Preclinical studies generating semiquantitative histopathology data are welcomed by regulatory agencies because, theoretically, reproducible and comparable results would be obtained with independent observers. But is that necessarily the case?
Many factors influence the interpretation and reproducibility of semiquantitative data. Histopathology results are dependent on establishing appropriate scoring criteria. Semiquantitative scoring, with categorical ranking of observed changes, is also subject to interpretation limitations due to the relatively broad classification of changes within each defined category.
Precise definition of scoring criteria is important when using semiquantitative methods. Even published standards do not necessarily ensure comparable results between pathologists. ISO-10993-6 Annex E standards are considered to be appropriately rigorous when applied to the host response to an implanted device, but this standard only provides guidance with room for individual pathologist interpretation.
For example, a host cellular response often varies in intensity along the host/implant interface. Scoring inflammation severity defined by the number of cells per microscope field sounds rigorous. But should the score be assigned based the maximal observed inflammationwhich may only be present in one area? Or should the score be assigned based on the average inflammation present? And if so, how many microscope fields should be used for averaging? Without these critical details being specified, different pathologists could end up with different scores for the same slide sectionand they would all be correct scores.
In summary, semiquantitative scoring is only as valid as the applied scoring criteria and careful definition of how the scoring criteria are to be applied to the slide section is necessary to ensure reproducible results between pathologists.