Analysis of factors related to adjuvant chemotherapy decision in early breast cancer patients with intermediate recurrence score.

作者：Feilin, Qu; Jiayi, Wu; Xiaosong, Chen; Ou, Huang; Jianrong, He; Li, Zhu; Weiguo, Chen; Yafen, Li; Kunwei, Shen

来源：Journal of Clinical Oncology, 2017, 35(15_suppl): e12032-e12032.

DOI：10.1200/jco.2017.35.15_suppl.e12032

<jats:p> e12032 </jats:p><jats:p> Background: The 21-gene Recurrence Score (RS) assay has been routinely used to guide systemic chemotherapy in patients (pts) with estrogen receptor positive, node-negative early breast cancer (EBC). However, there has been less clarity in recommendation of adjuvant chemotherapy (ACT) in pts with intermediate RS (IRS). According to the definitions of IRS in retrospective NSABP B14, B20 trials and prospective validation of TAILORx trial, we adopted the broaden cut-offs of 11-30 in our study. We sought to analyze the factors related to ACT decision in this subset of pts. Additionally, the role of Adjuvant!Online (AOL) was evaluated in ACT decision. Methods: A cohort of 564 pts with RS of 11-30 was retrospectively analyzed from January 2014 to October 2016 in Ruijin Hospital Shanghai Jiaotong University School of Medicine. The Oncotype DX RS, a RT-PCR 21-gene assay, was performed on RNA extracted from formalin-fixed paraffin-embedded tissue. The AOL was used to determine pts’ clinical risk. Predictive factors of chemotherapy usage in different clinical risk catogories were also assessed. Results: 267 (47.3%) pts received chemotherapy. Age, tumor grade, pathologic type, pT, pN, molecular subtype and RS were significantly correlated with ACT decision ( p &lt;0.05). These factors were all independent predictors of ACT usage ( p&lt;0.05) in mutivariable model. AOL was successfully measured in 504 (89.4%) pts, of whom 279 (49.5%) were at low clinical risk and 225 (39.9%) had high clinical risk. The distribution of RS correlated significantly with AOL clinical risk catogories. The predictive factors of ACT administration in high and low clinical risk subgroups were highly consistent with the overall population, except pT in high clinical risk pts. Discordances between clinical risk and ACT decision were found in 158 pts (28.0%), probably due to different age, molecular subtype and RS. Conclusions: Age, tumor grade, pathologic type, pT, pN, molecular subtype and RS were independent predictors of ACT administration in EBC pts with IRS. AOL should not be used alone to aid chemotherapy decision in this subset of pts. </jats:p>

• 出版日期2017-5-20
• 单位上海交通大学