Background: The current study aimed to evaluate the predictive performance of the American Joint Committee on Cancer eighth edition staging system in patients with invasive breast cancer based on the Surveillance, Epidemiology, and End Results database.
Subjects, materials, and methods: Patients diagnosed with T1-2N0M0, estrogen receptor-positive, human epidermal growth factor receptor 2-negative breast cancer from 2010 to 2014 were retrospectively recruited in this analysis. Patients were reassigned to different stages according to the anatomic staging system (AS), prognostic staging system (PS), and prognostic and genomic staging criteria downstaging patients with recurrence score (RS) lower than 11 (PGS_RS11). Cox models were conducted for multivariate analyses, and likelihood ratio (LR) χ2, Akaike information criterion (AIC), and Harrell's concordance index (C-index) were calculated for the comparison of different staging systems. Additionally, adjustments were made to generate prognostic and genomic staging criteria downstaging patients with RS lower than 18 (PGS_RS18) and RS lower than 25 (PGS_RS25).
Results: PGS_RS11 was an independent predictor for breast cancer-specific survival, as were PS and AS. Adjusted for age and ethnicity, PGS_RS11 (AIC = 2,322.763, C-index = 0.7482, LR χ2 = 113.17) showed superiority in predicting survival outcomes and discriminating patients compared with AS (AIC = 2,369.132, C-index = 0.6986, LR χ2 = 60.80) but didn't outperform PS (AIC = 2,320.992, C-index = 0.7487, LR χ2 = 114.94). The predictive and discriminative ability of PGS_RS18 was the best (AIC = 2297.434, C-index = 0.7828, LR χ2 = 138.50) when compared with PS and PGS_RS11.
Conclusion: PGS_RS11 was superior to AS but comparable with PS in predicting prognosis. Further validations and refinements are needed for the better incorporation of RS into staging systems.
Implications for practice: Staging systems are of critical importance in informing prognosis and guiding treatment. This study's objective was to evaluate the newly proposed staging system in the American Joint Committee on Cancer eighth edition staging manual, which combined biological and genomic information with the traditional TNM classification for the first time to determine tumor stages of breast cancer. The superiority of the prognostic and genomic staging system was validated in our cohort and possibly could encourage the utility of genomic assays in clinical practice for staging assessment and prognosis prediction.
摘要
背景。本研究旨在基于监测、流行病学和最终结果数据库，评估美国癌症联合委员会第八版分期系统对侵袭性乳腺癌的预测能力。
受试者、材料和方法。回顾性分析 2010 至 2014 年间诊断为 T1‐2N0M0、雌激素受体阳性、人表皮生长因子受体 2 阴性乳腺癌患者的临床资料。根据解剖分期系统 (AS)、预后分期系统 (PS)、对复发评分 (RS) 低于 11 的患者进行降级分期的预后和基因组分期标准(PGS_RS11)，将患者进行重新分期。采用 Cox 模型进行多变量分析，计算似然比 (LR)χ2、Akaike 信息标准 (AIC) 和 Harrell 一致性指数(C 指数)对比各分期系统的效果。此外，调整生成对 RS 低于 18 的患者 (PGS_RS18) 和 RS 低于 25 的患者 (PGS_RS25) 进行降级分期的预后和基因组分期标准。
结果。PGS_RS11 与 PS、AS 一样，都是乳腺癌相关生存率的独立预测因子。校正年龄和种族，PGS_RS11(AIC = 2 322.763, C指数 = 0.748 2, LRχ2 = 113.17)相比 AS(AIC = 2 369.132, C 指数= 0.698 6, LR χ 2 = 60.80)在预测生存结局的结果和区别患者方面的表现更优，但与 PS(AIC = 2 320.992, C 指数 = 0.748 7, LR χ2 = 114.94)结果相当。与 PS 和 PGS_RS11 相比，PGS_RS18的预测和区别能力最佳(AIC = 2 297.434, C 指数 = 0.782 8, LR χ2 = 138.50)。
结论。PGS_RS11 在预测预后方面优于 AS，但与 PS 相当。为了更好地将 RS 融入分期系统，仍需要进一步验证和改进。
实践意义:分期系统对预测预后和指导治疗至关重要。本研究旨在评价美国癌症联合委员会第八版分期手册中新提出的分期系统，这一系统首次将生物学和基因组信息与传统 TNM 分类相结合，用于确定乳腺癌的肿瘤分期。研究结果证实，预后和基因组分期系统具有一定的优势，可将基因组分析推广到临床实践中，用于分期评估和预后预测。
Keywords: Breast malignancy; Multigene assay; Prognosis; Tumor staging.
© AlphaMed Press 2019.