Classification of asbestos and their nonasbestiform analogues using FTIR and multivariate data analysis

Our plain-English summary of Classification of asbestos and their nonasbestiform analogues using FTIR and multivariate data analysis, by T. Lee, S.E. Mischler, C. Wolfe (Journal of Hazardous Materials, 2024).

The study

• Authors: T. Lee, S.E. Mischler, C. Wolfe
• Journal: Journal of Hazardous Materials
• Published: 2024;469:133874
• Cited by: 7 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

This study presents a possible application of Fourier transform infrared (FTIR) spectrometry and multivariate data analysis, principal component analysis (PCA), and partial least squares-discriminant analysis (PLS-DA) for classifying asbestos and their nonasbestiform analogues. The objectives of the study are: 1) to classify six regulated asbestos types and 2) to classify between asbestos types and their nonasbestiform analogues. The respirable fraction of six regulated asbestos types and their nonasbestiform analogues were prepared in potassium bromide pellets and collected on polyvinyl chloride membrane filters for FTIR measurement. Both PCA and PLS-DA classified asbestos types and their nonasbestiform analogues on the score plots showed a very distinct clustering of samples between the serpentine (chrysotile) and amphibole groups. The PLS-DA model provided ∼95% correct prediction with a single asbestos type in the sample, although it did not provide all correct predictions for all the challenge samples due to their inherent complexity and the limited sample number. Further studies are necessary for a better prediction level in real samples and standardization of sampling and analysis procedures.

Published by Elsevier B.V.

In plain English

This laboratory study tested whether infrared spectroscopy (FTIR), combined with two statistical pattern-recognition methods (PCA and PLS-DA), could tell apart the six regulated asbestos types (actinolite, amosite, anthophyllite, chrysotile, crocidolite and tremolite) and separate them from their nonasbestiform analogues. It studied mineral reference samples, not people. The authors made respirable-size dust from each mineral and prepared 84 potassium bromide pellets and 120 PVC filter samples. Samples grouped clearly by mineral, with chrysotile furthest from the five amphiboles. Within each mineral, the asbestos form grouped apart from the nonasbestiform form. The Results section gave average classification errors of 0.007 for pellets and 0.046 for filters, while the Methods section gives 0.06 and 0.007. On 12 test samples of known make-up, the model's predicted compositions were not fully accurate. For example, a sample that was 100% nonasbestiform tremolite was predicted as 32% nonasbestiform tremolite and 20% tremolite. The authors concluded that classification is possible. They noted the small number of reference samples, the controlled laboratory conditions, and the need for further research and comparison with standard asbestos methods.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Laboratory method-development study using reference mineral samples, FTIR spectroscopy and multivariate data analysis (PCA and PLS-DA)
• Who was studied: Respirable-size samples of 12 reference minerals (six asbestos types and six nonasbestiform analogues): 84 potassium bromide (KBr) pellet samples (7 per mineral) and 120 PVC filter air samples (10 per mineral), prepared in a NIOSH laboratory in Pittsburgh, PA, USA
• Exposure: Asbestos types: actinolite, amosite, anthophyllite, chrysotile, crocidolite, tremolite. Nonasbestiform analogues: actinolite, grunerite, anthophyllite, antigorite, riebeckite, tremolite
• Main results:
  • In all clustering analyses, samples grouped distinctly by mineral, with chrysotile separating furthest from the five amphiboles
  • Within each mineral type, the asbestiform samples formed subclusters apart from the nonasbestiform samples
  • In the Results section, the cross-validated PLS-DA model for KBr pellet samples had an average classification error of 0.007 (a 0.7% chance of misclassifying a new sample)
  • In the Results section, PVC filter samples had an average classification error of 0.046. The authors interpreted this as a 95% probability of correctly classifying a pure sample that is one of the 6 mineral types in the model
  • The Methods section reports different cross-validation class-error averages for the PLS-DA models: 0.06 for KBr pellets and 0.007 for PVC filters
  • Challenge sample #1 (100% amosite) was predicted as 91% amosite, 2% nonasbestiform actinolite and 7% other minerals
  • For challenge sample #6 (100% nonasbestiform tremolite), the model failed to distinguish the asbestiform and nonasbestiform habits. It predicted 32% nonasbestiform tremolite, 20% tremolite, and smaller amounts of other minerals
  • Challenge sample #12 (49% tremolite, 51% nonasbestiform tremolite) was predicted as 25% tremolite, 29% nonasbestiform tremolite, 14% actinolite, 11% nonasbestiform actinolite, 8% crocidolite and 13% other minerals
  • Challenge sample #5 (100% nonasbestiform actinolite) was predicted as 35% actinolite and 46% nonasbestiform actinolite
  • Classification accuracy was better with KBr pellets than with PVC filters. Absorbance was more than about 2 times higher in KBr pellets than on PVC filters of similar sample mass
  • Mixed cellulose ester (MCE) filters were dropped from the study because they showed strong absorbance of their own
• The authors' conclusion: The authors concluded that classifying asbestos types with this method is possible, and that classifying asbestos versus its nonasbestiform analogues is also possible on PCA and PLS-DA score plots. They added that prediction in mixed samples needs improvement and that further research is needed to validate the method against current standard asbestos methods and on real samples
• Limitations the authors state:
  • The study was conducted under controlled laboratory conditions with a small number of well-characterized reference samples
  • The models were trained only on pure minerals, not on mixed samples, yet were used to predict mixed samples
  • The method does not provide morphology or fiber-size analysis, so additional microscope analysis would be needed
  • The study built a mass-based calibration, while current occupational exposure limits for asbestos are based on fiber number
  • Spectra were obtained only in transmission mode, which may carry uncertainty from pellet thickness
  • No well-characterized nonasbestiform reference materials are available for research

Topics

The National Library of Medicine indexes this study under subjects our site covers: Chrysotile.

Medical terms in this study, explained in our Lexicon: Non-Asbestiform, Regulated Asbestos, Respirable Fraction, Serpentine Group, Chrysotile, Amphibole Group.

Related studies

• Lung tumor cells: a multivariate approach to cell classification using two-dimensional protein pattern — H.R. Schmid, D. Schmitter, et al. (1995)
• Determination of low levels of free fibres of chrysotile in contaminated soils by X-ray diffraction and FTIR spectroscopy — E. Foresti, M. Gazzano, et al. (2003)
• Proteome screening of pleural effusions identifies galectin 1 as a diagnostic biomarker and highlights several prognostic biomarkers for malignant mesothelioma — F. Mundt, H.J. Johansson, et al. (2014)
• The Lost City Hydrothermal Field: A Spectroscopic and Astrobiological Analogue for Nili Fossae, Mars — E.S. Amador, J.L. Bandfield, et al. (2017)
• Assessment of the pathogenic potential of asbestiform vs. nonasbestiform particulates (cleavage fragments) in in vitro (cell or organ culture) models and bioassays — B.T. Mossman (2008)
• Identifying Thoracic Malignancies Through Pleural Fluid Biomarkers: A Predictive Multivariate Model — J.M. Porcel, A. Esquerda, et al. (2016)

Browse more research in Journal of Hazardous Materials, or return to medical research.