Experiences with observer variation in mesothelioma panels.

The voting system used in the Dutch Mesothelioma Panel and in those of the European Communities and the IARC in the UK to estimate the probability of the histopathological diagnosis of malignant mesothelioma is described. Five categories are used: A: definite B: probable C: possible D: improbable E: definitely not The voting system facilitates measurement of probability and comparison of opinions, and good or reasonable agreement between panel members is achieved. Second viewing of the slides usually increases agreement. The percentages of dubious diagnoses are less when sufficiently large amounts of tissue are available.