Fast Screening of Tuberculosis Patients Based on Analysis of Plasma by Infrared Spectroscopy Coupled with Machine Learning Approaches DOI Creative Commons
Lin Mei, Hsiao‐Chi Lu, Hui‐Wen Lin

et al.

ACS Omega, Journal Year: 2025, Volume and Issue: unknown

Published: March 20, 2025

Prompt diagnosis of tuberculosis (TB) enables timely treatment, limiting spread and improving public health for this disease. Currently, a rapid, sensitive, accurate, cost-effective detection TB still remains challenge. For purpose, we engaged transmission skill an attenuated total reflectance (ATR) technique coupled with Fourier-transform infrared spectrometry (FTIR) to study the IR spectra plasma samples from patients (n = 10) healthy individuals 10). To ensure high-quality spectral data, were collected in both ATR modes, each measurement consisting 256 scans at resolution 8 cm–1. mode, measurements repeated five times per sample, while ATR-FTIR three sample. These parameters carefully optimized through rigorous testing achieve highest possible signal-to-noise ratio patient sample analysis. Using method, obtained 100 20 mode 60 ensuring sufficient data robust Further, applied machine learning techniques analyze classify spectra; by means, differentiated those between ones. In work, modified transmission-FTIR setup improve absorption sensitivity focusing light on interface sample; while, used high-refractive-index crystal ZnSe as medium reflect signals scheme. Routinely, compared methods; their second derivative curves, notified that there had distinct differences protein lipid regions (3500–3000, 2900–2800, 1700–1500 cm–1) groups. classifiers─Logistic Regression (LR), Random Forest (RF), XGBoost (Xg)─we found Xg achieved accuracy 0.749, precision 0.703, recall 0.901, F1 score 0.790, AUC ROC curve 0.82 3500–2700 cm–1 region; additionally, practice showed possessed performance ∼ 80% accuracy. We randomly assigned participants (rather than individual scans) training 20% test sets train validate models (LR, RF, Xg). Based results, concluded spectroscopic method demonstrated its superior diagnosis. Thus, have absorption-FTIR spectroscopy is valuable tool sorting disease patients. The analysis plasmas can complement clinical evidence provides rapid accurate clinic.

Language: Английский

Automated Diagnosis and Phenotyping of Tuberculosis Using Serum Metabolic Fingerprints DOI Creative Commons
Yajing Liu, Ruimin Wang, Chao Zhang

et al.

Advanced Science, Journal Year: 2024, Volume and Issue: 11(39)

Published: Aug. 19, 2024

Abstract Tuberculosis (TB) stands as the second most fatal infectious disease after COVID‐19, effective treatment of which depends on accurate diagnosis and phenotyping. Metabolomics provides valuable insights into identification differential metabolites for However, TB phenotyping remain great challenges due to lack a satisfactory metabolic approach. Here, metabolomics‐based diagnostic method rapid detection is reported. Serum fingerprints are examined via an automated nanoparticle‐enhanced laser desorption/ionization mass spectrometry platform outstanding by its speed (measured in seconds), minimal sample consumption (in nanoliters), cost‐effectiveness (approximately $3). A panel 14 m z −1 features identified biomarkers 4 Based acquired biomarkers, models constructed through advanced machine learning algorithms. The robust model yields 97.8% (95% confidence interval (CI), 0.964‐0.986) area under curve (AUC) 85.7% CI, 0.806‐0.891) AUC In this study, serum biomarker panels revealed develop tool with desirable performance phenotyping, may expedite implementation end‐TB strategy.

Language: Английский

Citations

4
Porous PtCu Alloys Decode Plasma Metabolic Fingerprints for the Recognition of Severe Community‐Acquired Pneumonia DOI Open Access
Kexin Meng, Yao Shen, Dongni Hou

et al.

