分子分型对1～3个腋窝淋巴结转移早期乳腺癌患者局部或区域复发风险的预测价值

Predictive value of molecular subtyping for loco-regional recurrence in early breast cancer patientswith one to three positive axillary lymph nodes

目的研究1～3个腋窝淋巴结转移早期（病理N1期）乳腺癌分子分型与术后局部或区域复发（LR）间关系，探讨改进个体化辅助放疗指征。方法回顾分析1998—2009年本院手术的547例病理T1～2N1M0期乳腺癌根治术后未行放疗患者。根据免疫组织化学、荧光原位杂交检测结果分为LuminalA、LuminalB、HER-2过表达和三阴型，比较LR复发率（LRR）及LR生存率（LRFS），并结合临床病理特征对其LR风险进行分组分析。Kaplan-Meier法计算LRR、LRFS并Logrank法检验和单因素预后分析，多因素预后分析采用Cox模型。结果LuminalA、LuminalB、HER-2过表达和三阴型分别占30．0％、48．6％、9．3％和12．1％。随访率97．1％，随访时间满5、10年者分别为334、127例。单因素分析显示HER-2过表达型、三阴型LR风险比LuminalA型高，5年LRR分别为19．0％、14．9％与5．3％（X2=4．28、5．02，P：0．026、0．015），LRFS分别为73．5％、80．6％与91．1％（X2=7．27、4．77，P=0．005、0．021）。多因素分析显示HER-2过表达型、三阴型、年龄≤35岁、pT：期病变是LRR及LRFS的不良预后因素（X2=2．29、2．08、18．22、6．86，P=0．020、0．016、0．001、0．005及X2=1．90、1．41、8．58、3．94，P=0．006、0．025、0．002、0．039）。有以上0、1和i〉2个危险因素者10年LRR分别为4．3％、14．1％和31．9％（X2=28．03，P=0．000）。结论分子分型有助于个体化区别1～3个腋窝淋巴结转移病理N．早期乳腺癌患者间LR风险，具有多个危险因素者应接受术后放疗。

Objective To investigate the relationship between molecular subtypes of breast cancer and postoperative loco-regional recurrence （LR） in early breast cancer patients with 1--3 positive axillary lymph nodes （ pN1 ） and to improve the individualized indications for post-mastectomy radiotherapy （PMRT） in these patients. Methods A total of 547 patients with pT1-2 N1M0 breast cancer, who received masteetomy between December 1998 and December 2009 in Sun Yat-sen University Cancer Center, were retrospectively analyzed. None of them received adjuvant radiotherapy after mastectomy. The patients were divided into luminal A group, luminal B group, HER-2-overexpressing group, and triple-negative group according to the molecular subtypes of breast cancer determined by immunohistochemistry and fluorescence in situ hybridization. The patients in different groups were compared in terms of LR rate （LRR） and LR-free survival （LRFS） , and the risk factors for LR were analyzed in combination with clinical and pathological features. The Kaplan-Meier method was adopted to calculate LRR and LRFS;the Logrank test was used for survival difference analysis and univariate prognostic analysis. The Cox proportional hazards model was used for multivariate prognostic analysis. Results The luminal A group, luminal B group, HER-2-overexpressinggroup, and triple-negative group accounted for 30. 0% , 48.6%, 9. 3% , and 12. 1% , respectively, of all the patients. The follow-up rate was 97. 1% ;334 patients were followed up for at least 5 years, and 127 were followed up for at least 10 years. Univariate analysis showed that, compared with the luminal A group, the HER-2-overexpressing group and triple-negative group had significantly higher 5-year LRRs （19. 0% vs 5.3% , X2 = 4. 28, P = 0. 026 ; 14. 9% vs 5.3% , X2=5.02, P = 0. 015） and significantly lower LRFSs （73.5% vs 91.1%, X2=7.27, P=0.005;80.6% vs 91.1%,X2=4.77, P=0.021）. Multivariate analysis revealed that HER-2 overexpression, triple-negative phenotype, age of ≤ 35 years, and stage pT2 were poor prognostic factors for survival （ LRR and LRFS） （ X2 = 2. 29, 2. 08, 18.22, and 6. 86, P = 0. 020, 0. 016, 0. 001, and 0. 005;X2 = 1.90, 1.41, 8.58, and 3.94, P = 0. 006, 0. 025, 0. 002, and 0. 039）. The 10-year LRRs of patients with 0, 1, and ≥2 of the above risk factors were 4. 3% , 14. 1% , and 31.9%, respectively （ X2= 28.03, P = 0. 000 ）. Conclusions Molecular subtyping is helpful for individualized evaluation of LR risk in early breast cancer patients with 1--3 positive axillary lymph nodes （oN,）. PMRT should be recommended for the patients with 2 or more risk factors for LR.