detection limits below 2 CFU/ml. Using a 342-isolate single-cell Raman database and a 1DResNet model, the method achieves 95.1% lab accuracy. In a 305-patient clinical study, it shows 95.4% agreement with culture and 98.5% sensitivity, supporting next-generation AMR diagnostics.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages19

Reading Time48 min

AccessFree to download

Background: AMR diagnostic bottlenecks

Need for rapid, culture-free testing

Proposed platform

Microfluidic enrichment (dialysis-dielectrophoresis)

Raman micro-spectroscopy and single-cell fingerprints

Deep learning model (1DResNet)

Performance results

Lab accuracy from Raman database

Clinical validation vs traditional culture

Clinical impact and limitations

Need for broader validation for implementation

Article [https://doi.org/10.1038/s41467-025-66996-y](https://doi.org/10.1038/s41467-025-66996-y) Rapid culture-free diagnosis of clinical pathogens via integrated microﬂuidicRaman micro-spectroscopy Received: 14 December 2024 | Accepted: 19 November 2025 | | --- | | | | Check for updates | Yuetao Li 1,2,8, Jiabao Xu 1,3,8, Xiaofei Yi4,5, Xiaobo Li1, Yanjun Luo4, Andrew Glidle1, Phil Summersgill2, Simon Allen2, Tim Ryan2, Xiaochen Liu6, Wei Yu7, Xiaobing Chu7, Shiyu Chen7, Qian Zhang7, Xiaogang Xu 5, Xiaoting Hua 6, Qiwen Yang7, Julien Reboud 1, Yunsong Yu 6 , Wei E. Huang 3 , Jonathan M. Cooper 1 & Huabing Yin 1 Antimicrobial resistance (AMR) is a critical global health challenge, demanding rapid and accurate diagnostics to guide timely antimicrobial therapy. Current diagnosis is hindered by prolonged culturing and difﬁculties detecting low pathogen loads. Here, we present a culture-free diagnostic platform that integrates microﬂuidics, Raman micro-spectroscopy, and deep learning to deliver “sample-to-report” testing within 20 min. The microﬂuidic enrichment system employs dialysis-dielectrophoresis (DEP) technology to rapidly isolate pathogens directly from clinical samples with a detection limit as low as <2 colony forming unit (CFU)/ml. Combining a single-cell Raman ﬁngerprint database of 342 clinical isolates from 29 bacterial and 7 fungal species with a 1DResNet deep learning model, our approach achieved 95.1% accuracy in lab settings. Validated in a 305-patient clinical study involving primary urine and other clinical samples, it demonstrated 95.4% agreement with traditional culture methods and 98.5% sensitivity in diagnosing infections. While broader validation is needed for clinical implementation, the integrated, rapid diagnosis pipeline, as well as broad-spectrum detection, offer a promising solution for next-generation diagnostics for combating AMR. Antimicrobial resistance (AMR) has become one of the top threats to human health. Globally, deaths due to AMR are expected to rise from~5 million in 20191 to ~10 million by 2050 and the GDP loss is estimated to be around $100 trillion by 20502. While the development of AMR is inevitable due to natural selection, its rapid acceleration is primarily driven by the overuse and incorrect prescription of antibiotics. A signiﬁcant contribution to this crisis is the lack of rapid diagnostic tools, which slows timely identiﬁcation of microbial infections. Current diagnostic standards, require at least two days to provide results due to lengthy sample culture ( > 24 hours) and the need for pure isolates for pathogen identiﬁcation (e.g., MALDI-TOF mass spectrometry)3 and further sub-culturing for antibiotic susceptibility testing4. Thus, broad spectrum antibiotics are empirically administered to patients in the ﬁrst instance, which could lead to poor patient outcomes, and foster the development of antibiotic resistance5. 1James Watt School of Engineering, University of Glasgow, Glasgow, UK. 2Epigem Ltd, Redcar, UK. 3Department of Engineering Science, University of Oxford, OX1 3PJ, Oxford, UK. 4Shanghai D-band Medical Technology Co., LTD, Shanghai, P. R. China. 5Huashan Hospital, Fudan University, Shanghai, P. R. China. 6Department of Infectious Diseases, Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University, Hangzhou, China. 7Department of clinical laboratory, Peking Union Medical College Hospital, Peking Union Medical College, Beijing, P. R. China. 8These authors contributed equally: Yuetao Li, Jiabao Xu. e-mail: [yvys119@zju.edu.cn](yvys119@zju.edu.cn); [wei.huang@eng.ox.ac.uk](wei.huang@eng.ox.ac.uk); [Jon.Cooper@glasgow.ac.uk](Jon.Cooper@glasgow.ac.uk); [huabing.yin@glasgow.ac.uk](huabing.yin@glasgow.ac.uk) Diagnostic time is crucial for combating AMR infections, especially in life-threatening conditions like sepsis and pneumonia. To shorten turnaround time for diagnosing infections, molecular-based technologies such as polymerase chain reacti

FAQs

Why is culture-based diagnosis insufficient for AMR management?

Culture-based methods take at least two days due to prolonged culturing and require pure isolates for identification and susceptibility testing. This delay forces empiric broad-spectrum antibiotic use and can worsen outcomes and resistance development.

How does the integrated microfluidic system isolate pathogens from clinical samples?

It uses a dialysis-dielectrophoresis (DEP) microfluidic enrichment approach to rapidly isolate pathogens directly from samples, achieving detection limits as low as <2 CFU/ml.

What diagnostic performance did the culture-free platform achieve in clinical testing?

In a 305-patient clinical study using primary urine and other samples, it showed 95.4% agreement with traditional culture methods and 98.5% sensitivity for diagnosing infections.

Sep 30, 2026

Uploaded by Lucas Martin