44P Prognostic Value of Detection Method by Breast Cancer (BC) by Phenotype

ABSTRACT Introduction Mammographic screening has shown to reduce BC mortality, but approximately 30% of the mortality reduction remains unexplained. Method of detection has independent prognostic value and we here assess its value in function of BC phenotypes. Patients and methods Primary operable invasive BC patients treated at the Multidisciplinary BC Centre UZ-Leuven (01/2000 – 12/2009) were included. Preoperative systemic therapy (n = 407), metastases at diagnosis (n = 228), male patients (n = 28) or external surgery (n = 530) were exclusion criteria. Details on treatment and pathology are described in detail elsewhere (Brouckaert et al., Breast Cancer Res Treat., 2009 May; 115(2): 349-58). BCs were allocated to screen detected (n = 1610) when diagnosed through organized or opportunistic screening (missing n = 91) and to palpable when palpable by the treating physician (missing n = 57). We defined BC phenotypes as Luminal A (ER and/or PR positive HER2 negative, grade 1/2), Luminal B1 (idem, grade 3), Luminal B2 (ER and/or PR and HER-2 positive), HER-2 like (ER/PR negative, HER-2 positive) and Triple negative (2011 St. Gallen criteria). Multivariable Cox proportional hazard models were used adjusting for age, size, grade, nodal status, ER, PR and HER2 receptor status and adjuvant therapy (see Table 1 ). The proportional hazard assumption for the Cox model was evaluated by including time-dependent covariates in the models, which were included when significant. Results Median follow-up 6.1 years. Table 1 . Adjusted HR (Screening versus non-) Distant metastasis free survival (p-value) Breast cancer specific survival (p-value) All (n = 4227) 0.42 (0.03) 0.49 (0.13) Luminal A (n = 2338) 0.79 (0.29) 0.63 (0.2) Luminal B1 (n = 924) 0.58 (0.03) 0.69 (0.23) Luminal B2 (n = 289) 0.37 (0.07) 0.53 (0.39) HER-2 like (n = 173) 0.39 (0.09) 0.46 (0.24) Triple negative (n = 412) 0.44 (0.04) 0.23 (0.01) Conclusion BC screening may be criticized to reduce mortality by simply diagnosing indolent tumors (length bias). However, the prognostic value of method of detection seems to have increasing value in increasingly more aggressive BC phenotypes. Method of detection may also prove to be a new and independent prognostic tool to assess outcome in triple negative BC. References on demand. Disclosure All authors have declared no conflicts of interest.