Proficiency Testing for Mismatch Repair Immunohistochemistry in Canada
Emina Torlakovic, Blaise Clarke, Catherine J Streutker, Sara Hafezi-Bakhtiari, Steve E Kalloger, Mulligan Anna Marie, Robert Wolber, Blake Gilks, Aaron Pollett. University Health Network, University of Toronto, Toronto, ON, Canada; St. Michael Hospital/University of Toronto, Toronto, ON, Canada; Vancouver Costal Health, Vancouver, BC, Canada; University of British Columbia, Vancouver, BC, Canada; Mt. Sinai Hospital/University of Toronto, Toronto, ON, Canada
Background: Current methods (clinical history and tumor morphology) to predict MMR deficiency (dMMR) are suboptimal. In view of recommendations of reflex MMR-IHC screening policies and in recognition of the growing predictive role of dMMR we undertook to determine the current quality indicators of dMMR testing in Canadian laboratories. This is part of an endeavor by the Canadian Association of Pathologists/Association Canadienne des pathologistes (CAP-ACP) National Standards Committee for Complex Laboratory Testing to introduce a Canadian consensus statement and national guideline on dMMR testing. The EQA run was conducted by the Canadian Immunohistochemistry Quality Control, which is partly supported by Canadian Partnership Against Cancer (CPAC).
Design: TMA was constructed from the 39 tumor samples with known germline status. Histologically normal tissues were included as additional tissue cores from each sample. Unstained slides were sent to 14 laboratories that were identified as performers of the IHC testing for MLH1, MSH2, MSH6, and PMS2. Expert assessment was conducted by five experienced pathologists. Percent correct results, false-positive and false-negative rate, specificity and sensitivity were calculated.
Results: Success with histologically normal tissues had linear-by-linear association with specificity with tumor samples (p=0.002). False absence of nuclear staining (FANS) in the tumor strongly correlated with the absence of staining in the normal tissue (r=0.959, p<0.0001, Pearson's R for all four tests). Sensitivity and specificity of each test are shown in Table 1.
|Sensitivity||88-100 (96)||100 (100)||93-100 (99)||90-100 (99)|
|Specificity||64-100 (96)||75-100 (93)||38-98 (73)||19-100 (86)|