Lymphocytic infiltrate is associated with favorable biomarkers profile in HER2-overexpressing breast cancers and adverse biomarker profile in ER-positive breast cancers.

The value for lymphocytic infiltration (LI) has been increasingly recognized for tumor assessment. In breast cancer, however, the overall significance of LI remains poorly defined, probably due to its heterogeneity. A large cohort of breast cancer was evaluated for the degree of LI and its association with traditional pathologic factors, biomarker expression, and cancer subtypes. The number of CD8 cytotoxic effector and FoxP3 regulatory T cell (Treg) was evaluated in those cases with high LI. High LI was associated with negative ER and PR but positive HER2 and EGFR expression (p < 0.001 for all). In ER-positive cancers, high LI was associated with poor prognostic features including higher grade, the presence of necrosis, and lymphovascular invasion (LVI) (p = 0.007 for LVI and <0.001 for the others). Conversely, LI correlated with smaller tumor size, a good prognostic feature (p = 0.046) in HER2+ ER-cancers. These observations suggested LI may show opposite prognostic values in different breast cancer subgroups. Interestingly, when the phenotype of LI in these subgroups was evaluated, a strong positive association with intratumoral accumulation of Treg was found in ER-positive cancers (p = 0.003, Rs = 0.319), while the opposite was observed in HER2+ ER-cancers (p < 0.001, Rs = -0.427). Also, in ER-positive cancers, positive associations between peri- and intra-tumoral distribution were found with both CD8 and Tregs (CD8: p < 0.001, Rs = 0.547; Treg: p = 0.001, Rs = 0.460). Nonetheless, in HER2+ ER-cancers, such strong association was found with CD8 (p < 0.001, Rs = 0.766) but not Tregs. The results may implicate a differential intratumoral migration of LI in different subtypes of breast cancer. In summary, the clinical value of LI in breast cancers could be subtype-dependent. In ER-positive cancers, high LI correlated with biologic parameters associated with poor prognosis, whereas in HER2 positive cancers, LI correlated with biologic parameters of favorable prognosis.

P-cadherin and vimentin are useful basal markers in breast cancers.

Basal-like breast cancer (BLBC) is the breast cancer subtype defined by gene profiling and generates keen clinical interest. Immunohistochemical panels using basal cytokeratins and epidermal growth factor receptor are widely adopted for its identification. Nonetheless, there are concerns about the risk for missing some true BLBCs. Both P-cadherin and vimentin have been proposed as BLBC markers, but their usefulness for BLBC classification has not been well documented. In this study, we evaluated by immunohistochemistry their expression in a large cohort of breast carcinoma. Cancers expressing vimentin or P-cadherin showed BLBC-related morphological features (high grade, presence of necrosis, and lymphocytic infiltration; P < .001 for all except P = .006 for vimentin with lymphocytic infiltration) and immunohistochemical profile (P < .001 for all markers tested except P = .007 for vimentin with human epidermal growth factor receptor 2). Concordantly, they were significantly associated with BLBC (P < .001 for both). Nonetheless, they did not appear to be good stand-alone BLBC markers. Compared with the commonly used reference panel, the specificity (95.9%) and sensitivity (43.1%) of coexpression of vimentin and P-cadherin were better than most single markers or their combinations tested. Moreover, their coexpression was significantly associated with basal features in non-BLBCs and worse disease-free survival in triple-negative breast cancers (hazard ratio, 2.232; P = .027). This raised the possibility that the vimentin and P-cadherin combination can be used to identify BLBC especially those that were missed by the commonly used basal cytokeratins and epidermal growth factor receptor panel. Together, P-cadherin and vimentin could be adjunctive to the commonly used immunohistochemical surrogates for BLBC identification.