Does biologic subtype and pathologic response after neoadjuvant chemotherapy predict locoregional recurrence?

65 Background: The relative contribution of biological subtype and response to neoadjuvant chemotherapy (NAC) to locoregional recurrence (LRR) is uncertain. We aim to determine if these factors identify a high risk population for LRR. Methods: 233 patients received anthracycline/taxane-based NAC, mastectomy and postmastectomy radiation therapy (PMRT) in 2000-2009 for Stage II-III breast cancer. 53% (n=123) were HR+ (ER or PR+/HER2-), 23% (53) HER2+ (HER2+/HR+ or HR-), and 24% (57) TN (HR-/HER2-). 76% of HER2+ received trastuzumab. Median PMRT dose was 50 Gy to chest wall and regional nodes. Pathologic complete response (pCR) rates were compared using Fisher's exact test. Rates of LRR and distant recurrence (DR) were estimated by Kaplan-Meier methods. Cox regression analysis was performed. Results: Median follow-up was 62 months (range 7-161) with 21 LRR, 84 DR and 58 deaths. pCR rate and 5-year LRR rates were 14% and 7% in the entire cohort, respectively. Significantly more TN and HER2+ patients achieved pCR than HR+ patients (Table 1, p=0.003). TN patients had higher 5-year LRR rate compared to HR+ and HER2+ patients (18% vs. 4% and 6%, p=0.02). The 5-year LRR rate was 0% in pCR patients versus 9% in non-pCR patients (p=0.06). In patients without pCR, TN subtype was associated with increased LRR (23% at 5-year vs. 4% HR+ and 7% HER2+; p=0.001). TN patients without pCR were also associated with increased DR (48% at 5-year vs. 29% HR+ and 30% HER2+, p=0.02). On univariate analysis, TN subtype (HR=2.0, p=0.008), pathologic stage (HR=2.2, p=0.02), and pN+ status (HR=9.3, p=0.03) were associated with increased LRR. Conclusions: Although response to NAC strongly correlates with breast cancer subtype, patients with HR+ and HER2+ breast cancer had favorable rates of LRR regardless of response to NAC, perhaps because of additional postoperative targeted therapy. In contrast, while no LRR was seen in TN patients with pCR, those with poor response to NAC had significantly higher LRR risk, underscoring the need for potential new treatment strategies to improve local control in this population. [Table: see text]