Inter-observer agreement in reporting HER 2 Neu protein over expression by immunohistochemistry.

Introduction: HER 2 Neu protein overexpression and its detection by immunohistochemistry (IHC) has become quiet critical because of its relevance in regards to Herceptin treatment. This peer review was done at a tertiary care center, which aimed at determining the inter-observer variation among five pathologists and evaluating the degree of agreement between them.

Aims: The aim of our study was to determine the reproducibility of HER 2 Neu system of reporting in breast cancer cases and determine inter-observer variability among five pathologists at a tertiary care center. To compare the results with similar studies done at other centers.

Settings And Design: Retrospective descriptive study.

Materials And Methods: Hematoxylin and Eosin (H and E) and IHC stained slides of 104 cases of carcinoma breast, on which HER 2 Neu status had been reported were reviewed. The time period for selection was from January 2010 to December 2011 (2 year period). Five pathologists reviewed the H and E and IHC slides independently and scored the results on a specially designed work sheet. Kappa values for inter-observer variation and Cornbach's alpha for internal consistency were calculated.

Statistical Analysis Used: SPSS 20.0 (IBM). not known.

Results: Complete agreement was seen between all five pathologists in 70 cases (70/104) = 67%. Agreement between four pathologists was seen in 78 cases (78/104) = 75%. Agreement between three pathologists was seen in 92 cases (92/104) = 88%. The global value for kappa co efficient for agreement between two pathologists was 0.706 and Cornbach alpha for internal consistency of reporting in the department was 0.987. CONCLUSION/KEY MESSAGES: Our departmental peer review indicated that there is good inter-observer concordance (agreement between two pathologists) and there is strong overall internal consistency of reporting for HER 2 Neu reporting by IHC. Our results are comparable to International reported data of similar studies.