P4-06-06: Morphological Categories of ICC-Detected CK+ Cells in Bone Marrow Have Different Prognostic Impact in Breast Cancer.

Abstract

Abstract Introduction: Morphological evaluation increases the specificity of immunocytochemical analysis (ICC) of disseminated tumor cells (DTC) in bone marrow (BM). In the Oslo1 cohort of primary breast cancer (BC) patients (pts), ICC-positive cells detected in BM were prospectively classified into 4 morphological groups: Tumor cells (TC) and un-interpretable cells (UIC i.e. degenerated/fragmented cells), both significantly associated with reduced survival; and hematopoietic cells (HC) and questionable HC (Q-HC), not affecting outcome (Borgen 1999, Naume 2004). The present study revisits these categories, comparing their significance according to therapy and tumor subtype. Materials and methods: Mononuclear cells (MNC) from BM collected at surgery of 761 pts were analyzed by standard ICC, including 2×106 BM MNC stained for cytokeratin by AE1AE3 monoclonal antibodies, and corresponding isotype specific negative controls (neg co; 2×106 BM MNC for most). Cells were classified as TC, UIC, Q-HC or HC. Primary tumor analyses included pTpN stage, grade, hormone receptor (HR) and HER2. Patients were followed for median 99 months. Survival analyses were performed by Kaplan Meyer and Cox regression analysis. Results: Frequency of TC in specific test (AE1AE3) was 16.9%, UIC 14.0%, Q-HC 23.4% and HC 27.2%. By analyzing the entire cohort and numerically correcting for similar cells in neg co, only cells classified as TC or UIC predicted systemic relapse. Analysis of no adjuvant treated (NO-ADJ; n=390) and adjuvant treated (ADJ; n=361) pts separately, revealed that in NO-ADJ pts only those harboring Q-HC experienced reduced survival (DDFS: HR 2.2, 95% CI 1.2−4.1, p=0.01; BCSS: HR 2.7, 95% CI 1.3−5.5, p=0.006). InADJ pts, only TC and UIC predicted relapse (HR 1.8, 95% CI 1.2−2.9, p=0.004) and death (HR 2.0, 95% CI 1.3−3.1, p=0.003). In multivariate analysis for NO-ADJ pts, presence of Q-HC by AE1AE3 remained significant (DDFS: HR 2.1, 95% CI 1.1−4.1, p=0.031; BCSS: HR 2.4, 95% CI 1.1−5.1, p=0.024). By subgroup analysis according to HR and HER2 (where available), prognostic impact of TC and UIC were observed only in HR+/HER2− subgroup (n=538; DDFS:HR 1.9, 95% CI 1.2−3.2, p=0.006; BCSS: HR 3.3, 95% CI 1.9−5.6, p>0.001), and not in HER2+ (n=80), or triple negative pts (n=117). In contrast, presence of Q-HC was associated with worse outcome in triple negative pts (n=115; DDFS:HR 2.9, 95% CI 1.4−6.1, p=0.005; BCSS: HR 2.4, 95% CI 1.1−5.2, p=0.026), with no significant prognostic impact in the other two pts groups. Interestingly, pts habouring ≥ 3 HC in AE1AE3 and in neg co together (8.5% of pts), had improved DDFS (HR 0.60, 95% CI 0.41−0.88, p=0.008) and BCSS (HR 0.57, CI .036-0.91, p=0.018) compared to those with fewer or no HC (p=0.029). Only 4.7% of pts with ≥ 3 HC experienced metastasis, versus 20.8% of the other. Conclusion: Morphological DTC subgroups may differ in clinical significance according to biology/primary tumor subtype and therapy status. This emphasizes the importance of separate analyses of DTC categories and further DTC characterization. False positive cells, probably in the B-cell lineage, might signify an immune-related favorable clinical outcome. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P4-06-06.