Advanced Healthcare Materials, Journal Year: 2025, Volume and Issue: unknown

Published: March 3, 2025

Abstract Rapid and accurate recognition of severe community‐acquired pneumonia (CAP) would facilitate the optimal intervention. Currently, diagnosis CAP is commonly based on criteria established by Infectious Disease Society America (IDSA)/American Thoracic (ATS), which include 2 primary 9 secondary criteria, making process cumbersome time‐consuming. Here, a porous PtCu alloy‐assisted laser desorption/ionization mass spectrometry (LDI MS) designed for extraction plasma metabolic fingerprints (PMFs), coupling with machine learning CAP. The alloys particle size exhibit excellent sensitivity, reproducibility, universality metabolite detection, due to structure, promising photoelectric effect, improved melting surface structure. Further, nanoplatform successfully records PMFs within seconds, using only 0.5 µL native plasma. Machine 69 individuals produces diagnostic model an area under curve (AUC) 0.832. Particularly, three biomarker panel demonstrates enhanced efficiency (AUC 0.846), outperforming reported biomarkers 0.560–0.770). Notably, can be completed in ≈35 min. work affords rapid precise method management through analysis.

Language: Английский

Citations

0
Single Test‐Based Diagnosis and Subtyping of Pulmonary Hypertension Caused by Fibrosing Mediastinitis Using Plasma Metabolic Analysis DOI Creative Commons
Yating Zhao,

Chunmeng Ding,

Hongling Su

et al.

Advanced Science, Journal Year: 2025, Volume and Issue: unknown

Published: March 6, 2025

Pulmonary hypertension (PH) often leads to poor survival outcomes and encompasses diverse subtypes with distinct underlying causes. Specifically, PH resulting from fibrosing mediastinitis (FM-PH) presents significant diagnostic challenges due nonspecific symptoms overlap of clinical characterization other subtypes, leading frequent misdiagnosis delayed treatment. Moreover, the complex procedures impose a burden on FM-PH patients, many whom already experience mobility difficulties. This study represents single test-based diagnosis FM-PH, using plasma metabolites obtained through ferric particle-enhanced laser desorption/ionization mass spectrometry analysis. Distinct metabolic alterations in are identified compared healthy controls achieving an area under curve (AUC) 0.987 for 0.728 differentiating subtypes. By addressing existing gaps strategies, this research highlights potential analysis elucidating landscape PH.

Language: Английский

Citations

0
Fast Screening of Tuberculosis Patients Based on Analysis of Plasma by Infrared Spectroscopy Coupled with Machine Learning Approaches DOI Creative Commons
Lin Mei, Hsiao‐Chi Lu, Hui‐Wen Lin

et al.

ACS Omega, Journal Year: 2025, Volume and Issue: unknown

Published: March 20, 2025

Prompt diagnosis of tuberculosis (TB) enables timely treatment, limiting spread and improving public health for this disease. Currently, a rapid, sensitive, accurate, cost-effective detection TB still remains challenge. For purpose, we engaged transmission skill an attenuated total reflectance (ATR) technique coupled with Fourier-transform infrared spectrometry (FTIR) to study the IR spectra plasma samples from patients (n = 10) healthy individuals 10). To ensure high-quality spectral data, were collected in both ATR modes, each measurement consisting 256 scans at resolution 8 cm–1. mode, measurements repeated five times per sample, while ATR-FTIR three sample. These parameters carefully optimized through rigorous testing achieve highest possible signal-to-noise ratio patient sample analysis. Using method, obtained 100 20 mode 60 ensuring sufficient data robust Further, applied machine learning techniques analyze classify spectra; by means, differentiated those between ones. In work, modified transmission-FTIR setup improve absorption sensitivity focusing light on interface sample; while, used high-refractive-index crystal ZnSe as medium reflect signals scheme. Routinely, compared methods; their second derivative curves, notified that there had distinct differences protein lipid regions (3500–3000, 2900–2800, 1700–1500 cm–1) groups. classifiers─Logistic Regression (LR), Random Forest (RF), XGBoost (Xg)─we found Xg achieved accuracy 0.749, precision 0.703, recall 0.901, F1 score 0.790, AUC ROC curve 0.82 3500–2700 cm–1 region; additionally, practice showed possessed performance ∼ 80% accuracy. We randomly assigned participants (rather than individual scans) training 20% test sets train validate models (LR, RF, Xg). Based results, concluded spectroscopic method demonstrated its superior diagnosis. Thus, have absorption-FTIR spectroscopy is valuable tool sorting disease patients. The analysis plasmas can complement clinical evidence provides rapid accurate clinic.

Language: Английский

Citations

0