Tissue microarray-based immunohistochemical study can significantly underestimate the expression of HER2 and progesterone receptor in ductal carcinoma in situ of the breast.

The accuracy of immunohistochemical (IHC) analysis on tissue microarray (TMA)-based studies largely depends on the uniformity of the staining pattern for a given antibody and minimal intratumor heterogeneity of a given tumor. Our study was designed to investigate the concordance of expression in TMA and whole sections of estrogen receptor (ER), progesterone receptor (PR) and HER2 using IHC analysis for ductal carcinoma in situ (DCIS) of the breast. Seventy-five consecutive cases of DCIS were retrieved, reviewed and used to construct the TMA. IHC analysis of the expression of ER, PR, and HER2 were performed on TMA and whole sections of the corresponding cases, and the results were compared. The specificity and sensitivity for TMA-based assays were 87.0, 75.9, 90.6 and 90.4%, and 76.1, 27.3 for ER, PR and HER2, respectively. The concordance and discordance were 89.3, 76.0 and 72.0%, and 6.7, 13.3 and 16.0% for ER, PR, HER2, respectively. The kappa values were 0.83, 0.89 and 0.42 for ER, PR and HER2, respectively. The non-concordance rates were inversely related to core number, with 46.67, 22.67 and 11.56% for one core, two cores, and three cores, respectively, per marker per case (p < 0.001), but not associated with tumor size. Our results showed that the intratumor heterogeneity and the number of cores have a great impact on the results of TMA-based studies. Increasing the number of tissue cores per case may help improve the accuracy and concordance with whole section results. Although TMA remains an effective tool for translational research, we should be cautious in our interpretation of these results.