Effect Modifiers of Low-Dose Tamoxifen in a Randomized Trial in Breast Noninvasive Disease.

Abstract

Low-dose tamoxifen halved recurrence after surgery in a phase III trial in breast noninvasive disease without increasing adverse events. We explored the effect of low-dose tamoxifen in clinically relevant subgroups, including menopausal status, estradiol levels, smoking, body mass index, and proliferation of baseline lesion. Incidence of invasive breast cancer or ductal carcinoma in situ was the primary endpoint. HRs and interaction terms were estimated using Cox models. A favorable HR and 95% confidence interval (CI) could be demonstrated for postmenopausal status (HR = 0.30; 95% CI, 0.11-0.82 vs. HR = 0.73; 95% CI, 0.30-1.76 in premenopausal women; P interaction = 0.13), women with estradiol less than 15.8 pg/mL, presence of menopausal symptoms at baseline, and never smoking (P interaction = 0.07), although the interaction P value was >0.05 for all characteristics. Efficacy was similar in all body mass index categories. Tumors with Ki-67 above the median level of 10% had a greater benefit (HR = 0.27; 95% CI, 0.09-0.81) than those with Ki-67 ≤10% (HR = 1.58; 95% CI, 0.45-5.60; P interaction = 0.04). The efficacy of low-dose tamoxifen seems to be greater in postmenopausal women and in women with lower estradiol levels. Benefits appear to be larger also in women with menopausal symptoms, never smokers, and tumors with Ki-67 >10%. Our results by menopausal status provide important insight into low-dose tamoxifen personalized treatment, although caution is necessary given their exploratory nature. Observation of an improved response in tumors with Ki-67 >10% is consistent but the use of the marker in this setting is investigational.See related commentary by Fabian, p. 3510.