Application of Image Analysis for Daily Monitoring of Positive Controls in Clinical Immunohistochemistry
Heather Neufeld, Richard Berendt, Dragana Pilavdzic, Bayardo Perez-Ordonez, Barry Ziola, Blake Gilks, John Garratt, Carol Cheung, John Krahn, Svein Carlsen, Emina Torlakovic. University of Saskatchewan, Saskatoon, SK, Canada; University of Alberta, Edmonton, AB, Canada; McGill University, Montreal, QC, Canada; University of British Columbia, Vancouver, ON, Canada; Lions General Hospital, Vancouver, BC, Canada; University of Toronto, Toronto, ON, Canada
Background: Digital Image Analysis (IA) is a relatively new application for the evaluation and quantification of clinical immunohistochemistry (IHC) tests. Calibrated standardized positive controls are especially desirable for Class II IHC markers that are quantitative in nature. This study compares the use of image analysis by evaluating i) the effect of tissue surface area on reproducibility, and ii) the use of cell line-based vs. tissue-based positive controls in clinical IHC practice for estrogen receptor (ER) and progesterone receptor (PR).
Design: The Aperio Scanscope IA system was used to evaluate tissue-based controls made from human tissues. Formalin-fixed paraffin embedded cell line-based controls from selected human breast cancer cell lines were also stained and compared. Groups of sequential, currently used, on-slide IHC controls of human tissue-based controls from four major Canadian academic laboratories were examined to observe the effects of tissue types (breast, endometrium, cervix) and surface area of tissue samples. IA results of whole tissue sections were compared with results derived from only a portion of tissue sections.
Results: A statistically significant variation of H-scores was observed between independent IHC runs using all tissue-based controls for both ER and PR (F-test, p<0.01 and F-test, p<0.01 respectively). Comparison with cell line-based control samples showed that the variation was largely due to tissue heterogeneity, with little to no contribution by tissue type. However, tissue size was an important determinant of variation with larger tissue samples exhibiting less variation than smaller tissue samples.
Conclusions: IA can be used for daily monitoring of IHC QC. However, the selection of control type and tissue surface area are important factors in determining control performance. Cell line-based controls were superior to human tissue-based controls due to their greater homogeneity. However, if tissue controls are monitored by IA, then IA evaluation should be performed on larger (rather than smaller) tissue sections in order to reduce the effects of tissue heterogeneity on results.
Category: Quality Assurance
Tuesday, March 5, 2013 9:30 AM
Poster Session III # 244, Tuesday Morning