Abstract P6-09-01: Identification of tumors with an indolent disease course: MammaPrint ultralow signature validation in a retrospective analysis of a Swedish randomized tamoxifen trial

Abstract

Abstract Background Better tools are needed to identify breast cancer patients at very low risk of dying from their disease. We applied a previously specified 'Ultralow risk' threshold for the FDA-cleared MammaPrint 70-gene expression score to predict long-term absence of breast cancer-specific mortality in a Swedish randomized trial of breast cancer patients treated with tamoxifen (Tam) versus not and followed for more than 25 years. Methods Between 1976-1990 the Stockholm Tamoxifen (STO) trial enrolled and randomized node negative breast cancer patients with tumor size less than 30 mm to receive 2 years of Tam versus not, without regard to hormone receptor status. In the Tam-treated arm, patients without relapse at 2 years were further randomized to receive 3 additional years of Tam versus no additional endocrine therapy. From the original STO randomized trial cohort, about half (778 cases) had remaining formalin-fixed paraffin-embedded primary tumor blocks for additional tumor characterization and MammaPrint analysis. In this validation dataset, which now has >25 years follow-up, breast cancer-specific survival was assessed between MammaPrint scored 'Ultralow risk', 'Low risk' or 'High risk' categories by Kaplan-Meier analyses and multivariate Cox proportional hazard modeling, adjusting for treatment, age, period of diagnosis, tumor size, grade, receptor (ER, PR, HER2) and Ki-67 status. Results: In this unscreened patient population MammaPrint scored 15% of patients as 'Ultra-low risk', 43% as 'Low risk' and 42% as 'High risk'. At 20 years, a statistically significant difference in survival between risk categories was seen for all patients (log rank, P=0.0001), the Tam treated arm (log rank, P=0.0088) and the untreated arm (log rank, P=0.014). Ultra-low risk patients have significantly lower risk of disease-specific death relative to Low and High risk patients in our multivariate Cox analysis. Among Ultra-low risk patients there were no deaths out to 15 years in the Tam-treated arm; by 20 years, disease-specific survival in Tam-treated and untreated arms were 94% and 90%. The majority of Tam-treated patients received only 2 years of treatment, with ∼35% going on to receive 5 years of treatment. All Ultra-low risk cases were HR+HER2-. 78% were luminal A by PAM50; but only 31% of luminal A were Ultra-low risk. Invasive ductal (no-special-type) carcinomas were the most frequent, but lobular, tubular, invasive papillary and invasive cribriform subtypes were enriched and mucinous types were absent. Conclusions: Node-negative, T<3cm breast cancer cases designated as 'Ultralow risk' by the 70-gene MammaPrint assay have minimal risk of metastatic death within the first 15 years if given a short (2+ year) course of adjuvant Tam. While a small but very late (>15 y) metastatic recurrence risk exists if left untreated, 'Ultralow risk' status could identify women for whom lumpectomy alone and a short course of adjuvant endocrine therapy is sufficient. Given the high frequency (∼50%) of such 'Ultralow risk' tumors detected in modern screening cohorts (e.g. MINDACT), a substantial fraction of newly diagnosed women could benefit from more targeted and less aggressive local and adjuvant therapy. Citation Format: Esserman LJ, Thompson CK, Yau C, van 't Veer LJ, Borowsky AD, Tobin NP, Nordenskjöld B, Fornander T, Stål O, Benz CC, Lindström LS. Identification of tumors with an indolent disease course: MammaPrint ultralow signature validation in a retrospective analysis of a Swedish randomized tamoxifen trial. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P6-09-01.