Abstract
The Magiscan 2 (M-2) image analysis system with asbestos fiber counting software was evaluated. The M-2 takes a video image from a standard phase contrast light microscope used for human (manual) counting of asbestos fibers, processes the video image and then analyzes the particle shapes to count the number of fibers. Operator attention is required to the extent of placing the slide in the microscope, selecting the areas for analysis and focusing the microscope. A set of Occupational Safety and Health Administration (OSHA) quality control field samples was analyzed by the M-2, and the results were compared to the OSHA counts. The data indicated good precision by the M-2 and reasonably good agreement with the OSHA counts. A subset of these samples also was analyzed by a group of 12 proficient laboratories. The M-2 results agreed well with the inter-laboratory means. Finally, the performance of the M-2 was observed for individual fibers. The M-2 did not always agree with the observer (e.g., sometimes missing the thinner fibers, sometimes breaking fibers apart and counting the segments as separate fibers), but was generally specific for fibers. It also was noted that chains of small particles and edges of certain types of large particles sometimes were detected as fibers. Thus, for samples containing such features, the M-2 may show significant biases. The M-2 has several advantages over manual counting of occupational asbestos exposure samples. The M-2 system gives precision at least as good as that of manual counters, is less prone to inter-laboratory bias, is somewhat faster than manual counters, reduces operator fatigue and requires less operator expertise.