P1-06-02: Correlation between Gene Variants in CYP19 (Aromatase) and TCL1A with Disease and Tolerability Endpoints in the ATAC Trial.

Aim: To assess whether candidate polymorphisms in CYP19 (aromatase) and TCL1A were associated with recurrence, musculoskeletal symptoms (MSKs) or hot-flushes in primary estrogen receptor-positive breast cancer in postmenopausal women treated with tamoxifen or anastrozole. Background. Recently, variants rs10459592 and rs4775936 in CYP19 (aromatase) have been reported to be associated with aromatase inhibitor (AI) response in patients with metastatic breast cancer (Park, et al, CCP, 2011) and variants in CYP19 and at the 3’ end of TCL1A (rs11849538) with AI-induced arthralgia (Ingle, et al, JCO, 2010). These findings require validation before clinical application. The ATAC trial was a prospective randomized double-blind placebo-controlled trial that compared the adjuvant use of anastrozole versus tamoxifen for 5 years. The trial9s detailed efficacy and safety data, long term (10-year) follow-up and high event rate make it an ideal setting for pharmacogenetic single nucleotide polymorphism (SNP) validation studies. Here we report our findings testing for correlations with 3 SNPs in CYP19 and TCL1A with clinical outcomes including any disease recurrence, distant recurrence, MSKs, and hot-flushes. Methods: Genotypes were determined using ABI Taqman assays without knowledge of the patient9s treatment assignment or outcomes. Hazard ratios (HR) and corresponding 95% confidence intervals (CI) were estimated by the proportional hazards regression model and were adjusted for: age, tumor size, grade, and nodal status (for survival analyses) and age, hormone replacement therapy, body mass index, and smoking (for toxicity analyses). Results: 807 patients were genotyped for CYP19 SNPs rs10459592 and rs4775936 and 842 for TCL1A SNP rs11849538. The rs10459592 genotype frequencies were 16.6% Wt/Wt, 51.5% Wt/Vt, and 32% Vt/Vt. There was a trend towards decreased recurrence in Vt/Vt and Vt/Wt compared to Wt/Wt (HR=0.77 [0.50−1.22] and 0.68 [0.42−1.09], p[trend]=0.1), similar for both anastrozole- and tamoxifen-treated patients and also similar for distant recurrence. For the same SNP, there was a trend toward more MSKs in Vt/Vt and Vt/Wt compared to Wt/Wt (HR=1.22 [0.80−1.88] and 1.47 [0.94−2.30], p[trend]=0.08), primarily due to the effect in the tamoxifen cohort (HR=1.88 [0.89−3.99], p[trend]=0.1). The rs4775936 genotype frequencies were 26.0% Wt/Wt, 52.1% Wt/Vt and 21.9% Vt/Vt. Trends in the same direction as those for rs10459592 were found with rs4775936, albeit less significant. The TCL1A genotype frequencies were 77.5% Wt/Wt, 20% Wt/Vt and 2.5% Vt/Vt. No association was observed between TCL1A genotypes and recurrences or MSKs, however, the small number of Vt/Vt patients on tamoxifen had greater incidence of hot flushes (HR=1.99 [0.9−3.91], p=0.09). Conclusion: These data do not support an association between the TCL1A 3’ variant and AI-induced MSKs independent of severity. Our data do support the CYP19 SNP rs10459592 and the loosely linked rs4775936, which previously were associated with letrozole efficacy, as probably being associated with better disease outcome on anastrozole and tamoxifen and with worse drug side effects. The degree of association requires definition in further data sets. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-06-02.