Could treatment with tamoxifen be superior to aromatase inhibitors in early-stage breast cancer after pharmacogenomic testing? A modeling analysis

Abstract

502 Background: Adjuvant endocrine treatment with upfront aromatase inhibitors yields improved disease-free survival compared to tamoxifen in unselected women with estrogen receptor-positive breast cancer. However, levels of endoxifen, the active tamoxifen metabolite, are known to vary with the number of mutant alleles of the cytochrome P450 CYP2D6 enzyme. We created a Markov model to examine whether the optimal treatment strategy for patients with wild type genetics (wt/wt) might differ from that for patients with homozygous (*4/*4) or heterozygous (wt/*4) mutations. Methods: Annual recurrence risks with aromatase inhibitors and tamoxifen in unselected women were taken from the BIG 1–98 trial. Based on data from Goetz et al., we assumed that the percent of patients who are homozygous and heterozygous for mutations are 6.8% and 21.1%, respectively, and that the hazard ratio (HR) for increased cancer recurrence on tamoxifen among *4/*4 carriers is 1.86 relative to patients with wt/wt or wt/*4 genotypes. Because the efficacy of tamoxifen among wt/*4 patients is not known, we tested the full possible range, from efficacy same as wt/wt patients (Ehet=0) to that of *4/*4 mutation carriers (Ehet=1). Results: In the unselected group, 5-year disease-free survival (5-DFS) with aromatase inhibitors and tamoxifen was 0.840 and 0.813, respectively. Under baseline assumptions, tamoxifen is more effective than aromatase inhibitors among wt/wt patients as long as the effect in heterozygotes is at least 54% (Ehet=0.54) of that in *4/*4 patients. With increasing HR for *4/*4 patients, tamoxifen estimates exceed those of aromatase inhibitors in the wt/wt cohort even at lower assumed Ehet ratios (see table ). Conclusion: In patients without CYP2D6 mutations, modeling suggests that initial treatment with tamoxifen could be superior to treatment with aromatase inhibitors, supporting the use of genetic testing to determine the optimal treatment strategy. [Table: see text] No significant financial relationships to disclose.