Histopathological Variables on a Scale to Best Refine Prognosis in Breast Cancer Patients.
Isabella Castellano, Silvia Beatrice, Caterina Marchio', Luigi Chiusa, Elena Allia, Anna Vandone, Riccardo Arisio, Antonio Durando, Michela Donadio, Anna Sapino. University of Turin, Italy; Ospedale San Giovanni Battista, Molinette, Turin, Italy; Sant'Anna Hospital, Turin, Italy
Background: Breast cancer encompasses different lesions with heterogeneous clinical behaviour. Universally acknowledged prognostic factors are size, grade of differentiation, lymph-node status, estrogen and progesterone receptor (ER/PR) expression, proliferation index and HER2 expression. Recently molecular signatures have been proposed as better predictors of outcome, at least in ER+ patients. We here investigate whether a score based on routine histopathological/immunophenotypical parameters would discriminate between two categories of patients: those with good prognosis from those with poor prognosis.
Design: 1162 breast cancer cases diagnosed between 1994 and 2004 were retrieved and reviewed. Histological grade, vascular invasion, size, margins were assessed. Complete clinical follow-up data were collected. Immunohistochemistry for ER, PR, HER2, Ki-67, androgen receptors was performed. Univariate and multivariate analyses were sequentially used to identify among all clinico-pathological parameters those which were independent predictors of overall and disease-free survival (OS/DFS). Score values ranging between -1 and 2 were applied to each of the independent predictors, depending on their power of statistical correlation with OS and DFS at Cox analysis. A second-round univariate analysis followed.
Results: First-round multivariate analysis identified size, vascular invasion, number of metastatic lymph-nodes and percentage of androgen receptor positivity as independent prognostic factors. Score values were assigned to each parameter as follows: ≤ 3 metastatic lymph-nodes=1; >3 metastatic lymph-nodes=2; size <15mm=1; size>15mm= 2; absence of vascular invasion=1; presence of vascular invasion=1; AR negative tumour=0; AR positive tumour= -1. The final “weighted-score” ranged between 1 and 5. At second-round univariate analysis values of weighted-score between 1 and 3 identified a group of patients with good prognosis, whereas values of 4 and 5 corresponded to subgroups of patients with poorer prognosis.
Conclusions: A weighted-score based on routine histopathological/immunohistochemical parameters significantly discriminates between two categories (good versus poor prognosis/low-risk versus high-risk of recurrence) and may represent a valuable tool to best refine prognosis in breast cancer patients.
Tuesday, March 1, 2011 9:30 AM
Poster Session III # 41, Tuesday Morning