淋巴结清扫数目对食管鳞癌患者预后的影响分析

Impact of the number of lymph node examined on the prognosis of esophageal squamous cell carcinoma

目的：探讨淋巴结清扫数目对食管鳞癌患者预后的影响。方法：采用回顾性病例对照研究方法。收集2005年1月至2013年3月天津医科大学肿瘤医院收治的628例行根治术治疗的食管鳞癌患者的临床病理资料。患者均行经右胸食管癌根治术。观察指标：（1）手术和术后病理学检查情况。（2）随访和生存情况。（3）淋巴结清扫数目对食管鳞癌患者预后的影响。（4）影响食管鳞癌患者预后的因素分析。（5）亚组分析。术后通过门诊、电话和信件方式进行随访，了解患者生存情况。随访时间截至2018年2月。偏态分布的计量资料以M（范围）表示。采用受试者工作特征（ROC）曲线确定淋巴结清扫数目评估患者预后的最佳截点值。采用Kaplan-Meier法绘制生存曲线并计算生存率，生存分析和单因素分析采用Log-rank检验，多因素分析采用COX比例风险模型。结果：（1）手术和术后病理学检查情况：628例患者中，472例行Ivor-Lewis手术，156例行McKeown手术；肿瘤直径≤3.5 cm 284例，肿瘤直径〉3.5 cm 344例。 628例患者共清扫11 139枚淋巴结，平均淋巴结清扫数目为18枚/例（2～78）枚/例，中位淋巴结清扫数目为16枚/例。628例患者中，肿瘤分化程度高分化、中分化、低分化分别为48、469、111例，肿瘤浸润深度 T0～1、T2、T3、T4a期分别为30、119、260、219例，淋巴结转移程度N0、N1、N2、N3 期分别为349、173、69、 37例，淋巴结转移率分期rN0、rN1、rN2、rN3期分别为349、184、54、41例。（2）随访和生存情况：628例患者均获得术后随访，随访时间为3～144个月，中位随访时间为36个月。628例患者术后1、3、5年生存率分别为82.4%、53.7%、41.3%。（3）淋巴结清扫数目对食管鳞癌患者预后的影响：ROC曲线结果显示淋巴结清扫数目评估患者预后的最佳截点值为16枚。628例患者以淋巴结清扫数目16枚为截点，淋巴结清扫数 目〈16枚和≥16枚患者术后5年生存率分别为36.7%和45.1%，两者生存情况比较，差异有统计学意义（χ^2=9.527，P〈0.05）。进一步将淋巴结清扫数目≥16枚患者根据其淋巴结清扫数目中位值23枚分为淋巴结清扫数目为16～23枚和〉23枚患者，其结果显示：淋巴结清扫数目〈16枚、16～23枚和〉23枚患者的术后5年生存率分别为36.7%、41.2%和50.3%，3者生存情况比较，差异有统计学意义（χ^2=10.588，P〈0.05）。其中淋巴结清扫数目〈16枚与16～23枚患者比较，差异有统计学意义（x2=4.419，P〈0.05）；淋巴结清扫数目16～23枚与〉23枚患者比较，差异无统计学意义（χ^2=1.413，P〉0.05）。淋巴结清扫数目≤ 23枚与〉23枚患者术后5年生存率分别为38.6%和50.3%，两者生存情况比较，差异有统计学意义（χ^2=5.885，P〈0.05）。（4）影响食管鳞癌患者预后的因素分析。单因素分析结果显示：患者年龄、吸烟史、体质量指数、肿瘤直径、淋巴结清扫数目、肿瘤浸润深度、淋巴结转移程度、淋巴结转移率分期是影响食管鳞癌患者预后的相关因素（χ^2=5.454，4.875，7.669，10.691，10.588，30.612，59.780，76.565，P〈0.05）。多因素分析结果显示：年龄、肿瘤直径、淋巴结清扫数目、肿瘤浸润深度、淋巴结转移率分期是影响食管鳞癌患者预后的独立因素（风险比=1.268，1.300，0.762，1.354，1.357，95%可信区间：1.034～1.556，1.038～1.629，0.662～0.878，1.183～1.549，1.089～1.692，P〈0.05）。（5）亚组分析：在279例伴有淋巴结转移患者中，淋巴结清扫数目〈16枚、16～23枚、〉23枚患者术后5年生存率分别为23.7%、 19.4%、39.5%，3者生存情况比较，差异有统计学意义（x2=8.397，P〈0.05）。其中淋巴结清扫数目〈16枚与16～23枚患者比较，差异无统计学意义（χ^2=0.389，P〉0.05）；淋巴结清扫数目为16～23枚与〉23枚患者比较，差异有统计学意义（χ^2=5.425，P〈0.05）。淋巴结清扫数目≤23枚与〉23枚患者术后5年生存率分别为21.9%和39.5%，两者生存情况比较，差异有统计学意义（χ^2=7.942，P〈0.05）。在349例无淋巴结转移患者中，淋巴结清扫数目〈 16枚、16～23枚、〉23枚患者术后5年生存率分别为45.6%、60.3%、59.2%，3者生存情况比较，差异有统计学意义（χ^2=9.755，P〈0.05）。其中淋巴结清扫数目〈16枚与16～23枚患者比较，差异有统计学意义 （χ^2=8.208，P〈0.05）；淋巴结清扫数目16～23枚与〉23枚患者比较，差异无统计学意义（χ^2=0.284，P〉0.05）。淋巴结清扫数目≤23枚与〉23枚患者术后5年生存率分别为51.1%和59.2%，两者生存情况比较，差异无统计学意义（χ^2=1.147，P〉0.05）。结论：淋巴结清扫数目是影响食管鳞癌患者预后的独立因素，推荐至少清扫16～23枚淋巴结。对于伴有淋巴结转移患者，应清扫〉23枚淋巴结；对于无淋巴结转移患者，超过23枚的广泛淋巴结清扫不能显著改善患者预后。

