Detecting Cardiovascular Diseases and Diabetes Mellitus Through Fingernails Using Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy and Machine Learning

Wilson, M, Al-Jumeily, D ORCID: 0000-0002-9170-0568, Tang, L, Birkett, J ORCID: 0000-0002-5682-512X, Khan, I ORCID: 0000-0002-4206-7663, Abbas, I and Assi, S ORCID: 0000-0002-5142-9179 (2025) Detecting Cardiovascular Diseases and Diabetes Mellitus Through Fingernails Using Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy and Machine Learning. Analytical Letters. pp. 1-20. ISSN 0003-2719

Preview

Text Detecting Cardiovascular Diseases and Diabetes Mellitus Through Fingernails Using Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy.pdf - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (1MB) | Preview

Abstract

Cardiovascular diseases (CVDs) and diabetes mellitus (DM) represent a global concern for the public and often result in severe medical and/or economic consequences. Traditional disease detection techniques such as blood work and cardiac catheterization are not only invasive and intrusive, but also require high expenses for equipment, facilities and training. This urges for the use of alternative methods such as infrared (IR) spectroscopy, which has shown great success in the detection of cancer, Fabry disease and kidney disease. Through the combination of machine learning algorithms (MLAs) and the alternative biological matrices (fingernails), infrared (IR) spectroscopy has the potential to replace traditional techniques, especially in low- and middle-income countries, where the economic and medical consequences created by CVDs and DM are at their highest. Thus, this work explored the detection of CVDs and DM through fingernails using attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy, chemometrics and MLAs such as correlation in wavenumber space (CWS), principal component analysis (PCAs) and self-organizing maps (SOMs). Spectral interpretation of fingernail spectra revealed the presence of key endogenous compounds such as amino acids, lipids and proteins, as well as disease-related compounds including glucose, high-density lipoproteins (HDLs) and homocysteine. Moreover, the applied MLAs demonstrated the feasibility of ATR-FTIR spectroscopy for the classification of healthy vs. diseased fingernails. Of these MLAs, PCA and SOM were more accurate than CWS where the latter showed high number of mismatches between fingernail spectra of varying health/disease status. However, PCA and SOM demonstrated more accurate clustering between healthy and diseased fingernails. This in turn confirmed that ATR-FTIR and MLAs were accurate in detecting CVDs and DM in fingernails non-destructively.

Item Type:	Article
Uncontrolled Keywords:	3401 Analytical Chemistry; 34 Chemical Sciences; Networking and Information Technology R&D (NITRD); Cardiovascular; Diabetes; Machine Learning and Artificial Intelligence; 3 Good Health and Well Being; 0301 Analytical Chemistry; 0399 Other Chemical Sciences; Analytical Chemistry; 3401 Analytical chemistry
Subjects:	R Medicine > RA Public aspects of medicine > RA1001 Forensic Medicine. Medical jurisprudence. Legal medicine
Divisions:	Pharmacy and Biomolecular Sciences
Publisher:	Taylor and Francis Group
Date of acceptance:	11 July 2025
Date of first compliant Open Access:	13 August 2025
Date Deposited:	13 Aug 2025 10:37
Last Modified:	13 Aug 2025 10:45
DOI or ID number:	10.1080/00032719.2025.2534604
URI:	https://researchonline.ljmu.ac.uk/id/eprint/26935

View Item