Abstract P1-11-02: Pathologic Response and Survival after Neoadjuvant Therapy for Breast Cancer: A 30-Year Single-Center Study

Abstract

Abstract Background: Triple-negative (TN) and HER2-positive (HER2+) breast cancers have usually a worse prognosis than the luminal subtypes despite an initial chemosensitivity. The aim of this single-center study was to compare pathologic complete response (pCR) rates, disease-free (DFS) and overall (OS) survivals rates after neoadjuvant therapy according to both HER2 and hormonal status. Patients and Methods: Between 1978 and 2008, 461 patients were treated with neoadjuvant therapy at Georges-François Leclerc Cancer Center. Patients were classified in 3 groups: TN (defined by both estrogen/progesterone receptor negative and HER2 negative), HER2+ (3+ in immunohistochemistry or 2+ with amplification in fluorescence in situ hybridization) and HR+ (estrogen and/or progesterone receptor positive without HER2 overexpression or amplification). Median follow-up lasted 7.1 years [range: 0.49-29.8]. Number of events was 213 for DFS calculation and number of deaths was 150 for OS calculation. Response rates were compared using Chi2-tests. Survivals were calculated according to Kaplan-Meier and compared using log-rank tests. Univariate and multivariate Cox proportional hazards models were performed. The multivariate models were internally validated using bootstrapping (400 replications). Results: 86 of 461 were TN (19%), 125 were HER2+ (27%) and 250 were HR+ (54%). Patients with TN and HER2+ breast cancer were younger (p=0.032), had more inflammatory cancer (p=0.033) and aggressive tumors (SBR 3, P<0.001). Pathologic complete response rate (grades 1 and 2 of Chevallier's classification) was significantly higher for TN (22.4%) and HER2+ (29.6%) than in for HR+ (3.6%) (P<0.001). In univariate analysis, the following characteristics were related to a higher pCR rate: smaller clinical size (p=0.029), higher grade tumor (p=0.001) and HER2+ or TN status (P<0.001). In multivariate analysis, only tumor grade (p=0.022) and hormonal/HER2 status (p=0.003) were independently associated with pCR. Median DFS was 4.4years for TN, 7.8y for HER2+ and 9y for HR+ (p=0.003, logrank test). In HER2+ patients, neoadjuvant trastuzumab was associated with a higher DFS (8.65 vs. 3.24y, p=0.002). Patients who achieved a pCR had a higher DFS (p=0.015) than those with only partial pathologic response (median=6.9y vs. 12.7y). In multivariate analysis, pCR remained significant (HR (bootstrapping) = 0.5 [IC95%, 0.28-0.91], p=0.023). Median OS was 6.4y for TN, 15.1y for HER2+ and 13.1y for HR+ (P<0.001, logrank test). Patients who achieved a pCR had a higher OS (p=0.004) than those with non-pCR (median=not reached vs. 10.9y). In multivariate analysis, OS was lower for patients>50years (HR (bootstrapping)=1.79 [IC95%, 1.23-2.60], p=0.002), in TN subgroup (HR=2.46 [IC95%, 1.5-4.03], p=0.001) and when pCR was not achieved (HR=0.28, [IC95%, 0.12-0.66], p=0.003). Conclusion: After neoadjuvant therapy, TN breast cancers have a worse prognosis despite their initial chemosensitivity with a high pCR rate. HER2+ have a lower DFS than HR+/HER2- breast cancers but a better OS, mainly due to anti-HER2 targeted therapies. Pathologic complete response is a strong independent prognostic factor after neoadjuvant therapy for breast cancer. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P1-11-02.