Objective：To investigate the effect of the number of lymph nodes exmined （NLNE） on the prognosis of esophageal squamous cell carcinoma （ESCC）. Methods：The retrospective casecontrol study was conducted. The clinicopathological date of 628 ESCC patients who underwent radical resection in the Tianjin Medical University Cancer Institute and Hospital from January 2005 to March 2013 was collected. Patients underwent radical resection of ESCC through right thorax Observation indicators： （1） surgical and postoperative pathological exminations; （2） followup and survival situations; （3） effect of NLNE on the prognosis of ESCC; （4） factors analysis affecting prognosis of ESCC patients; （5） subgroup analysis. Followup using outpatient exmination, telephone interview and mail was performed to detect postoperative survival up to February 2018. Measurement data with skewed distribution were described as M （range）. Receiver operating characteristic （ROC） curve analysis was used to determine the appropriate cutoff of the NLNE. The survival curve and survival rate were respectively drawn and calculated by the Kaplan-Meier method, and the survival analysis was done by the Log-rank test. Multivariate analysis was done by the Coxproportional hazard model. Results：（1） Surgical and postoperative pathological exminations： 472 and 156 patients underwent respectively Ivor-Lewis and Mckeown operations. There were 284 patients with tumor diameter ≤ 3.5 cm and 344 patients with tumor diameter 〉 3.5 cm. The total NLNE was 11 139 for all of the 628 patients, with an average NLNE of 18 per case（range, 2-78 per case） and a median NLNE of 16 per case. Of 628 patients, high, moderate and lowdifferentiated tumors were respectively detected in 48, 469 and 111 patients; staging T0-1, T2, T3 and T4a of depth of tumor invasion in 30, 119, 260 and 219 patients; N0, N1, N2 and N3 of degree of lymph node metastasis in 349, 173, 69 and 37 patients; rN0, rN1, rN2 and rN3 of rate of lymph node metastasis in 349, 184, 54 and 41 patients. （2） Followup and survival situations： all the 628 patients were followed up for 3-144 months, with a median time of 36 months. The 1, 3 and 5year survival rates were 82.4%, 53.7% and 41.3%, respectively. （3） Effect of NLNE on the prognosis of ESCC： ROC curve showed that the appropriate cutoff value of the NLNE was 16. Using NLNE = 16 as a cutoff value, 5year survival rate was respectively 36.7% in patients with NLNE 〈 16 and 45.1% in patients with NLNE ≥ 16, with a statistically significant difference in survival （χ^2=9.527, P〈0.05）. According to a median NLNE of 23, the patients with NLNE ≥ 16 were further divided into patients with 16 ≤ NLNE ≤ 23 and NLNE 〉 23. Results showed that 5year survival rate in patients with NLNE 〈 16, 16 ≤ NLNE ≤ 23 and NLNE 〉 23 was respectively 36.7%, 41.2% and 50.3%, with a statistically significant difference in survival among them （χ^2=10.588, P〈0.05）, between patients with NLNE 〈 16 and 16 ≤ NLNE ≤ 23 （χ^2=4.419, P〈0.05）. There was no statistically significant difference between patients with 16 ≤ NLNE ≤ 23 and NLNE 〉 23 （χ^2=1.413, P〉0.05）. Fiveyear survival rate in patients with NLNE ≤ 23 and NLNE 〉 23 was respectively 38.6% and 50.3%, with a statisctically significant difference （χ^2=5.885, P〈0.05）. （4） Factors analysis affecting prognosis of ESCC patients： results of univariate analysis showed that age, smoking history, BMI, tumor diameter, NLNE, depth of tumor invasion, degree and rate of lymph node metastasis were related factors affecting the prognosis of ESCC patients （χ^2=5.454, 4.875, 7.669, 10.691, 10.588, 30.612, 59.780, 76.565, P〈0.05）. Results of multivariate analysis showed that age, tumor diameter, NLNE, depth of tumor invasion and rate of lymph node metastasis were independent factors affecting the prognosis of ESCC patients [HR=1.268, 1.300, 0.762, 1.354, 1.357, 95% confidence interval （CI）： 1.034-1.556, 1.038-1.629, 0.662-0.878, 1.183-1.549, 1.089-1.692, P〈0.05]. （5） Subgroup analysis： among 279 patients with lymph node metastasis, 5year survival rate in patients with NLNE 〈 16, 16≤ NLNE ≤ 23 and NLNE 〉 23 was respectively 23.7%, 19.4% and 39.5%, with a statistically significant difference among them （χ^2=8.397, P〈0.05）, between patients with 16≤ NLNE ≤ 23 and NLNE 〉 23 （χ^2=5.425, P〈0.05）. There was no statistically significant difference between patients with NLNE 〈 16 and 16 ≤ NLNE ≤ 23 （χ^2=0.389, P〉0.05）. Fiveyear survival rate in patients with NLNE ≤ 23 and NLNE 〉 23 was respectively 21.9% and 39.5%, with a statisctically significant difference （χ^2=7.942, P〈0.05）. Among 349 patients without lymph node metastasis, 5year survival rate in patients with NLNE 〈 16, 16 ≤ NLNE ≤ 23 and NLNE 〉 23 was respectively 45.6%, 60.3% and 59.2%, with a statistically significant difference among them （χ^2=9.755, P〈0.05） and between patients with NLNE 〈 16 and 16 ≤ NLNE ≤ 23 （χ^2=8.208, P〈0.05）. There was no statistically significant difference between patients with 16 ≤ NLNE ≤ 23 and NLNE 〉 23 （χ^2=0.284, P〉0.05）. Fiveyear survival rate in patients with NLNE ≤ 23 and NLNE 〉 23 was respectively 51.1% and 59.2%, with no statisctically significant difference （χ^2=1.147, P〉0.05）. Conclusions：The NLNE is an independent factor affecting the prognosis of ESCC patients, and at least 16 to 23 lymph nodes should be dissected. For patients with lymph node metastasis, and more than 23 lymph nodes should be dissected. For patients without lymph node metastasis, more than 23 lymph nodes dissection cannot obviously improve the prognosis of patients.