National Cancer Centre Singapore, Department of Radiation Oncology, Singapore.
National Cancer Centre Singapore, Department of Radiation Oncology, Singapore.
Breast. 2018 Aug;40:38-44. doi: 10.1016/j.breast.2018.04.013. Epub 2018 Apr 17.
We aim to validate the AJCC 8th edition prognostic staging system for breast cancer in an Asian setting.
Clinico-pathologic information and cancer-specific survival (CSS) outcomes of 6287 stage I to III patients with invasive breast cancer who underwent upfront surgery at SingHealth institutions in Singapore from 2006 to 2014 were analyzed. Survival distributions for the different staging systems were estimated by the Kaplan-Meier method and compared using the log-rank tests. Multivariable Cox proportional hazards models were used, with Akaike Information Criterion (AIC) and Harrell's Concordance Index (C-index) to compare both staging systems. Among patients with positive hormone-receptor status, 84.8% received endocrine therapy. Among the cohort, 60.3% of received chemotherapy; 82.1% of node positive patients received chemotherapy and 86.0% of HER2-enriched patients in whom chemotherapy was also indicated received adjuvant HER2-targeted therapy. Ninety-seven percent of patients received anthracyclines and/or taxanes containing chemotherapy regime.
The median follow up was 64 months. 2921 patients (46.5%) were discordant between the anatomic and prognostic systems of which 363 (5.8%) were upstaged and 2558 (40.7%) were down-staged. For all patients, stages in both the prognostic and anatomic systems were discriminating for 5-year CSS. Controlling for age, ethnicity and receipt of chemotherapy, the prognostic staging system model (AIC = 7538.87, C = 0.79) presented slightly better explanation and concordance of survival times than the anatomic staging system model (AIC = 7607.31, C = 0.77).
The prognostic staging system was better than the anatomic staging system in predicting outcomes but the anatomic system remains relevant due to its ease of use.
我们旨在验证 AJCC 第 8 版乳腺癌预后分期系统在亚洲人群中的适用性。
分析了 2006 年至 2014 年在新加坡 SingHealth 机构接受初始手术的 6287 例 I 期至 III 期浸润性乳腺癌患者的临床病理信息和癌症特异性生存（CSS）结果。通过 Kaplan-Meier 方法估计不同分期系统的生存分布，并使用对数秩检验进行比较。使用多变量 Cox 比例风险模型，使用 Akaike 信息准则（AIC）和 Harrell 一致性指数（C 指数）比较两种分期系统。在具有阳性激素受体状态的患者中，84.8%接受了内分泌治疗。在该队列中，60.3%接受了化疗；82.1%的淋巴结阳性患者接受了化疗，86.0%的 HER2 富集患者也接受了辅助 HER2 靶向治疗。97%的患者接受了含蒽环类和/或紫杉类的化疗方案。
中位随访时间为 64 个月。2921 例（46.5%）患者在解剖和预后系统之间存在不一致，其中 363 例（5.8%）被升级分期，2558 例（40.7%）被降级分期。对于所有患者，两种预后和解剖系统的分期均能区分 5 年 CSS。在控制年龄、种族和接受化疗的情况下，预后分期系统模型（AIC=7538.87，C=0.79）在解释和一致性方面略优于解剖分期系统模型（AIC=7607.31，C=0.77）。
预后分期系统在预测结局方面优于解剖分期系统，但由于其易于使用，解剖系统仍然相关。
