EUVIMEDEUVIMED entdecken
Crossref

DNN-PURE: A deep neural network approach to paper-based urea sensing

Souvik Biswas, Arijit Pal, Koel Chaudhury, Soumen Das

Biosensors and Bioelectronics · 2026-11 · 311 · 118980

DOI: 10.1016/j.bios.2026.118980 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Biosensors and Bioelectronics, 2026-11. Verfasst von Souvik Biswas und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Medical Science and Technology, Indian Institute of Technology Kharagpur, West Bengal, 721302, India; Department of Biomedical Engineering,…. Hauptthemen laut MEDLINE: Biosensing Techniques, Urea. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 37 Literaturangaben. Gefördert von: Department of Health Research, Ministry of Health and Family Welfare, Indian Council of Medical Research.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

A low-cost, disposable paper-based enzymatic sensor has been developed for the quantitative estimation of urea concentration in serum samples. The sensor utilizes a novel detection mechanism, where ion conduction through a localized sensing region is assessed to measure changes in impedance, enabling highly sensitive detection. The sensing platform is fabricated by functionalizing polyaniline (PANI) over a chromatography paper substrate, enhancing electrical conductivity and allowing covalent immobilization of urease via glutaraldehyde cross-linking. Additionally, material characterization via scanning electron microscopy (SEM), Raman spectroscopy, and X-ray photoelectron spectroscopy (XPS) validated the successful PANI deposition and enzyme immobilization. Impedimetric characterization demonstrates high sensitivity, achieving a clinically relevant detection range of 0.01 mg/ml to 1 mg/ml, with a limit of detection (LOD) of 0.01 mg/ml (0.166 mM), sensitivity of 6.977 units of normalized impedance change per mg/mL and a resolution of 0.04 mg/ml. To enhance resolution and minimize human error, a deep neural network (DNN) regression model was implemented, trained on impedimetric responses recorded at four key frequencies (100 Hz, 1 kHz, 10 kHz, and 100 kHz). The DNN model demonstrated superior predictive capabilities, achieving an accuracy exceeding 99%, significantly outperforming traditional manual calibration methods (94.71%) and improved resolution of 0.02 mg/ml. The robustness of the sensing platform was validated using Bland-Altman analysis, while spiked serum sample evaluations confirmed the sensor's efficacy in biologically relevant environments. The sensor is also tested with a clinical sample to verify its real-world application. The integration of deep learning not only enhances precision and resolution but also enables automated and real-time urea quantification, making this sensing platform a promising candidate for point-of-care diagnostics and biomedical applications.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Biosensing TechniquesUreaAniline CompoundsNeural Networks, ComputerHumansPaperLimit of DetectionUreaseEnzymes, ImmobilizedElectric ConductivityEquipment Design

Stichwörter der Autor:innen

Deep neural networkElectrical equivalent modelImpedimetric sensorPaper sensorPolyanilineUrea sensor

Substanzen

UreaAniline CompoundspolyanilineUreaseEnzymes, Immobilized

Autor:innen und Institutionen

  • School of Medical Science and Technology, Indian Institute of Technology Kharagpur, West Bengal, 721302, India; Department of Biomedical Engineering, National Institute of Technology Raipur, Chhattisgarh, 492010, India.
  • School of Medical Science and Technology, Indian Institute of Technology Kharagpur, West Bengal, 721302, India; Department of Engineering, University of Exeter, Exeter, EX4 4QJ, UK.
  • School of Medical Science and Technology, Indian Institute of Technology Kharagpur, West Bengal, 721302, India.
  • School of Medical Science and Technology, Indian Institute of Technology Kharagpur, West Bengal, 721302, India. Electronic address: sou@smst.iitkgp.ac.in.

Förderung

  • Department of Health Research
  • Ministry of Health and Family Welfare
  • Indian Council of Medical Research

Zitieren

APA

Biswas, S., Pal, A., Chaudhury, K., & Das, S. (2026). DNN-PURE: A deep neural network approach to paper-based urea sensing. Biosensors and Bioelectronics, 311, 118980. https://doi.org/10.1016/j.bios.2026.118980

Vancouver

Biswas S, Pal A, Chaudhury K, Das S. DNN-PURE: A deep neural network approach to paper-based urea sensing. Biosensors and Bioelectronics. 2026;311:118980. doi:10.1016/j.bios.2026.118980

BibTeX
@article{Biswas2026,
  author = {Biswas, Souvik and Pal, Arijit and Chaudhury, Koel and Das, Soumen},
  title = {DNN-PURE: A deep neural network approach to paper-based urea sensing},
  journal = {Biosensors and Bioelectronics},
  year = {2026},
  volume = {311},
  pages = {118980},
  doi = {10.1016/j.bios.2026.118980}
}
Kostenlos starten →