Asbestos fibre counting by image analysis--the performance of the Manchester Asbestos Program on Magiscan

Our plain-English summary of Asbestos fibre counting by image analysis--the performance of the Manchester Asbestos Program on Magiscan, by L.C. Kenny (The Annals of Occupational Hygiene, 1984).

The study

• Author: L.C. Kenny
• Journal: The Annals of Occupational Hygiene
• Published: 1984;28(4):401-15
• Cited by: 8 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Manchester University's Asbestos fibre-counting Program (MAP) now performs satisfactorily on the Magiscan image-analyser for most industrial samples. Further work to develop the software is unlikely to produce proportionate improvement in performance. From the available data, MAP counts are not distinguishable within the range of results from visual-counting laboratories either internationally, when using the European Reference Method (ERM), or nationally, using in-house methods or the ERM. When compared with those visual counters employed by the Health and Safety Executive (HSE) who are trained and practised in the ERM, MAP results are not distinguishable at concentrations above approximately 0.3 fibres ml-1. MAP counts relatively high at concentrations below 0.3 fibres ml-1, but is still within the confidence limits established from quality control with HSE's in-house method. MAP performance on routine samples taken after delagging operations has not yet been investigated fully. HSE are now using the MAP in conjunction with Magiscan for routine monitoring of asbestos concentrations, taking these factors into account.

PubMed has no abstract for this study. This is the authors' abstract, copied from the article.

In plain English

This paper, from the UK Health and Safety Executive (HSE), evaluated how well the Manchester Asbestos Program (MAP), software running on a Magiscan image analyser, counted airborne asbestos fibres on membrane filter slides compared with human microscope counters. On eight slides circulated among fourteen laboratories in Europe and North America, MAP counts fell inside the visual range on every slide it counted except one, a UICC chamber amosite sample. In a UK exchange scheme, forty of eighty visual-counting laboratories counted below MAP and forty above. HSE also collected factory samples from asbestos textiles (175 slides), friction products (136) and asbestos cement (90). On these, MAP counted higher than HSE's Cricklewood visual counters at low densities. Above about 0.3 fibres per millilitre the two were indistinguishable. The author concluded that MAP may be regarded as equivalent to the European Reference Method for industrial samples. The author noted that confidence limits for that method had only been estimated, and that performance on routine delagging samples had not been fully investigated.

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: Methods evaluation comparing an automated image-analysis fibre-counting program with visual (microscope) counts, using international and national interlaboratory exchanges and HSE's own routine samples
• Who was studied: Slides of airborne asbestos samples. Eight slides were counted in an international exchange of 14 laboratories in Europe and North America. Slides were also counted in the UK RICE exchange scheme, with 80 visual-counting laboratories in its second round. HSE factory samples collected from early 1982: 175 slides from asbestos textiles, 136 from friction products and 90 from asbestos cement.
• Exposure: Airborne asbestos fibre samples on membrane filters, mostly from industrial sources. One international slide was a UICC chamber amosite sample used for animal exposure studies. The comparison was between counting by the MAP/Magiscan system and visual counting, mainly by the European Reference Method (ERM).
• Main results:
  • International exchange: MAP counts were inside the range of the 14 visual laboratories on all slides except slide 1, the UICC amosite sample (MAP 212 against a visual range of 327-686).
  • International exchange: MAP was inside the range of the three laboratories experienced in the ERM on five of the seven slides it counted, which was better than HSE's visual counters at that time.
  • UK RICE exchange, second round: of eighty visual-counting laboratories, forty had mean counting levels below MAP and forty above. There was a factor of ten in the laboratories' mean counting levels.
  • Different Magiscan systems agreed closely. The median ratio of the T&N Magiscan to the HSE Magiscan was 1.1, with 75% of ratios for 102 individual slides between 0.74 and 1.6. For Manchester University's Magiscan, the median ratio was 1.0, with 75% of ratios between 0.75 and 1.3.
  • On RICE slides, MAP generally undercounted on dense slides from non-industrial sources. It overcounted on very sparse slides from all sources.
  • On HSE samples from all three industries, the relationship was curvilinear. MAP counted higher at low densities and came close to the visual counts around 0.3 to 1.0 fibres per millilitre. More than 90% of points fell within the 90% confidence limits from HSE's 1978-1982 quality-control data.
  • Against the estimated ERM confidence limits, MAP was indistinguishable from the Cricklewood team's visual counts above about 0.3 fibres per millilitre. Below that, a substantial portion of MAP results were above the estimated upper 90% limit.
  • Agreement of individual HSE counters with MAP, measured as the share of count ratios between 0.5 and 2.0. Textiles: P 82% (17 slides), R 100% (24), M 88% (24). Friction products: P 78% (18), R 89% (18), M 85% (20). All three counters did better on textile samples.
• The authors' conclusion: The author concluded that, on industrial samples, MAP counts were indistinguishable from ERM counts by a typical counter in a competent laboratory, and that MAP may be regarded as an equivalent method to the ERM for samples from industrial sources. HSE integrated MAP into its visual-counting team from 1 January 1984.
• Limitations the authors state:
  • No ERM quality-control data were yet available, so the ERM confidence limits were only estimated.
  • It could not be determined whether the non-linearity at low densities was due to the visual counts, the Magiscan counts, or both.
  • Each HSE sample was counted by only one visual counter.
  • MAP performance on routine samples taken after delagging operations had not yet been fully investigated.
  • Later results showed differences between Magiscan systems on some types of sample, which appeared to be mainly due to the microscope used.

Related studies

• A summary of asbestos fibre counting experience in seven countries — G.W. Gibbs, P. Baron, et al. (1977)
• A comparison of fibre counting stopping rules in the application of the membrane filter method — N.P. Crawford, W.M. Maclaren, et al. (1987)
• A proposal for harmonising laboratory performance assessment criteria in national asbestos fibre counting schemes — M.C. Arroyo, J.M. Rojo (2001)
• Performance of asbestos fibre counting laboratories by transmission electron microscopy — B. Préat (2001)
• Comparison of NIOSH and AIA methods for evaluating asbestos fibres: effects of asbestos type, mounting medium, graticule type and counting rules — E. Gonzalez-Fernandez, F.R. Martín (1986)
• A comparison of fibre counting across three European national proficiency testing schemes — A.D. Jones, M.C. Arroyo, et al. (2005)

Browse more research in The Annals of Occupational Hygiene, or return to medical research.