Interpretation of Immunohistochemistry Results of HER2/neu Expression in Invasive Breast Cancer: Evaluation of Interobserver and Intraobserver Agreement

The accurate determination of HER-2 in invasive breast cancer has become a critical issue, particularly in context of the results of Herceptin adjuvant therapy. The aim of this study was to evaluate inter- and intraobserver reproducibility of assessment of HER2/neu immunostaining in invasive breast cancer This study was cross sectional and the convenience sampling method was applied. IHC for HER2/neu was performed on 107 archival invasive breast carcinoma tissues. Protein overexpression was scored independently by 4 pathologists using the criteria recommended by DAKO (0 to 3+ reflecting the intensity and pattern of the immunohistochemical signal present in 10% or more of tumor cells.). After a session in order to train for better evaluation and standardization of the scoring criteria, the samples were evaluated for the second time. A minimum of two months elapsed between the two examinations. The discrepancies in scoring results were reviewed by the 4 observers using a multihead microscope to obtain consensus. The inter- and intraobserver agreements were calculated by kappa statistics. Complete agreement was achieved in 33 (31%) and 40 (37%) of 107 cases in the first and second examinations, respectively. Generally, interobserver agreement was moderate (Kappa=0.44 in the first examination and 0.51 in the second one), with intermediate categories (1+, 2+) showing fair agreement and extreme categories (0, 3+) showing moderate agreement. Intraobserver agreement was also moderate (from the minimum kappa=0.55 to the maximum of 0.60). Considering three groups as negative (0/1+), weakly positive (2+) and strongly positive (3+), interobserver agreement improved but still staying in the moderate level (kappa=0.54 and 0.60 in the first and second time, respectively) with score 2+ group showing the least level of agreement. However, with dividing the cases into two groups as negative (0/1+) and positive (2+/3+), interobserver agreement improved to substantial level (kappa=0.68 and 0.75 in the first and second time, respectively).Also, with this consideration, intraobserver agreement improved to substantial level (0.71 to 0.84). By consensus there were 42 cases (39.2%) that were scored 0, 19 cases (17.8%) scored 1+, 27 cases (25.2%) scored 2+, and 11 cases (10.3%) scored 3+. For 8 cases (7.5%) no consensus was achieved in 6 of them the problem was placing them into 2+ or 3+ category and in 2 cases the problem was placing them into 0 or 1+ category. Our study showed that reproducibility (interobserver agreement) is moderate for evaluation of HER2 status in invasive breast carcinoma by IHC using DAKO criteria. However, for discrimination between negative and positive state, reproducibility (that is in substantial level) is better.