Effect modifiers in a randomized phase III trial of low-dose tamoxifen in breast preinvasive disease.

Abstract

1500 Background: Low-dose tamoxifen (babytam) at 5 mg/day for 3 years decreases local or contralateral recurrence by 52% in women with hormone sensitive breast pre-invasive neoplasia after surgery (DeCensi et al JCO 2019). Here we report the results of exploratory analyses to assess whether the benefit of babytam varies among subgroups of patients defined by individual characteristics. Methods: Post-hoc subgroup analyses were performed according to a mixed approach based on the test for interaction and biological plausibility. Incidence of invasive breast cancer or DCIS was the primary endpoint. HRs were estimated using Cox proportional-hazards modeling. Results: Age at menopause, smoking status and Ki-67 exhibited a significant interaction with treatment. Specifically, the effect of babytam was greater in women aged > 50y (n = 293, HR = 0.27, 95%CI: 0.10-0.73) than in women aged ≤50y (n = 207, HR = 0.86, 0.35-2.07), p-interaction = .09. Never smokers (n = 307) had a greater benefit than former (n = 68) or current smokers (n = 97): HR = 0.28, 0.11-0.70 vs HR = 0.57, 0.09-3.45 vs HR = 1.51, 0.41-5.64, respectively (p = .05). Tumors with Ki-67 above the median level of 10% (n = 133) had a greater effect (HR = 0.27, 0.09-0.81) than Ki-67 ≤10% (n = 145, HR = 1.58, 0.45-5.60, p = .04). Weaker statistical interactions (p > .1) were also found for waist circumference and hot flashes (HF) at baseline. Women with waist circumference ≥89 cm (metabolic syndrome, n = 208) had a greater effect (HR = 0.22, 0.07-0.78) than women < 89 cm (n = 228, HR = 0.61, 0.25-1.46). Compared with placebo and no HF, babytam effect was stronger in women with HF (HR = 0.13, 0.02-0.96) than in women on babytam and no HF (HR = 0.50, 0.24-1.03) or placebo and HF (HR = 0.72, 0.31-1.69, log-rank p-trend = .004). Additional subgroups according to obesity, family history of breast or ovarian cancer, alcohol use, extent of surgery, radiotherapy for DCIS, ER and HER2 expression, positive margins and treatment compliance showed no significant heterogeneity of treatment. Conclusions: Exploratory analyses showed a trend to a higher effect of babytam in women aged 50 or older, never smokers, women with hot flashes or abdominal obesity and tumors with Ki-67 above 10%. Our results provide insight into the efficacy of babytam towards a personalized preventive approach. Clinical trial information: NCT01357772.