Expression of retinoblastoma protein in breast cancer metastases to sentinel nodes: evaluation of its role as a marker for the presence of metastases in non-sentinel axillary nodes, and comparison to p16INK4a.

The aim of this study was to evaluate the role of retinoblastoma protein (pRb), alone and in combination with p16, as a predictive marker for metastases in non-sentinel nodes in cases where the sentinel node showed metastatic breast carcinoma. Paraffin blocks of lymph nodes from 48 patients with metastatic breast carcinoma were immunostained with a monoclonal antibody to retinoblastoma protein (PharMingen). Results were compared with known prognostic parameters of the primary tumor, estrogen and progesterone receptor status, proliferation index, and p16 (DAKO) expression. Lymph nodes from 38 of the 48 (79%) cases were pRb positive. There was no correlation of pRb staining alone with the primary tumor parameters studied or the proliferative index of the metastatic tumor. In 16 patients with both a sentinel node biopsy and an axillary lymph node dissection, 8 (50%) had metastatic breast carcinoma. The sentinel nodes of three of these eight patients (38%) were pRb negative (positive predictive value of 60% vs. 73% for p16). The remaining eight patients (50%) had no metastases in non-sentinel nodes, even though their sentinel nodes had metastatic breast carcinoma; six of these eight patients (75%) were pRb positive (negative predictive value of 55% vs. 83% for p16). pRb and p16 staining results combined showed that pRb-negative/p16-positive cases were associated with non-sentinel node metastases (positive predictive value of 100%) as well as poor prognostic parameters. Patients with the opposite staining profile (pRb positive and p16 negative) were mostly without non-sentinel node metastases (negative predictive value of 75%). Cases negative for both pRb and p16 were consistently associated with a better prognostic phenotype and absence of additional axillary node metastases. In conclusion, the presence or absence of pRb in sentinel nodes is of little predictive value for non-sentinel node metastases unless taken in conjunction with the presence of p16 staining. Instead, it appears to enhance the positive predictive value of p16 in determining the presence of non-sentinel node metastases. Due to the limited subgroup sample size in this study, clinical guidelines cannot be suggested as yet, but further research focused on the pRb-negative/p16-positivie and pRb-negative/p16-negative phenotypes may yield beneficial results.