The 8th edition of the American Joint Committee on Cancer staging system provide improved prognostic accuracy in T1-2N1M0 postmastectomy breast cancer patients / 中华肿瘤杂志

ABSTRACT

Objective@#To validate whether the prognostic stage groups by the 8th edition of the American Joint Committee on Cancer (AJCC) staging system provides improved prognostic accuracy in T1-2N1M0 postmastectomy breast cancer patients compared to 7th edition.@*Methods@#a total of 1 823 female patients with T1-2N1M0 breast cancer who underwent mastectomy and axillary lymph node dissection without neoadjuvant chemotherapy were analyzed and restaged according to 8th edition. Univariate analysis of prognostic factors was evaluated by using log-rank test. Multivariate analysis was estimated by using the Cox proportional hazards model. The prognostic accuracy of the two staging systems was compared using receiver operating characteristic (ROC) analyses and the concordance index (C-index).@*Results@#5-year locoregional recurrence rate (LRR) for the whole group was 6.0%, 5-year distant metastasis (DM) rate was 11.5%, 5-year disease-free survival (DFS) was 85.0%, and 5-year overall survival (OS) was 93.1%. Cox analysis showed that 7th edition of the AJCC staging system and progesterone receptor status were independent risk factors for LRR, DM, DFS and OS (P<0.05). Compared with stage by 7th edition, 1 278(70.1%) were assigned to a different prognostic stage group 1 088 (85.1%) to a lower stage and 190 (14.9%) to a higher stage. LRR, DM, DFS and OS were significantly different between prognostic stage ⅠA, ⅠB, ⅡA, ⅡB and ⅢA according to 8th edition of the AJCC staging system(P<0.001). Prognostic stage had significantly higher C-indexes and provided better estimation of prognosis compared to stage by 7th edition of the AJCC staging system (P<0.001).@*Conclusion@#The prognostic stage groups of 8th edition AJCC staging system has superior prognostic accuracy compared to 7th edition in T1-2N1M0 breast cancer, and has better clinical therapeutic guidance value.